Check out Symmetric Chess, our featured variant for March, 2024.


[ Help | Earliest Comments | Latest Comments ]
[ List All Subjects of Discussion | Create New Subject of Discussion ]
[ List Latest Comments Only For Pages | Games | Rated Pages | Rated Games | Subjects of Discussion ]

Comments/Ratings for a Single Item

LatestLater Reverse Order EarlierEarliest
Sac Chess. Game with 60 pieces. (10x10, Cells: 100) [All Comments] [Add Comment or Rating]
💡📝Kevin Pacey wrote on Wed, Aug 5, 2020 05:33 PM UTC:

Actually, I don't know if (10x10) International Checkers is more popular than standard 8x8 checkers at present. The 10x10 variant just kind of got that name somehow.

Having 'International Chess' as 10x10 would simply be following with that precedent, besides the thought that 10x10 may indeed be the best board size, as John Brown would have it in the second link I gave before.


H. G. Muller wrote on Wed, Aug 5, 2020 05:26 PM UTC in reply to Kevin Pacey from 05:08 PM:

Well, as long as it would not be more popular than 8x8 orthodox Chess, any claim to the name 'International Chess' would be outright ridiculous.

That you want to restrict the competition to 10x10 is as arbitrary as only taking chess variants into account that have a Cannon as one of the pieces.


💡📝Kevin Pacey wrote on Wed, Aug 5, 2020 05:08 PM UTC in reply to H. G. Muller from 07:06 AM:

Interesting. The wiki on International Draughts mentions early on that it is synonymous with 'Polish Draughts' or 'International Checkers':

https://en.wikipedia.org/wiki/International_draughts

I'd agree that to be called International Chess, such a variant really should be already popular internationally. Yet, I would have it as 10x10 (like for its checkers namesake), so it couldn't be orthodox (8x8) chess.

Otherwise, like for International Checkers, I'd think that besides being 10x10 there should be no pieces added/dropped to the board once the game begins. So, perhaps the first such 10x10 variant that becomes popular enough internationally should grab the title by being re-named 'International Chess' (at least until it begins to have some sort of steep decline, like chess itself might one day).

At the moment the only contender for the moniker (10x10) 'International Chess' I'm aware of might be (10x10) Grand Chess, if only since it has been promoted a lot - but it is not yet near enough popular in most nations. Finally I'd note that I've seen somewhere on this CVP website (on a rules page for a 10x10 CV that uses rotating spearmen) the claim that 'many scholars' agree that if 8x8 chess is to be further developed, it will be by a 10x10 CV:

https://www.chessvariants.com/large.dir/contest/cenchess.html


H. G. Muller wrote on Wed, Aug 5, 2020 07:06 AM UTC:

I thought that 'International Chess' is already used as a synonym for orthodox Chess, to contrast it to variants with a clear regional binding, such as Xiangqi, Shogi, Makruk. (Which are also known as Chinese / Japanese / Thai Chess.) No one outside the tiny community of CV players would ever say 'orthodox Chess'.

And I never heard anyone speak about 'International Checkers'; everyone seems to call that 'International Draughts'. In Dutch we just call that game 'Dammen', no doubt a distorted form of the French name 'Jeu de Dames', and what you know as Checkers we call 'Amerikaans/Engels Dammen'.

The logical requirement for being called 'International' is that it should be significantly popular in a large fraction of all countries, without one country or continent having a much larger fraction of its population play it than any other. If 80% of the people playing it would be distributed (approximately proportionally) over European countries, it would be 'European' rather than 'International'. That doesn't really seem to depend on how large the board is.

So to be called 'International' it would have to be the one item of its kind that has the largest global spread. I suppose you could quantify the concept of global spread by defining it as the fraction of the population that plays it in the country that is half-way down the list that orders countries by this fraction.

Of course 'kind' is only loosely defined, but it refers to what the 'International' predicate is applied to. In 'International Chess' that would be any game that qualifies as Chess. If you want to restrict it to 10x10 variants, you would have to call it 'International 10x10 Chess'.


💡📝Kevin Pacey wrote on Wed, Aug 5, 2020 03:09 AM UTC:

Thanks for your description of my style, Carlos. I've tried to play in many styles (notably for chess) over the years, but I still find that at lower levels of skill than elite, aggression (if available in the position) often pays off most heavily.

There is a question I've had for some time now. 10x10 checkers (at least one variant of it) has been called 'International Checkers'. When naming Sac Chess, though it uses all the classic compound pieces, I avoided calling it 'International Chess' because Sac Chess was totally unproven as yet (besides that it would have been immodest). I did mention on my Chess Federation of Canada Discussion Board blog though that were Sac Chess ever to take off, I wouldn't mind if it were re-named to that instead.

Anyway, my question would be, has anyone more knowledgeable ever given thought to what characteristics a 10x10 CV should have, were it ever to be thought worthy of being called 'International Chess' (a name I find a bit ambiguous, incidentally, much like 'International Checkers').


Carlos Cetina wrote on Mon, Aug 3, 2020 06:40 PM UTC:

Thanks, Kevin, I already read the game. It shows very well the dangers that exist since the opening. Playing carelessly Black enters after just 4 moves into a near-deadly whirlwind. Your style of play is clearly very aggressive trying to do as much damage as possible in the minimum amount of time.

I share 2 games that I played vs the applet imitating its style of play. The first testifies that the AI does make castlings. The second took a long time because I intentionally sought to reach a final of kings and pawns to test pawn's promotion on both sides.

1.Ng4 Jb7 2.Nd4 Ji7 3.c4 Nd7 4.Jc3 Ng7 5.h4 Mc9 6.Jh3 Mh9 7.e4 Jg6 8.f4 Jd6 9.i4 Nf6 10.b4 Md7 11.Mb3 Cc9 12.Mi3 c7 13.Bf3 Ne6 14.Be3 Jde7 15.Nxe6 Jxe6 16.Nxf6 Jexf6 17.d4 Mhg7 18.g4 h6 19.Cd3 Bh8 20.Cg3 O-O 21.O-O Ac8 22.f5 Jgf7 23.e5 J6e7 24.Ac2 Af9 25.f6 Jexf6 26.exf6 Mxf6 27.Cgf5 Mfe6 28.Cxi8+ Ki9 29.Cxg7 Bxg7 30.Mj5 Ch8 31.Jj4 Ad6+ 32.M°g3 Ade7 33.Jd5 Ad7 34.Jxe6 Jxe6 35.Cg6 Bxd4 36.Cxh8+ Bxh8 37.Ji6 Bxb2 38.Jxj8+ Kh8 39.Ai8#

1.Jb4 Jb7 2.Ji4 Ji7 3.Jh4 Ng7 4.Ni4 f6 5.Jc4 Mh9 6.Nb4 Af8 7.f5 Jd6 8.Jxd6 Axd6 9.Af3 Nd7 10.Ad5 Mc9 11.Axd6 Bxd6 12.e4 Jh7 13.Bf4 Cg9 14.Bxd6 Cxd6 15.Ch2 Mf8 16.Cd2 Ne5 17.d4 Nf7 18.Cf4 Mcd7 19.Cxd6 Mxd6 20.Mc2 Cc9 21.Me3 Cd7 22.g4 Ne5 23.Ac2 Cb6 24.dxe5 Cxe3+ 25.Qxe3 Mdf7 26.e6 Me7 27.e5 fxe5 28.Cxe5 Mexe6 29.Bf3 Me7 30.O-O Jf6 31.M°g3 Jxe5 32.M°xe5 Mff7 33.Nc6 Md7 34.Jg6 Mg9 35.Jf6 Mf8 36.Qxj8 b7 37.Nd4 d6 38.M°g3 Mxf6 39.Qg5 Mgf7 40.Qf4 Sg9 41.Ng5 M7d7 42.a5 Qxa5 43.Ra2 Qc5 44.Rxa8 Af8 45.M°b5 Ne6 46.fxe6 Mxf4 47.M°xc5 Mxc5 48.M°xf4 Axf4+ 49.Ki2 Mxg5 50.Ad2 Sf8 51.Axf4 Sxf4 52.Mh2 M°d8 53.Sd2 Rg9 54.Sf2 Sf6 55.b5 M°ee9 56.Be2 g6 57.Sxf6+ M°xf6 58.Rf2 M°eg7 59.b6 Sc9 60.Nb5 Mxe6 61.Nc7 Me4 62.Sc1 Mg3+ 63.Kj2 Rb8 64.Rxb8 Sxb8 65.Nxd9 Sc9 66.Nc7 Se9 67.Bb5 Sf8 68.c4 Md3 69.Sf1 M°ff7 70.Rf3 Me5 71.Sd1 M°ff6 72.Sxd6 M°xi3+ 73.Kxi3 Mxf3 74.Sd9+ Kf10 75.Sd10+ Kf9 76.Mxf3 Sxf3 77.Bxe8+ Kf8 78.Sf10+ Rf9 79.Sxf9+ Ke7 80.Sxf3 i6 81.Sf7+ M°xf7 82.Bxf7 Kxf7 83.g5 i5 84.Ki4 Kg7 85.Kxi5 Kh7 86.j5 Ki7 87.h4 Kh7 88.j6 Ki8 89.Kh6 h7 90.Ne8 Kh8 91.j7 Kg8 92.Ki7 Kf7 93.Nc7 Kg7 94.j8 Kg8 95.Ki8 Kg7 96.j9 Kf7 97.j10=A Kg8 98.c5 h6 99.gxh6 Kf7 100.h7 Kf8 101.h8 Kg8 102.h9 Kf8 103.h10=A+ Kf7 104.Af8+ Kxf8 105.Ag7+ Kxg7 106.Ki7 Kg8 107.Kh6 Kf7 108.Nb5 Kf6 109.Nd6 c6 110.Nxb7 Kf7 111.Nd8+ Kf6 112.b7 Ke7 113.Nxc6+ Kf6 114.b8 Kf7 115.b9 Ke6 116.b10=A Kf7 117.Nd8+ Kf6 118.c6 Kf5 119.c7 Kf6 120.c8 Kf5 121.c9 g5 122.c10=A gxh4 123.Acd9 h3 124.Adc10 h2 125.Acd9 h1=A+ 126.Ah5+ Axh5+ 127.Kxh5 Kf4 128.Ab4+ Kf5 129.Ad4#


💡📝Kevin Pacey wrote on Mon, Aug 3, 2020 02:10 AM UTC:

@Carlos:

I posted a game of mine vs. the Sac Chess applet in my previous reply, in case you missed it (it had my name given as 'panther').


panther wrote on Sun, Aug 2, 2020 06:22 PM UTC in reply to Carlos Cetina from 05:48 PM:

I tried to play the applet set at 4 or 3 ply, but it seemed too slow right at move one (I also was asked to stop a Script each time, early on, but that did not help). So, I played it at 2 ply (the default). I don't know how to post the moves I played with the diagram so as to be able to replay it (if that's possible), but here's the moves that I cut and pasted after the game (I had the White side - unfortunately the machine played somewhat blindly - maybe giving some small hope that huge CV Sac Chess can resist even some other engines for a while, vs. truly strong humans):

  1. f5 Ji7 2. e5 Jj9 3. Ng4 b7 4. Bd5 Mb8 5. Bxg8 Kxg8 6. Nf6 Kf9 7. Qi6 h7 8. Nxh7 Jh8 9. Nxi9 Jh7 10. Nxg10 Jxi6 11. Nxe9 Kxe9 12. O-O Ch8 13. Bf4 Jj6 14. d5 Ji6 15. Nd4 a6 16. Rbe2 Jg7 17. Rgf2 c6 18. e6 d6 19. Bxd6 Md8 20. Bc7 cxd5 21. Bxd8 Kxd8 22. e7 fxe7 23. Ne6 Kd7 24. Axd5 Kc8 25. Rf4 Jh6 26. Adc5 Kb8 27. Aa7#

Carlos Cetina wrote on Sun, Aug 2, 2020 05:48 PM UTC:
I just did the test and the result is that the Applet says ***I resign!*** highlighted in red, so I don't think you have to worry, HG, because there could be a bug.

On the other hand, I already changed the definition of the King from KisO2 to KisjO2.

I use Chrome on Windows 10.

I would very much like, Kevin, if you would share a game of yours vs the Applet!

H. G. Muller wrote on Sun, Aug 2, 2020 05:12 PM UTC:

Still something strange happened, because it would never play a move when it says resign. That it doesn't print 'I resign' is indeed no big deal, but if that doesn't work, there is no telling what else might not work, so it had me worried a bit. Great that it works now, let's hope it stays that way, for everyone.


💡📝Kevin Pacey wrote on Sun, Aug 2, 2020 04:52 PM UTC in reply to H. G. Muller from 08:59 AM:

@ H.G.: Today I tried the sequence 1.Nb4 Ka7 (illegal) 2.Qa6 (illegal) followed by clicking on 'Play it!' and then 'Move'. This time I noticed it printed out 'I resign' (highlighted in red), a bit ABOVE Carlos' diagram.

I suppose it's possible I missed 'I resign' on most of my earlier test tries simply because I was looking for the words 'I resign' below the diagram instead, i.e. where the moves 1.Nb4 Ka7 2.Qa6 were printed out. Otherwise, I'm still not sure why Black's king took my queen consistently on the first day I tried the test. Anyway, I can live with it even if there is a bug somewhere, rather than my hallucinating. Thanks for the help!


H. G. Muller wrote on Sun, Aug 2, 2020 08:59 AM UTC in reply to Kevin Pacey from 06:12 AM:

If nothing happens at all, the script might have stopped because encountered an error, and in that case it would be possible to get information about it. Just hit the F12 key after you pressed 'Move' and nothing happened, and an area will open at the bottom of your browser window, with a number of tabs in it. Select the tab 'Console' if that is not already selected. Error messages should appear in there. You can close the area again by another F12.


💡📝Kevin Pacey wrote on Sun, Aug 2, 2020 06:12 AM UTC:

I wrote something and discovered I wasn't logged in at the time I finished the post. Hopefully it can be recovered by an editor.


panther wrote on Sun, Aug 2, 2020 06:11 AM UTC in reply to H. G. Muller from 05:34 AM:

Now when I try entering 1.Nb4 Ka7 (illegal) 2.Qa6 (illegal) and then click on 'Play it!' followed by clicking on 'Move', nothing happens at all, not even an 'I resign' message. I tried this 3 times tonight and it happened 3 times that way.

I suppose when I originally saw the Black king take my Queen the first day I tried this test, something may have been different.

On the other hand, I have been told I have schizophrenia, since once in a while I see things incorrectly (such as words) now and then (for example) - though I've never seen a chess or CV position incorrectly, when I've been awake, as far as I know.

I'm not sure how to check which exact version of MS Windows Explorer I have as my browser, if that would help instead.


H. G. Muller wrote on Sun, Aug 2, 2020 05:34 AM UTC in reply to Kevin Pacey from 12:02 AM:

I am also running Windows 7 (on a desktop). I tried it with MS Internet Explorer (update version 11.0.170), but I still get the 'I resign' message. This makes it a complete mystery.

Does anyone else have this problem with the Diagram?

That the piece values differ slightly from run to run is normal; the diagram determines those by sampling a number of randomly generated positions.


💡📝Kevin Pacey wrote on Sun, Aug 2, 2020 12:02 AM UTC in reply to H. G. Muller from Sat Aug 1 07:16 AM:

@ H.G.: I'm using a MS Explorer browser (from my Windows7 laptop).

When I did what you suggested as another test, the value displayed for the king was '311 (c00)', so it would seem that it's considered royal, according to what you wrote.

[edit: Note that when I did the test a second time, the pieces were given slightly different values for some reason. For the king, '316 (c00)' was given, for example.]


H. G. Muller wrote on Sat, Aug 1, 2020 07:16 AM UTC in reply to Kevin Pacey from 01:51 AM:

That is odd indeed. What web browser are you using? I tried it with FireFox and Chrome, and in both cases it said "I resign" without moving.

One other test we can do: when you open the piece table by clicking on "here" under the diagram, and then click on the "move" header of the move column, it should switch to displaying the piece values it estimated. (Don't take those too seriously! They are often very inaccurate, but good enough for playing against someone who desn't know the piece values either.) Behind the King it should also print "(c00)", though, to indicate that it is royal. One of the explanations for what you see is that it somehow did not get through that the King is royal.


💡📝Kevin Pacey wrote on Sat, Aug 1, 2020 01:51 AM UTC in reply to H. G. Muller from Fri Jul 31 11:29 AM:

That's odd. Just now I entered the moves 1.Nb4 (square lit up as usual) Ka7 (illegal) 2.Qa6 (illegal) and then clicked on 'Play It!', and next I clicked on 'Move'. At that point the Black king took my queen on a6.


H. G. Muller wrote on Fri, Jul 31, 2020 06:58 PM UTC:

Note that the castling move that is now defined for the King (isO2) describes castling with the furthest piece in that direction. Which in this case is the Judge. If you would want it to castle with the Rook you would have to specify isjO2 . (I used the j for this, which normally means 'jumping', because in a sense you jump over the Judge while looking for the castling partner.)

For user input the diagram is not very pedantic, and allows you to castle with the Rook anyway, but it allows every other illegal move as well, so that doesn't mean a whole lot. As it is the AI would never castle with the Rook, however.


H. G. Muller wrote on Fri, Jul 31, 2020 11:29 AM UTC in reply to Kevin Pacey from 03:15 AM:

This is definitely not as it is supposed to behave. The chekmate detection does seem to work, though, as can be tested by first playing the (illegal) moves 1.Nb4 Ka7, then switching on the AI and playing 2.Qa6. It then says "I resign" without playing a move. After it lost its King, and it is its turn to move, it should say "It seems I have already lost".

No idea why that wouldn't have worked in the game you played. That it doesn't see mate-in-1 coming is normal; the default lookeahead depth is only 2 ply + quiescence search, and for mated-in-1 the King is only captured on ply 4. (So 3 ply + QS would be enough to see it.)

In any case the play will be quite weak; it is intended to be a demo for people who just have seen the rules for the first time, to get an idea what the game is like. Setting it to 4 ply would make it slightly tougher, but even then a strong player should not have much dificulty beating it. (An for large games like SAC Chess it might think too long.)


💡📝Kevin Pacey wrote on Fri, Jul 31, 2020 03:15 AM UTC:

I played it a complete game, though I didn't save it. It may have been on a very low number of ply look-ahead (i.e. the default level). I don't know if it's a bug, but when it let me checkmate it in one move, the losing king took the mating piece attacking it one square away, and it continued the game after I took its king (at which point I abandoned the game). Presumably playing it using several ply look-ahead higher would give me a much tougher time.


Carlos Cetina wrote on Fri, Jul 31, 2020 02:57 AM UTC:

OK, you are welcome.


💡📝Kevin Pacey wrote on Fri, Jul 31, 2020 02:35 AM UTC:

Thank you Carlos! I will try to play it sooner or later!


Carlos Cetina wrote on Thu, Jul 30, 2020 08:41 PM UTC:

Kevin:

I configured the Muller's applet to play your variant. Hope you like it and find it interesting to face its AI. For me it is a great resource to enjoy doing something that I like.

 

files=10 ranks=10 promoZone=1 promoChoice=A Q M C S R M° J B N graphicsDir=/graphics.dir/alfaerie/ squareSize=52 graphicsType=gif pawn:P:ifmnDfmWfceF:pawn:a3,b3,c3,d3,e3,f3,g3,h3,i3,j3,,a8,b8,c8,d8,e8,f8,g8,h8,i8,j8 knight:N:N:knight:c2,h2,,c9,h9 bishop:B:B:bishop:d2,g2,,d9,g9 rook:R:R:rook:b2,i2,,b9,i9 queen:Q:Q:queen:e2,,e9 cardinal:C:BN:cardinal:b1,i1,,b10,i10 marshall:M:RN:chancellor:a1,j1,,a10,j10 amazon:A:QN:amazon:d1,g1,,d10,g10 judge:J:WFN:pegasus:a2,j2,,a9,j9 missionary:M°:BW:promotedbishop:e1,f1,,e10,f10 sailor:S:RF:promotedrook:c1,h1,,c10,h10 king:K:KisjO2:king:f2,,f9

I share the next game I played vs it driving me the white side.

1.Ng4 Jb7 2.Nd4 Ji7 3.h4 Nd7 4.Jh3 Ng7 5.c4 Mc9 6.Jc3 Mh9 7.Mc2 Jc5 8.Mh2 Jg6 9.b4 Jd6 10.f4 a6 11.a5 Nf6 12.e5 Nxg4+ 13.Jxg4 Jde7 14.Be4 Jgf7 15.M°h3 Md7 16.Nf5 Nxf5 17.Jxf5 Cc9 18.Jxe7 Mxe7 19.Jd5 Mb7 20.Bc3 Mg7 21.Kf1 Ch9 22.Jc5 Mc7 23.d5 h6 24.e6 Je7 25.Bxg7 fxg7 26.Jb6 Af8 27.Jxc7 bxc7 28.Qf3 Rb7 29.M°c3 Bh8 30.M°e5 Ca7 31.M°d4 Cb8 32.Cc3 i6 33.M°e5 Sg9 34.Cd4 Ri7 35.Mg4 d6 36.exd7 cxd7 37.f5 d6 38.M°e6 Jd8 39.Mg6 Ah7 40.g5 Cf8 41.Rf2 hxg5 42.Mxg5 Ah6 43.i5 Aj7 44.Ce2 g6 45.f6 Ch7 46.M°e7 Cxg5 47.Axg5 Rxe7 48.fxe7+ Jf7 49.Axi6 Axi6 50.Qxi6 Bxb2 51.Mxb2 Qg7 52.Sg1 Ae9 53.Bxg6 Jf6 54.Rxf6+ Qxf6 55.Sf2 Ag7 56.Cd4 Axi6 57.Sxf6+ Sf8 58.Sxf8+ Kg10 59.Axi6 Sc9 60.Ai10#

King, Queen, Rook, Bishop, Knight and Pawn are labeled as usual. The remain thus:
Amazon: A
Marshall: M
Cardinal: C
Sailor: S
Missionary: M°
Judge: J

I renamed the Archbishop as Cardinal so I could use the A to label the Amazon.


Carlos Cetina wrote on Wed, Jul 15, 2020 01:07 AM UTC:

Thank you very much!


Greg Strong wrote on Tue, Jul 14, 2020 07:53 PM UTC:

Ok, that's fixed.


Carlos Cetina wrote on Tue, Jul 14, 2020 06:41 PM UTC:

@Greg:

I just noticed that in both standard and alternate piece sets the labels for Sailor (R+F) and Missionary (B+W) are reversed.

They are currently incorrectly labeled thus:

Alt text for a graphic image = M

Alt text for a graphic image = S

They must be correctly labeled thus:

Alt text for a graphic image = S

Alt text for a graphic image = M

Could you please take a look at it?


💡📝Kevin Pacey wrote on Sun, Jul 12, 2020 08:21 PM UTC:

Okay, thanks. I hadn't realized the implications of that default checkbox.

Incidentally, Tim O'Lena recently submitted a couple of Zog-related submissions to be looked over by editor(s), in case you missed that.


🕸Fergus Duniho wrote on Sun, Jul 12, 2020 08:09 PM UTC:

I just unchecked Uses HTML for you. You should check that only if you are completely depending on HTML to describe your page's formatting.


💡📝Kevin Pacey wrote on Sun, Jul 12, 2020 06:22 PM UTC:

I slightly edited the rules page for Sac Chess, by noting in the introduction that there are (now) 3 rules enforcing presets available.

Unfortunately, it seems the Sac Chess rules page formatting has become messed up, once more.


💡📝Kevin Pacey wrote on Sat, Jul 11, 2020 02:47 AM UTC:

Yes, it appears the formatting issues with all my submissions that were messed up previously (somehow) have been nicely resolved, by one or more of the editors it may well be.

The only possible exception might be with the Notes Section of my Full House Hexagonal Chess rules page - unfortunately I cannot recall if I left sparse/no spacing between what perhaps should be independent paragraphs in the Notes section to that rules page, so perhaps fixing it further, if it should be fixed further, might take some guesswork or arbitrary editing.


Carlos Cetina wrote on Fri, Jul 10, 2020 05:53 PM UTC:

You are welcome, Kevin.

The formatting issues seem to have been fixed. It only remains to mention that the available presets are three.


💡📝Kevin Pacey wrote on Mon, Jun 29, 2020 12:04 AM UTC:

@ Carlos:

Looks good, thanks!

@ Greg:

The Sac Chess rules page is not the only submission of mine that has gotten messed up, due to some formatting changes to CVP's database perhaps. I wasn't in a hurry to draw attention to all of those submissions, in case the onus somehow fell on me to modify what used to be well-formatted submissions of mine.


Carlos Cetina wrote on Sun, Jun 28, 2020 11:39 PM UTC:

@Greg,

The key at the bottom that describes the pieces shows Alfaerie images.

 

So is. I already corrected it. I even put links to the three presets separately so that the player who is going to launch an invitation can choose the one he/she likes.


H. G. Muller wrote on Sun, Jun 28, 2020 09:53 PM UTC:

Yeah, I know. I have to be very careful never to switch off 'Source code', once I have started to do advanced things. Or they would all simply disappear. Embedded JavaScript disappears completely, Diagram definitions become line-filling text, the 'id' names of HTML elements are all forgotten...


Greg Strong wrote on Sun, Jun 28, 2020 09:32 PM UTC:

The formatting probably got messed up somewhere along the way.  It can happen with the editor if you aren't careful.  Fortunately, now that revisions are saved, it will be possible to go back if it happens ...


H. G. Muller wrote on Sun, Jun 28, 2020 09:15 PM UTC:

Some sections are very poorly formatted; even the images are just part the lines. And division into paragraphs is sorely missing in a very long text.


Greg Strong wrote on Sun, Jun 28, 2020 08:56 PM UTC:

You're welcome.

The key at the bottom that describes the pieces shows Alfaerie images.  So I think the preset should default to Alfaerie, particularly since the dragon king/dragon horse are non-standard.  A player can switch to abstract if he desires.


Carlos Cetina wrote on Sun, Jun 28, 2020 08:33 PM UTC:

@Greg,

 Thank you very much!

@Kevin,

The new preset page is already updated. Please let me know if you want us to write any particular text there. You don't have to worry about past game logs being affected because I have used another settings name.

 


Greg Strong wrote on Sun, Jun 28, 2020 05:25 PM UTC:

Ok, I have created a set group for Sac Chess with two Alfaerie options and an Abstract set option.  These use the proper notations for Sac Chess so be sure to update the game courier presets and change the settings name so that existing game logs don't get broken.

Note for Fergus: The graphics Kevin has used for dragon horse and dragon king may not be correct.  The cross within the graphic usually represents royal pieces but I don't see any other options for these pieces.


Carlos Cetina wrote on Sun, Jun 28, 2020 01:02 PM UTC:

I have never seen the word 'compound' used in sequential meaning; I would say the Sissa is an (isosceles) hook mover. And I would not say the Tai Shogi Hook Mover is a (conjunctive) compound of two Rooks. The word 'or' usually implies 'and', and if you consider this operations on the move sets, the 'conjunctive compound' of R and B would have no moves (as B and R have no moves in common), and the conjunctive compound of K and R would be the Wazir, etc. There doesn't seem a case where this cumbersome way of describing more elementary move sets is useful, as they tend to all have simple names of their own. Note that the Sissa can neither move as a Rook, nor as a Bishop.

It seems to me the addition of 'disjunctive' serves no other purpose here then sow confusion in a case that otherwise would be correctly understood with 100% certainty.

Thank you very much, HG, for the clarification. Soon I will make the correction in what corresponds to the Sac Chess preset. It will take more time to do it in other texts because they are many.

What is paradoxical and anecdotal about this case is that I believed that making that distinction introduced clarity!

In passing, it should be borne in mind that 90% of what I write in English is "formatted"  [written, thought, said] by the Google translator.


H. G. Muller wrote on Sun, Jun 28, 2020 12:50 PM UTC:

Sissa is a compound of rook and bishop but its move is not disjunctive, it is conjunctive.

Conjunctive = A and B

Disjunctive = A or B

Am I right?

I have never seen the word 'compound' used in sequential meaning; I would say the Sissa is an (isosceles) hook mover. And I would not say the Tai Shogi Hook Mover is a (conjunctive) compound of two Rooks. The word 'or' usually implies 'and', and if you consider this operations on the move sets, the 'conjunctive compound' of R and B would have no moves (as B and R have no moves in common), and the conjunctive compound of K and R would be the Wazir, etc. There doesn't seem a case where this cumbersome way of describing more elementary move sets is useful, as they tend to all have simple names of their own. Note that the Sissa can neither move as a Rook, nor as a Bishop.

It seems to me the addition of 'disjunctive' serves no other purpose here then sow confusion in a case that otherwise would be correctly understood with 100% certainty.


💡📝Kevin Pacey wrote on Sat, Jun 27, 2020 07:13 PM UTC:

@ Greg:

I don't mind, as long as Carlos is happy (the presets for Sac Chess were made by him). Just as long as previously finished or ongoing game logs of Sac Chess are not somehow broken when the changeover to a new set is made, hopefully.


Carlos Cetina wrote on Sat, Jun 27, 2020 07:06 PM UTC:

Thanks Greg. I would suggest that players could choose customize the piece set between abstract and alfaerie.

The one who edited the preset in use was me with some guidance from Fergus.


Greg Strong wrote on Sat, Jun 27, 2020 06:52 PM UTC:

I can definitely make a custom piece set for Sac so that the Centaur is properly represented.  Regarding H.G.'s question why the pieces have strange notations, I am guessing it is because this game does not have a custom piece set, so Kevin used whatever notation was associated with an appropriate graphic in the existing Game Courier piece set that came closest.

Since I'm making a new piece set anyway, we can improve the notations as well, but that's up to Kevin.


Carlos Cetina wrote on Sat, Jun 27, 2020 06:40 PM UTC:

Sissa is a compound of rook and bishop but its move is not disjunctive, it is conjunctive.

Conjunctive = A and B

Disjunctive = A or B

Am I right?


H. G. Muller wrote on Sat, Jun 27, 2020 04:57 PM UTC:

And why the strange piece IDs, that seem to be totally unrelated to the piece name? 'T' for Amazon, 'G' for Sailor??? And what the heck is a 'disjunctive compound'? Is there more than one type of compound?


Carlos Cetina wrote on Sat, Jun 27, 2020 03:27 PM UTC:

Hello Greg,

I would like to know if there is any way to access the set of pieces that ChessV uses for Sac Chess, that is, to which group does it belong and the name of the particular set?

I am looking to change in the current preset of Sac Chess the image of the zebra for the most appropriate of the knight/guard.


🕸Fergus Duniho wrote on Sun, Apr 5, 2020 01:35 AM UTC:Excellent ★★★★★

I can't believe this game hasn't been reviewed yet. This is the best game I've played that includes an Amazon. I normally leave the Amazon out of my games, because it has the power to force checkmate by itself, and that has the potential to wreak a game. However, that hasn't been a problem with this game. This game includes several other weaker compound pieces that help make it unsafe to move the Amazons out too early. To get to the point where you could use an Amazon to force checkmate against a King, you have to do lots of maneuvering of other pieces. Furthermore, the potential of the Amazon getting a bead on the King means that position is sometimes more important than material advantage. You can't count on winning just because you are ahead materially. If you find that you can't stop your opponent's Amazon, you may lose even if you're materially ahead. This makes the game more dynamic and exciting.


💡📝Kevin Pacey wrote on Sat, Aug 17, 2019 07:05 PM UTC:

Sac Chess still seems to be holding up well in practical testing on Game Courier, I'm happy to see.

A while back I toyed with the idea of making novel spinoff variant(s) that included the use of N+Guard+Bishop (technically NWB) compound pieces (I've called them Freemasons) and also what I've called Ships, that is N+Guard+Rook (technically NFR) compound pieces (using either an 8x10 or 12x12 board). I had, subsequent to inventing Sac Chess, had the feeling these two triple compound pieces perhaps ought to have been included in a Sac Chess-like game, somehow, as they perhaps logically complete the set of thematic compound pieces added to the standard FIDE piece types. On the other hand, there seems to be no precedent for the use of these triple compound piece types in previously fairly-widely played CVs, e.g. shogi includes the promoted B & R types, at least, while I recall Amazons had been fairly popular, in a variant of chess played in Russia long(?) ago - even centaurs have been employed (Courier-Spiel). So, I don't feel the urge to use the Freemason and Ship compounds as strongly as I once did, as I now see Sac Chess clearly uses the 'classic' compound pieces, as Carlos once put it.

A side issue arising from all this that occured to me is: what values should the Freemason and Ship piece types have? If we note that computer studies place Archbishops (let alone Chancellors) very close to a Q in value, and also place Amazon=Q+N in value (only), at least on 8x8, there doesn't seem much room to put Freemasons and Ships between Archbishops (or Chancellors) and Amazons on a numerical value scale, one might guess (at first thought anyway). I'm not sure what I'd speculate any future computer studies might (approximately) give for the values of Freemasons or Ships, but my own tentative valuation of an Amazon (as =Q+N+P) gives more room to fit Freemasons and Ships on a values scale (e.g. for armies on 8x8 or 10x10), at least. [edit: Fwiw, for 10x10 my own tentative estimates place Chancellor=N+R+P=3+5.5+1=9.5; Amazon=Q+N+P=10+3+1=14 and Ship=(N+F+P)+R+P=(3+0.75+1)+5.5+1=11.25, for example; if I only changed the value of an Amazon, to be equal to Q+N=10+3=13, then its value would seemingly be too close to that of a Ship's.]


💡📝Kevin Pacey wrote on Mon, Jun 19, 2017 05:37 AM UTC:

Hi Fergus

Yes, 3-fold repetition is a draw in Sac Chess. I had assumed (apparently incorrectly) that somehow by the form to enter a submission on CVP, and by some instances I looked at of other submitted variants that have been played, that a submitted variant has the same basic or auxilary rules as standard chess, unless otherwise stated by the inventor, at least if the variant seems clearly intended to be chess-like (Shogi & Chinese Chess variants for example, would be another story).

Thus, by another rule of standard chess, Sac Chess would also have agreed draws allowed (which Game Courier assumes for all variants), and also the 50 move rule that is used in chess (with some modern exceptions that make the rule to be for 100 moves instead according to FIDE, e.g. R+B+K vs. R+K, and many semi-specified positions with 2N+K vs.P+K, the latter of which, however, clearly couldn't be applied to Sac Chess' 10x10 board). I was actually wondering recently if the 50 move rule should be applied to larger board variants than 8x8, since even B+N+K vs. K might take more than 50 moves in the worse case with good play on 10x10, though I doubt it will in that particular basic endgame, or others that might arise in Sac Chess.

In general it's a bit tedious for inventors to try to cope with or unfailingly anticipate every little rule case that might come up. A solution could be simply to state in one of one's CVP submissions that the basic rules of chess apply unless otherwise stated, but that might have caused some confusion in the case of Sac Chess, as you can see from the above rule cases, if I didn't still take the trouble to elaborate at length on them as I have above. I'd say if any variant ever becomes popular enough, any future governing body it would have (like FIDE is to chess) might later fine-tune the rules or clarify rule omissions/ambiguities made by the inventor.

At least on Game Courier, the players afaik seem to sort out such rule issues without much trouble, or perhaps even consult the inventor as here. I would observe that some variant submissions, even of old variants that might even be played on Game Courier, have an author's comment that certain rules aren't made clear by the inventor, and in such cases iany rule enforcing preset programmers can decide on the exact rules they enforce in their presets. I don't know if Carlos took into account any of the rule cases (3 fold draw, 50 move rule) that I discussed at length above as they apply to the Sac Chess rule-enforcing preset that you're currently using, but if you care to you might assume for now he did not.

 

On a personal note, I'd intended to stop playing on Game Courier for at least the summer (if not longer) due to, not least, having no air conditioner (just a fan) in the room where my 'playstation' is in my small apartment, but I'm now looking into the chance of having an affordable and practical portable air conditioner.

Kevin


🕸Fergus Duniho wrote on Mon, Jun 19, 2017 01:40 AM UTC:

Is 3-times repetition a draw? That's what I'm presently going for in a game I'm playing.


💡📝Kevin Pacey wrote on Sat, Apr 29, 2017 05:22 PM UTC:

Here's a Sac Chess game I played vs. Carlos arguably not so long ago. It seems like the closest Sac Chess game played between people that I've seen to date. Perhaps it's a matter of taste, but I like that of the relatively few pieces left on the board at the end, all of the 4 Amazons still remain:

/play/pbm/play.php?game=Sac%2520Chess&log=panther-sissa-2016-289-852


💡📝Kevin Pacey wrote on Mon, Mar 27, 2017 07:30 PM UTC:

After some playtesting of Sac Chess, I am more confident it is playable, and may be a very rich game indeed, given it has so many pieces. What pleases me a lot is that while in chess an average game is 40 moves, with a pair of units traded on average every 5 moves (thus leaving an average of 8 pieces per side at the end of a game), in Sac Chess an average game so far on Game Courier seems to last about 60-80 moves, though at times being played out until mate (if done for chess, an average game of it might last about as long, too). It seems the rate at which a pair of units is traded in Sac Chess is about every 3.5 to 4 moves, which would also put an average game at around 80 moves, if such a game got down to 8 pieces per side (though it doesn't seem to get this far in practice, so far on Game Courier). Another nice thing about Sac Chess is that it seems to allow for more decisive games than is the case in chess, if played at a high level, I'd guess.

One thing I'm not sure about is how rich the early opening phase, or range of 'good' opening variations, can be for Sac Chess. Playing something like the Pirc Defence in chess seems less attractive here, since White's Queen is guarded by an Amazon in the start position, so White might early on hit Black's knight on the kingside with the appropriate centre pawn. Also, Open Sicilians in chess may be a little, or a lot, less attractive for White in Sac Chess. Then there are the ramifications of the Queen's Gambit Accepted in chess still waiting to be analyzed, if played in Sac Chess, and this could be really critical to the range of attractive/promising White openings. On the other hand, 'insipid' sidelines in chess may be much more interesting in Sac Chess (in regard to the early opening moves, I mean as always here). Even if White's initiative is less than in chess, this can be seen as a good thing, perhaps.

Regarding the names I chose for some of the pieces, I still like 'Missionary' and 'Sailor' as used (these seem even a bit appropriate, given what a man is crossed with, e.g. a bishop in the former case). However, I have some regret about not using 'Centaur' instead of 'Judge' (I somehow missed the Centaur entry on wikipedia, in spite of looking at its long list of fairy chess pieces several times for just such a precedent). On the other hand, maybe there's something nicer about using the name of a human occupation (Judge), rather than the name of a mythical monster; at least there were travelling judges in medieval times. Plus, I'd note that I don't completely get the consistency of why, e.g., an Amazon is called what it is, since no horse is involved in that name, and given that such a female warrior doesn't necessarily ride a horse (not only that, but a knight might be seen as a man riding a horse). But I digress.


💡📝Kevin Pacey wrote on Thu, Jul 14, 2016 06:45 PM UTC:

Carlos wrote some months ago: "I have just played a game versus the HG's Fairy-Max/Winboard/Sac Chess program..."

I have a question for HG: I wasn't aware till I searched recently that Fairy-Max has its own page on the web. Is Sac Chess one of the variant programs under Fairy-Max that any viewer/user can find available even nowadays? Maybe I didn't look hard enough, but I didn't see Sac Chess listed as such. Thanks in advance, Kevin.

P.S. I saw after making this post that you mentioned your Fairy-Max program for Sac Chess on Chess.com 4 months ago. Thanks for that, too!

💡📝Kevin Pacey wrote on Wed, Feb 3, 2016 11:25 PM UTC:
I've adjusted my tentative Sac Chess relative piece value estimates, in just a few cases, after realizing that I had overestimated the fighting value of a Sac Chess K.

Carlos Cetina wrote on Sun, Jan 31, 2016 10:31 PM UTC:
I have just played a game versus the HG's Fairy-Max/Winboard/Sac Chess program and translated it to a Game Courier log. Watch it here. Everyone draw their own conclusions.

[Play Tester played the role of the program.]


💡📝Kevin Pacey wrote on Sun, Jan 31, 2016 05:55 AM UTC:
I thought that tonight I better try unzipping that link you provided, H.G., in case I needed my sister-in-law's help tomorrow. It was pretty much as simple to unzip as you described in an earlier post. The only thing was I couldn't seem to save it as anything but a temporary download, so I'll have to download it each time I use it, unless I can figure out how not to. My used laptop was a gift from a friend, which might somehow complicate things due to any possible security measures set for him.

The program was set for 1 minute per side, and I initially tried playing at that speed. I clearly blundered a pawn for nothing very early, but for some reason the engine passed it up. Some more blunders followed (by me) and I soon allowed a mate. The next three games I tried 5 minutes, then 10, then 15. At one point I noticed I had lost on time, but was allowed to keep playing anyway (I didn't bother to figure out how to change that yet, if I wish to). Each of these 3 games I lost as well. Maybe I'm not yet used to the images for all the piece types, but the odd blunder (or once a mouse slip) kept occurring. I came closest to doing well in the 15 minute game (I always tried not to 'lose on time').

After that I set the program to have two engines play a 15 minute game. Again they each exceeded the limit and were awarded a further 15 minutes each on the clock all the same. As a spectator I thought the engines played conservatively, if not anti-positionally (IMHO). At one point Black seemed to go ahead in material considerably after about 40 moves with one trade only. Then trades came thick and fast, and eventually Black won a pawn ending two clean pawns ahead. The game only took about 80 moves (until mate), which would be about right for standard chess.

Thanks again for the link, H.G.

💡📝Kevin Pacey wrote on Sun, Jan 31, 2016 01:17 AM UTC:
Sorry H.G., I hadn't been in a hurry up till now, and didn't realize you were eager. I wanted to at least see how my 2 games with Carlos were going to develop before getting more adventurous by playing a computer. Looking at our game of Sac Chess, Carlos seems to be doing very well, which doesn't suggest my chances would be good against a program (against a good one, I half expected to score say 1/4, if it would be about 200 rating points higher than my standard chess rating). 

In the meantime, it is becoming more clear to me that looking for a computer-resistant chess variant is a futile exercise, especially now that 19x19 Go computer programs are demonstrably so much stronger than before. I had guessed that it might take up to 100 more years for Go programs to take over and eventually dominate that game, from people.

I also had wanted to try to see my Glinski's Hexagonal Chess game with Carlos to a conclusion, before trying out that link you gave. An irrational concern that I might somehow go wrong and mess up my laptop in some way, before that game had finished, had been with me. Anyway, I'll try to get playing that computer program at Sac Chess soon after my brother & his wife go home tomorrow.

H. G. Muller wrote on Sat, Jan 30, 2016 09:53 PM UTC:
Did you already have an opportunity to try the program at <a href="http://hgm.nubati.net/SacChess.zip">http://hgm.nubati.net/SacChess.zip</a> ?

💡📝Kevin Pacey wrote on Sat, Jan 30, 2016 08:53 AM UTC:
I've added the following to the end of the Notes section for my Sac Chess submission (to try to, in the most natural way I can imagine to date, accommodate those who dislike there being a single queen and two amazons in the setup position for Sac Chess):

"Note: to accommodate those who dislike the queen being clearly inferior to amazons both in their power AND in their number in Sac Chess at the start of a game, I can suggest the following fairly natural idea for a modified variant ('Royal Bevy Chess'), i.e. it has a similar setup and the same rules as for Sac Chess. Namely, in the setup position for Sac Chess, switch 2 queens for the 2 amazons, and switch an amazon for the single queen. This idea of having two queens and one amazon in a setup position may have first been used in 'Alekhine Chess', which is somewhat similar to Sac Chess, perhaps; here's a link to it:"

http://www.chessvariants.com/index/external.php?itemid=zAlekhineChess

💡📝Kevin Pacey wrote on Tue, Jan 26, 2016 02:31 AM UTC:

I originally had in mind that the person I seem to recall who disagreed with the game produced by the inventor of Alekhine Chess (as far as the number of its Amazons to be included) was an example of a person not pleased with a particular game (other than mine, which is somewhat similar to Alekhine Chess). Clearly not everyone can always be satisfied with any particular variant, and someone may want it altered even in an unspecified way. In Sac Chess, if one has a position on the board with all piece types included, such as in the Setup position, the Amazon is the only piece that can always (unless pinned to the King) take any sort of piece type that is threatening it (I think this is a major point that you're indicating). The game is also a bit heavy on pieces that can move at the least like both a king and a rook (or bishop), e.g. both Sailors, both Amazons, and the Queen (for each side) can move like a king or rook, at a minimum. It's a drawback (of the theme of compound pieces that I used), possibly.

Whether or not it would be better to alter the game is an open question, but at this stage it's already being played on Game Courier, and I think I like the way it's turned out so far. Perhaps I (or someone else) can make a variant based on Sac Chess at a later date that may prove popular. At the moment it's the only variant I've invented (out of 10, so far) that someone (namely Carlos) kindly decided to write a Game Courier program for.

[edit 28 March 2018: I'd note, too, that, e.g., in the historic variant Courier Chess, the queen used there is a ferz-like piece, clearly inferior in powers to the guard-like piece type that is also used in that game, so it seems that Sac Chess is in such a way at least not without one or more precedent(s), in variant(s) that proved at the least somewhat popular in the past, if only regionally perhaps.]


John Whelan wrote on Tue, Jan 26, 2016 01:53 AM UTC:
Re:  "you can't please everyone" - No, but it is not as though I ever demanded you do 1 Amazon & 2 queens.  It was 1 of several alternative suggestions.  And these were merely examples to suggest that the issue is not insoluble.

Re: "close to standard chess in spirit".  Seems to me the pieces of FIDE chess have a high degree of ability to threaten each other without being threatened back.  Even the Queen, the most powerful piece on the board, can be threatened by the 2 knights.  Seems to me a lot of that is lacking here, for all the variety of pieces.  If an opposing piece cannot threaten you back, it is probably less powerful.

💡📝Kevin Pacey wrote on Tue, Jan 26, 2016 01:18 AM UTC:
Fwiw, here's a link to Alekhine Chess, which uses 1 Amazon and 2 Queens: http://www.chessvariants.com/index/external.php?itemid=zAlekhineChess Also fwiw, I seem to recall somewhere one person commented that that game could be better with 2 Amazons and 1 Queen, so as to lessen the chance that all the Amazons might be traded off... you can't please everybody. I wanted to keep Sac Chess fairly close to standard chess in spirit, in a way I hope that's understandable, when adding all the compound pieces to fill in the spaces in each side's camp, on a 10x10 board. Including Amazons in a variant always seems to create some sort of a backlash eventually, but I couldn't resist doing so since they seemed a natural extension of the compounds that I used (which are all the crowned or knighted standard chess pieces).

John Whelan wrote on Tue, Jan 26, 2016 12:17 AM UTC:
One solution would be to have two Queens, and to put the Amazon in the place of honor next to the King (and perhaps change their names).  The only thing stopping this would be a desire to keep both the traditional names, and their traditional significance.

Another solution might be to give the King some kind of enhanced movement power while the Queen is in play, such as an ability to switch places with the Queen, or an ability to jump over a Queen.

Another solution might be to give all the knight-combo pieces, including the Amazon, limited range:   2 or 3 squares (distinguishing them from the King-combo pieces). This would leaving the Queen, Bishop and Rook as the only full-range pieces.  This would leave most of the pieces geographically localized, which fits the advice of some who have discussed computer resistant variants.

There are any number of possible solutions.

💡📝Kevin Pacey wrote on Sat, Jan 23, 2016 04:01 AM UTC:
In my submission I gave links from my Chess Federation of Canada website blog, including my entry there that covered Sac Chess in detail. Included in that entry was my proposal for two slight spinoff variants of Sac Chess to accomodate anyone who dislikes Amazons being clearly more powerful than a Queen. One spinoff variant ("Cash Chess") had two different piece compound pieces replacing each of the 2 Amazons in both camps. These were to be B+N+K ("Freemason") and R+N+K ("Ship") compounds, each a compound piece of my own invention (afaik). 

I wasn't keen on this spinoff variant myself, since there would be less near-symmetry in each camp, and there would be more piece types to remember in total. The other spinoff variant ("Royal Sac Chess") would let Queens have a small distinction over Amazons on at least some occasions. In this spinoff variant, in the event of stalemate or a three-fold repetition of moves (perpetual check being a case of such), if one side had more queens on the board than his opponent then he would win the game, else it would be a draw in the event of an equal number of queens.

I kind of liked this spinoff variant, as perhaps an important way to justify having Queens on the board alongside Amazons (which are clearly superior in their relative value AND movement capabilities on an empty board). On the other hand, it's using a rather artificial rule change that would drastically affect some games' final results. 

In the end, I decided that the reasons you gave for liking having a Queen, besides Amazons, were adequate enough, such as keeping a quaint royal tradition going in spite of my wanting to include many compound pieces in Sac Chess, including Amazons. The Amazon seems to get a bad rap because it is so powerful (it can mate a King on an empty board with no assistance, for instance). At least having a Queen on the board as a possible threat to trade itself for an Amazon gives a player with an Amazon one more thing to watch out for.

John Whelan wrote on Sat, Jan 23, 2016 02:28 AM UTC:
I am interested this variant because I am fascinated by large chess variants, and this variant is LARGE (Dragonchess beats it though).   One thing bothers me, though.  The Queen.  She retains her status as the only unique piece other than the King.  She retains her place of honor by the King's side.  But there is nothing special about her in this version.  She is trumped to two Amazons, and a number of other pieces come close to her in value.

H. G. Muller wrote on Mon, Dec 21, 2015 09:46 AM UTC:
> <i>https://en.wikipedia.org/wiki/Princess_(chess)#Value</i> <p> And guess who wrote that there! :-) <p> Note that arguments based on move counting on an empty board are often unreliable. Distant slider moves contribute very little to value, because during most of the game you will not be able to make them anyway due to blocking or board edges. And once the board population has dropped to where you can, they often lead to places where you have no business going anyway. In defending KRPKR it hardly matters how far away you can move the Rook from where the Pawn is. <p> The value of a piece is mostly determined by how efficiently a piece supports and combats Pawns in the end-game. <p> Another factor could be 'winning power' in often occurring Pawnless end-games, which determine how easily you can draw by acrificing a piece for the opponent's last Pawn to stop its promotion. E.g. Queen + minor (B or N) vs Queen is draw, making Q+minor+Pawn vs Q+minor also easy to draw. But Q+minor vs Chanchellor or Archbishop is a win. Q vs R is virtually always won, giving Q+P vs 2R a big advantage, but C vs R is only won if your King is not confined to its own board half by the Rook, and A vs R is a dead draw, making A+P vs 2R a hopeless proposition. More or less coincidental properties like that might have impact on the value, although it is different to separate cause from consequence here:one could also argue that some pieces win and others not against a given opposition is because they were generally inferior. That it is not possible to checkmate a bare King with a pair of Knights, but you can do it with a pair of Phoenixes, does not seem to be tracable to general inferiority of the Knight, however. (E.g. KRKN is generally a draw, but Rook vs Phoenix is a general win.)

💡📝Kevin Pacey wrote on Sun, Dec 20, 2015 11:48 PM UTC:
Thought I'd give the wikipedia link below, discussing the value of the Archbishop (Princess) fairy chess piece:

https://en.wikipedia.org/wiki/Princess_(chess)#Value

I'd note wikipedia mentions that computer self-play studies were used to establish a sort of value for this piece that Muller gave (though he nor anyone else seems to be credited for the computer studies, as far as I've read). Note also that links concerning the Chancellor (Empress) and Amazon are given by wikipedia.

Regardless of arguments about what the Archbishop's value might be in a particular chess variant, I'd note that wikipedia mentions that the computer studies led to the conclusion that the Archbishop increased its relative value (in comparison to that of a queen) on an 8x10 board, as opposed to the smaller 8x8 board. Without bringing in any notion of synergy necessarily, I can give one possible reason (or contributing reason) why that conclusion might be correct, in case no one has mentioned it. 

Namely, whereas on an 8x10 board the rook component of a queen benefits the queen as a piece by 2 extra squares covered (on an empty board) at all times, not depending at all on the location of the square, note that the knight component of the Archbishop has many more available squares to it (than on an 8x8 board) where it benefits the Archbishop as a piece, by up to 4 extra squares covered more than from a less favourable square nearby that a knight might have to settle for on the smaller 8x8 board.

I can note that on a 10x10 board (such as for Sac Chess), the rook component for a queen would by similar reasoning pick up 4 extra squares covered on an empty board (than if on an 8x8 board) whereas the knight component for an Archbishop often allows it to pick up up to (still) just 4 extra squares coverage for the Archbishop (than if on an 8x8 board), suggesting to me that a 10x10 board might not benefit an Archbishop like an 8x10 board apparently did, in terms of its value to that of a Queen. Perhaps inconclusive and murky pondering on my part, I'll admit, but it gave me pause.

💡📝Kevin Pacey wrote on Sat, Dec 19, 2015 10:56 PM UTC:
H.G. wrote:

"BTW, did you just place the text string the generator gave you in a non-HTML submission, or did you have to tick 'Using HTML tags' to make it accept the diagram?"

Being tech-challenged these days, my attempt to describe what I did would be:

I went to edit my Sac Chess submission, then in doing so noticed that just above the 'Setup' box (i.e. to show the starting position for a chess variant) there now exists a sentence that mentions the Diagram Designer. I then clicked (on its blue highlighted words) "Diagram Designer" and a new window opened up, i.e. that of the Diagram Designer. Perhaps it wasn't quite obvious to a newbie how to use it even in a minimal way, but its board shape setting happened to be for a "Square" shape (since the Diagram Designer by default was for standard chess). I tried setting "Columns" = 10 and the square board increased from 8x8 to 10x10, still showing the 32 chess pieces. 

I happened to decide to click on some hignlighted words for a setting ("Next Rank", I think), and revealed to me by a new window was a long documentation. This was actually useful enough to me in a way (with my minimal desire, i.e. to at least initially find any possible quick & dirty way to make a better diagram for my Sac Chess submission), because it finally dawned on me that the 26 characters of the alphabet each represent an available chess or fairy piece type in the given set (abstract pieces), and a character is to be used as part of the FEN string. After closing this window, I saw that the FEN string default was indeed for chess, and I could quickly tell how to use the FEN string for Sac Chess pieces (and its empty squares) instead. The slight difficulty after that was deciding which letters of the alphabet corresponded to desired Sac Chess pieces, though I eventually decided to use the standard chess piece abstract figures as a sort of theme I carried over when I decided on the letters of the alphabet to represent other (fairy chess) pieces in Sac Chess in the FEN string. 

At any time I could click on an "Update" button and the Diagram Designer would show the latest version of the board, with the pieces on it. When I had finished, I knew there was a box in the Diagram Designer that contains a brief code (it described it as HTML code), and the final code corresponded to the work I had done with the Diagram Designer. The instructions above this box said to simply copy the contents of the box to my Chess Variants webpage, so I copied the contents as the first step of a quick cut & paste. I closed the Diagram Designer window, i.e. returning to my submission webpage that I was editing. I then did the pasting of the Diagram Designer code into the "Setup" box of my submission webpage, which simply showed the text of the code as a result (I had before now deleted my old version of a Setup diagram that wasn't up to snuff). After I submitted this edited version of my Sac Chess submission, now whenever I look at the new version of it, the Diagram Designer diagram I specified actually shows up on my screen.

Edit: I've identified the squares pieces are on (by text) in my Sac Chess setup diagram now.

🕸Fergus Duniho wrote on Sat, Dec 19, 2015 01:41 PM UTC:
Abstract is a good choice for the piece set, though I'm biased. I will mention that three of the pieces you chose (besides the King) were designed as royal pieces for games with alternate royal pieces (mainly Fusion Chess and Cavalier Chess). The one that could most easily be confused for a royal piece is the one combining the King and Knight images. There is a nonroyal alternative that combines the Man and the Knight. This is called ct in the Abstract: All set. I'll add some non-royal alternatives for the others later.

Another thing that is important is to make a connection between the diagram and your descriptions. Including an image of each piece in the piece descriptions is a good way to do this. An alternate way to do this is to mention which spaces each piece starts on.

H. G. Muller wrote on Sat, Dec 19, 2015 12:33 PM UTC:
> <i>Please confirm I've used all of the right types of figures to represent all of the pieces and pawns in the diagram, as I was slightly uncertain what some of the figures represented</i> <p> It seems you did. But, because of the uncertainty you express here yourself, do you really think this is the best image set to represent Sac Chess? To me the Alfaerie images are a lot more recognizable. Under 'Compounds' the Diagram Generator seems to have symbols for all the pieces in your game, including the BK and RK compounds. <p> One problem now is that the part where you explain how the pieces move you refer to them by ID letter, but the IDs are no longer part of the diagram. Perhaps you should prefix those explanation lines by a list of squares on which these pieces start, so people can make the connection between image and name. <p> BTW, did you just place the text string the generator gave you in a non-HTML submission, or did you have to tick 'Using HTML tags' to make it accept the diagram?

H. G. Muller wrote on Sat, Dec 19, 2015 08:05 AM UTC:
> <i>For example, if one incorrectly sets the value of a rook (or, I would opine for argument's sake, even a bishop) exactly equal to the value of a knight, I'd imagine in a number of playtest games the side with an extra rook would erroneously trade it for the extra knight of the opposing side, say when thinking the position was approximately equal in all respects.</i> <p> Indeed, this is exactly what happens. The Rook side will needlessly squander a Rook for a Knight, and because the initial setup was likely to give the already two Pawns worth of compensation, would badly lose after that. So making the programs erroneously believe that a certain trade is equal is one of the major pitfals of this method. This especially holds for 1-on-1 trades, as opportunities for concerted multiple trading do not occur very frequently. The worst of those is if two values differ exactly a Pawn, as Pawns are abundant, and X + Pawn for Y opportunities are not so rare as the others. <p> But you know which values have been given to the pieces in the opponent armies, so you know which of those are close to the initial estimate of the value or one Pawn above/below it. And then I usually try to stay ~20cP away from those points. If needed you can make two test runs, one with a programmed value 20cP above what you think it should be, one 20cP below it. If the results are nearly equal there is no reason to distrust it. If they are different the one that used the programmed value closest to what the score outcome suggests will obviously be the more reliable, and if the programmed value is different from the score-outcome value, you repeat the test with the latter. If the value suggested by the score is very close to that of a piece in the opponent value, you can repeat the test against armies of another composition, lacking the offending piece. It is always better to base the value assignment not on a single imbalance, but on a variety of imbalances anyway. (E.g. not only play A vs Q and A+P vs Q, but also A+P vs 2R, A+P vs 2N+B and A vs R+N+P.) <p> > <i>Regarding when a preliminary value is assigned to an Archbishop that is at least slightly different than that of a Queen when pitting the two pieces against each other in playtesting (other material being equal at the start), ...</i> <p> Well, obviously predicting the desirability of some trades the wrong way around will lead to unnatural play, which might affect the statistics of other trading opportunities compared to 'natural' games, which then affects the outcome. This is all possible in theory. In practice, however, you will be able to see that when it happens. This is why pondering about it is not the same as actually doing it. You can repeat the test with all kind of different programmed values for A, and see how the result score varies by this. If it doesn't vary at all, <i>apparently</i> the problem does not occur in practice. And in the worst case the value implied by the scores does significantly depend in some way on the programmed value, and you have to search for self-consistency, i.e. the value you have to program to get the same value out of the score. <p> With the usual resolution I am aiming for (~0.2 Pawn), however, I have never seen that happen, though. Trying to get more precise values is probably meaningless, as you will start to resolve all the higer-order corrections to the model of additive piece values. E.g. some pieces might do better against Knights, other might do better against Bishops.

💡📝Kevin Pacey wrote on Sat, Dec 19, 2015 07:45 AM UTC:
I've changed the setup position diagram for my Sac Chess submission, using the Diagram Designer as you requested, Fergus. Please confirm I've used all of the right types of figures to represent all of the pieces and pawns in the diagram, as I was slightly uncertain what some of the figures represented (some of the 26 possibilities to chose from were mysteries to me, but I thought I found all of the ones applicable to Sac Chess). 

Eventually I'll redo diagrams in similar fashion where needed for all of my Chess Variants Pages submissions if I am able to figure out how to.

💡📝Kevin Pacey wrote on Sat, Dec 19, 2015 05:35 AM UTC:
H.G. wrote:

"...One would expect the playtest to be only meaningful when the values used are consistent, i.e. the programmed value used for deciding on trades are the same as the value that comes out based on the score percentage. But to my surprise, putting a moderately wrong value there hardly had any effect on the outcome at all. If I put Q=9.5 and C=9, and play an army with Q against an army with C, the Queen wins by ~58%. If I put Q=9.5, and C=10, the Queen still wins by 58%! The explanation is that both engines share the misconception. So one of the two sides will always try to avoid the trade, meaning that Q for C trades will be relatively rare. So the test mainly measures how much damage Q and C do to the other pieces. Although a wrong C value might lead to wrong 2-for-1 or 3-for-1 trades, the number of occasions where such a trade can be forced is relatively rare, especially if they are not exactly equal, so that one of the players will try to avoid them. So the most error-prone value assignment is actually the one where the value is exactly the same as that of another piece, or the sum of two other pieces (and wrongly so). So I usually avoid that.
..."


Fwiw, instances of assigning a preliminary value for a piece to the same value of that of another piece (and also being a wrong preliminary value) were in fact uppermost in my mind. For example, if one incorrectly sets the value of a rook (or, I would opine for argument's sake, even a bishop) exactly equal to the value of a knight, I'd imagine in a number of playtest games the side with an extra rook would erroneously trade it for the extra knight of the opposing side, say when thinking the position was approximately equal in all respects. If the number of such games is significantly large in the playtesting, this could seriously drive up the percentage of drawn games (let alone losses) in such games where a rook for knight advantage is erroneously thrown away through such a trade, substantially skewing the results of the playtesting.

Regarding when a preliminary value is assigned to an Archbishop that is at least slightly different than that of a Queen when pitting the two pieces against each other in playtesting (other material being equal at the start), for example, and the value assigned the Archbishop is at least slightly wrong, I am now wondering something similar to what was uppermost on my mind before. That is, if the effect of all resulting incorrectly avoided trades during playtesting (e.g. of Queen for Archbishop plus a certain number of pawn[s]) might be to at least drive up the number of resulting unnecessary draws (let alone losses), in a way that may not at a minimum favour the Queen even approximately appropriately as far as its final overall percentage score in playtesting when pitted vs. an Archbishop.

In short, I wonder if the results of such playtest games might even still be substantially skewed (setting aside the quality of the play by the engines). At the risk of stating the obvious, viewers can note that even if the Queen wins about the correct ratio of times vs. its losses, an incorrect (say too high) percentage of drawn games skews the overall results percentages if measuring the Archbishop. For example, if in 20 games the Queen wins 8 times, loses 4 times, with 8 draws, for an overall percentage of 60%, it has the same ratio of wins to losses if it wins 10 times and loses 5 times (i.e. with 5 draws), but in the latter case the Queen scores a better overall percentage (of 62.5%). The actual difference due to any playtesting that might be faulty might conceivably be quite greater percentagewise than for these example figures.

H. G. Muller wrote on Fri, Dec 18, 2015 09:08 PM UTC:
> <i>Another question I have is: assuming that the computer programs you used for the playtesting must have assigned a preliminary value to an Archbishop, if only for when deciding during calculations whether to exchange it for something else (e.g. a queen), what was that preliminary value?</i> <p> Your questions are very much to the point. One would expect the playtest to be only meaningful when the values used are consistent, i.e. the programmed value used for deciding on trades are the same as the value that comes out based on the score percentage. But to my surprise, putting a moderately wrong value there hardly had any effect on the outcome at all. If I put Q=9.5 and C=9, and play an army with Q against an army with C, the Queen wins by ~58%. If I put Q=9.5, and C=10, the Queen still wins by 58%! The explanation is that both engines share the misconception. So one of the two sides will always try to avoid the trade, meaning that Q for C trades will be relatively rare. So the test mainly measures how much damage Q and C do to the other pieces. Although a wrong C value might lead to wrong 2-for-1 or 3-for-1 trades, the number of occasions where such a trade can be forced is relatively rare, especially if they are not exactly equal, so that one of the players will try to avoid them. So the most error-prone value assignment is actually the one where the value is exactly the same as that of another piece, or the sum of two other pieces (and wrongly so). So I usually avoid that. <p> One must not be off too much, though: If you make A < B + N the B + N side will try to avoid trading 2 minors for A, but with 4 minors of him on the board and a piece as powerful as A, he will not be able to do that, and in many games the side with A will squander his initial advantage by striving for the wrong trade. So eventually I always do the tests with the piece value in the program set to nearly the same value as what eventualy comes out. That the outcome is so insensitive to that value just means that I get the consistent value much quicker, usually already on the second try. <p> One has to be quite careful with pieces of nearly the same value, however. E.g. Knights and (unpaired) Bishops on 8x8. If I set N=3 and B=3.5, the side with the Knights will moderately win. But if I set N=3.5 and B=3 for letting the program decide what to trade, the side with the Bishops will moderately win! The piece that you tell the program is stronger will in practice be weaker. The explanation is that the program with the piece with the highest <i>programmed</i> value will try to protect it from trading against the perceived weaker piece. So if it can avert a small positional loss (of, say, 30cP) only by avoiding the trade, it will prefer the small loss. But that happens time after time during the game. The unwillingness to trade hinders the piece that is imagined stronger to do its job well, and therefore makes the piece weaker. This draws the values that have to be programmed (or used by humans to determine which trades are favorable) very close together. You might want to protect a Bishop against trading for a Knight, but not at all cost, and not even at a cost that is a large fraction of the intrinsic value difference, as you likely will have to do it several times during a game. This is an example of where piece values are influenced by the composition of the opponent army, that goes beyond simple addition of values for individual pieces. <p> > <i>Decades ago I was somewhat into computer programming, and found a lot of time was involved, and eventually a kind of aversion developed. </i> <p> Well, nowadays using software is as far removed from programming as watching television is from building one. And it can be quite interesting to see on the screen how the programs battle each other in a variant, when you set them to a time control that is a good compromise between being boring and driving you crazy. At least, I enjoy it a lot. Often it is much more interesting than watching a soccer game! :-)

💡📝Kevin Pacey wrote on Fri, Dec 18, 2015 07:02 PM UTC:
Hi H.G.

Another question I have is: assuming that the computer programs you used for the playtesting must have assigned a preliminary value to an Archbishop, if only for when deciding during calculations whether to exchange it for something else (e.g. a queen), what was that preliminary value? If it was set to the same value exactly as a queen, that would affect decisions such an engine made at times during a game via its calculations, for any moments when say an exchange of an Archbishop for a Queen might have been possible. I'm wondering if that might affect the play and results of a program during playtesting as much as if it, say, valued a minor piece the same as a rook during its calculations for any available exchanging posssibilities.

In regard to playtesting Sac Chess myself, I'm still just reluctantly dipping my toes into the edge of the water as far as my experience with engines and modern computer software goes. Decades ago I was somewhat into computer programming, and found a lot of time was involved, and eventually a kind of aversion developed. I can't afford to be a complete luddite though when it comes to refusing to use computers at all these days, amd if I immerse myself fully into the world of chess variants I may try my hand at playtesting with engines as you suggest. However, I think I'd prefer to start with human vs. human playtesting of Sac Chess, in the New Year if not sooner.

H. G. Muller wrote on Fri, Dec 18, 2015 12:00 PM UTC:
> <i>I'm still wondering about the value of an Amazon asserted to be only Q + N, in your opinion (besides that of other people).</i> <p> Well, it was not so much an opinion as an observation. The disclaimer is that I did not play-test the Amazon nearly as extensive as I did for the Capablanca pieces. The only test I did was replace Q by Z and delete one of the Knights in compensation from the FIDE setup, and play some 1000 games. To my surprise the score was very close to 50%; I had expected there to be some synergy. <p> A possible explanation could be that the manoeuvrability at some point saturates when the piece gets too powerful. Because the set of squares that the piece can reach in a single move is already so large that what it can reach in two moves is mostly encompassed in it. The two-step N tours (N+N) are weighted into the value of the Knight, and the two-step Q tours in that of the Queen. The Amazon would have in addition two-step Q+N and N+Q tours. But it is doubtful whether this really adds much to what two-step Q+Q tours already do. (To show that quantitatively one would obviously have to make assumptions on a representative filling fraction of the board, as a Queen, or even a Rook on an empty board would already cover the entire board with its two-step tours. So simple arm-chair math won't cut it. But the play testing of course would exactly measure that.) <p> > <i>... still ought to be measured from scratch by testplaying using Sac Chess games, if anyone is willing.</i> You are absolutely right about that. Board size does have an effect, and in particular I found that diagonal sliders like B gain in value compared to orthogonal ones on wider board. The R-B difference decreases in going from FIDE to 10x8 Capablanca to Cylinder Chess. This because it gets more and more common that both forward moves attack the enemy position, rather than bumping into the board edge. The values I quoted were for a 10x8 board, though, which should be very similar to your 10x10 board with Pawns starting on 3rd rank. An extra rank behind the pieces were you virtually never can or want to go doesn't have much effect. <p> But why don't you do it yourself? You seem to have a computer, as you post here (and I cannot imagine you would type such long messages from a phone...). So you could just download WinBoard and Sjaak II, set it to play Sac Chess, specify a list with materially imbalanced start positions featuring Amazons versus Queens+Knights or Chancellors+Bishops, let it run overnight and see what they did. <p> > <i>Whether just hundreds of games is a statistically satisfactory playtest sample size, I am not sure</i> <p> With a draw rate of 32% typically for orthodox Chess in these tests the statistical error in 100 games equals 40%/sqrt(100) = 4%. So I do at minimum 400 games, for an error of 2%. If a Pawn is worth 16% score advantage, that corresponds to 1/8 of a Pawn, and the 95% confidence interval of the result, which is 2 standard deviations, to 1/4 of a Pawn. Usually I play 400 games with many different opposing piece combinations each, though. The original determination of the piece values in Capablanca Chess for the benefit of optimizing the Joker80 engine was done with 20,000 games (but that was for all pieces together, not just Archbishop). And indeed I average over the first-move advantage. <p> As to the strength of the computers, I found the obvious ways to vary that to have almost no effect, over a range of about 600 Elo. I should add that the programs play at a level where they would crush most humans, although probably not at GM level at these fast time controls. The score advantage to which a certain material advantage corresponds can vary with the level of play. But trying to convert an advantage of 100cP into a win, be it a plain Pawn or Q vs R+B, appparently requires approximately the same level of skill. So that the score excess drops similarly when you decrease the skill very much and play gets less accurate, meaning that in terms of Pawn values the result would still be the same. So your concerns are valid, but they can be (and have been) checked. <p> Note that if it would be really true that the relation between material imbalance and score would be dependent on quality of play, it would imply that the piece values are not 'constants of nature', but would be different for players of different skill. It makes no sense to value a Bishop pair more than a Knight pair when your skill with Bishops is so poor that you would always lose with them against two Knights, but would be able to draw with two Knights against two Knights. Yet no one suggested this would be true for the values of the orthodox pieces. <p> > <i>Larry Kaufmam is an International Master as far as chess goes...</i> <p> I thought he was a GM, although it seems that he deserved this title by winning the U.S. Championship for players over age 60, so I don't know how much that is worth. But that is not really relevant. What matters in this case is only that he knows how to count. He obtained the piece value by identifying material imbalances in a database of GM games (reamining in effect long enough to exclude transient tactics), and determining the win rate in which these resulted. The only thing that matters is the strength of the GMs that actually played those games, not that of the person observing them and counting their number of wins... <p> > <i>In any event, I have not heard of any reasonably strong human chess players changing their strategies in regard to trading bishops for knights in over the board play,</i> <p> Well, perhaps this is one of the reasons why computers are about 1000 Elo stronger than humans. Note that Larry Kaufman was responsible for the evaluation function in both Rybka and Komodo, which were the strongest engines in existence at the time he was involved in them. <p> Dutch soccer players also are firm believers that penalty kicks should be shot low in the corner, to avoid the chance that they would lift the ball over the goal. Statistics, however, shows that the chances that the goalkeeper will make a save there are several times larger, so that the overall scoring rate is far lower. Yet they keep aiming for the low corners. Needless to say that the Dutch national soccer team is virtually always knocked out when it comes to penalty shoot-outs... Old habits die hard, and people, even professionals with 7-digit salaries, are not always rational.

💡📝Kevin Pacey wrote on Fri, Dec 18, 2015 09:51 AM UTC:
H.G. posted some time ago:

"Some of your piece values are off, especially Archbishop, which is about C - 0.25 P = Q - 0.75 P (so 9.25 on your scale). The Amazon seems to be worth only Q+N, so 13 on your scale.
..."


I'm still wondering about the value of an Amazon asserted to be only Q + N, in your opinion (besides that of other people). Maybe it is since I am a fairy chess newbie, but I'm not clear on subsequent remarks you have made regarding synergy, in regard to them being fully in line with saying that an Amazon just = Q + N. Furthermore, assuming that value is the measurement your method yielded, that result is still a red flag for me presently, as far how infallible the method or its playtesting conditions might be. 

As I alluded to earlier, in particular the supreme double attacking powers of an Amazon make me value it more than just a Q + N. Currently I am more willing to accept that I gave a tentative value for an Archbishop that was too low, by contrast (though it may be that the value of an Archbishop in the context of Sac Chess, rather than when it is used just in an 8x8 variant approximating standard chess, still ought to be measured from scratch by testplaying using Sac Chess games, if anyone is willing). There is another red flag for me concerning the method, regarding comparing a bishop to a knight by such measurement, resulting in asserting they are of equal value. All about that further below.


H.G. posted more recently:

"The values were indeed measured by play-testing through self-play of computer programs. To measure the value of, say, an Archbishop, I set it up opening positions where one side has the Archbishop instead of a combination of other material expected to be similar in value (like Q, R+B, R+N+P, 2B+N, R+R). For any particular material imbalance the back-rank pieces are shuffled to promote game diversity. I then play several hundred games for each imbalance, to record the score. This is rarely exactly 50%, and then I handicap the winning side by deleting one of its Pawns, and run the test again. This calibrated which fraction of a Pawn the excess score corresponds to. E.g. Q vs A might end in a 62% victory for the Q, and if Q vs A+P then ends in a 54% victory for the A+P, I know the P apparently was worth 16%, so that the 62% Q vs A advantage corresponds to 0.75 Pawn. 

I tried this with two different computer programs, the virtually knowledgeless Fairy-Max, and the 400 Elo stronger Joker80. The results are in general the same (after conversion to Pawn units), and also independent of the time control. (I tried from 40moves/min to 40 moves/10min.) Typically they also are quite consistent: if two material combinations X and Y exactly balance each other (i.e. score 50%), then a combination Z usually scores the same against X and Y. 

The results furthermore reproduce the common lore about the value of orthodox Chess pieces. E.g. if I delete one side's Knights, and the other side's Bishops, the side that still has the Bishop pair wins (say) by 68%, and after receiving additional Pawn odds, loses by 68%. Showing that the B-pair is worth half a Pawn. Deleting only one N and one B gives a balanced 50% score, showing that lone Bishop and Knight are on the average equivalent. This is exactly what Larry Kaufman has found by statistical analysis of millions of GM games.
..."


First off, I think I see how the Archbishop's value was measured, in terms of being 0.75 Pawns less than a Queen, based on the percentages given. Whether just hundreds of games is a statistically satisfactory playtest sample size, I am not sure (note that in chess White is thought to have a standard statistical edge, by about 54% or 55% over Black, so I assume half the time the Archbishop was with White thoughout the playtests). It also could be important how highly rated the computer programs were. By way of illustration, in chess it takes a good degree of skill for human players to know how to defend against a queen using, say, R + B + P, in situations where they are worth at least the queen objectively, based on the current position on the board.

Larry Kaufmam is an International Master as far as chess goes, which puts him below Grandmaster or certainly world champion level, and such players have in the past and present certainly believed that though a bishop is close in value to a knight in terms of relative value, in general unless there is a special reason to prefer having a knight, situations favouring a bishop tend to happen more often - whether in actual game play, or in the many calculated variations that could have arisen from them (these alas do not appear in the playtesting process). So, if a grandmaster willingly gives up a bishop for a knight, he has reasons to do so based on other factors in the position. 

In any event, I have not heard of any reasonably strong human chess players changing their strategies in regard to trading bishops for knights in over the board play, based on Kaufman's result from his method. In regard to the millions of games Kaufman looked at, I am not sure all the chess Grandmasters in history have played close to a million games yet, especially against just each other. I have chess game databases with over a million games in them, but they include vast numbers of games played by players who were below Grandmaster level at the time.

H. G. Muller wrote on Fri, Dec 18, 2015 09:25 AM UTC:
Indeed, piece values are a concept that assumes an additive model for material evaluation, where the value of your army is the sum of the individual piece values. But in reality this additivity is only a (usually pretty good) approximation. The effective value of a piece is affected by what other friendly or enemy pieces are on the board. This is well known from the Bishop pair bonus, and even more dramatically demonstrated by the fact that 3 Queens lose so badly from 7 Knights (in the presence of Pawns). <p> It turns out that under 'normal' circumstances, of a varied army of opposing pieces, with widely spread values, such non-additive effects are pretty small. The dominant effect here is that stronger pieces devaluate by the presence of weaker opponent pieces, as the latter force a trade-avoiding strategy on them, reducing their usefulness. But this can be accounted for as a second-order correction to the base values of the pieces, which then makes these base values even more universally accurate. <p> The base values themselves are also not just the sum of contributions from the individual moves, as the difference in value between Q and R + B shows. Usually adding move sets causes some synergy. This is even true if the added sets are without obvious defects. E.g. the Phoenix, which moves one step orthogonally or jumps two diagonally has 8 move targets, like the Knight, and neither of them is color-bounded or suffers from the Pawn hurdle. So not surprisingly they have very similar value, ~3. If you combine their moves, however, the resulting piece has 16 unblockable move targets, which corresponds to a value of 7-7.5, i.e. a synergy of about 1 Pawn. <p> Such a synergy is to be expected, as the moves of set A help to aim the moves of set B on their target, and vice versa. A quantitative indicator for this is the number of squares you could reach in 2 moves. E.g. a 'Narrow Knight', which only has the two forward-most and two backward-most moves of the Knight, can reach 4 squares in one move and 8 squares in two moves (plus a return to its starting point). Combine it with the complementary 'Wide Knight' to a normal Knight, and it can reach 8 squares in one move, and 32 in two. So although the number of moves simply added, the number of two-step tours quadrupled. Now what you can do in two moves is not as damaging as what you can do in one, as the opponent can see it coming and gets a turn in between to take conter measures, but it is not totally unimportant. <p> Apart from this 'manoeuvrability' synergy, there are the synergies due to moves repairing each other's defects that you mentioned, like color binding or Pawn obstruction. An alternative, equivalent way to look at them is as a penaly on the unfavorable combination of moves that had the defect. E.g. color binding, and in particular higher-order color binding, can be very damaging to the performance. A Shatranj Elephant, which jumps two diagonally, can reach only 8 squares on the board, and is almost worthless (~0.5 Pawn) despite its 4 moves. A piece that moves one step diagonally forward or straight back would be worth more, even though it only has 3 moves. A non-royal King (Commoner) suffers from lack of 'speed', and adding a move with a longer stride to it would be worth a lot. These kind of deficits typically occur in pieces with very few moves. <p> As to the Ferz vs Wazir value: if you take a piece with very many moves, so that taking away a few will not introduce severe deficits in terms of speed or color binding, you can disable individual moves to measure how the value suffers from it. This revealed that forward moves contribute approximately twice as much as sideway or backward moves, (and that captures contribute about twice as much as non-captures). This explains why the Ferz is better (at least in pairs): it has two forward moves, and the Wazir only one. The Wazir also suffers from the Pawn hurdle. In fact, it turns out that all Rook-like pieces I tested so far (several limited-range Rooks, of which the Wazir is the most extreme case), starting them on the back rank behind a closed wall of Pawns makes their value come out 0.25 Pawn lower than when you start them on an open file or before the Pawns. Wazirs starting on d3/e3/d6/e6 test as about 125 centi-Pawn, while on the back-rank they hardly beat a Pawn. Also normal Rooks in the starting setup tend to test as 4.75, rather than 5. This can be seen as an 'open-file bonus', which is sort of implied in the end-game. (And the classical piece values are end-game values!) Note that a Wazir also has lower 'speed' than a Ferz. To catch a Passer the Ferz has to be in the same square area as a King would have to be, but the Wazir has to be inside a triangle half the area.

John Whelan wrote on Thu, Dec 17, 2015 10:21 PM UTC:
Muller, thank you for your analysis.  You gave me some things to think about.  However, I'm still not sure we're entirely on the same page.

You mention the Wazir, and how it's barely worth more than a pawn.  So I asked myself, why would that be?  I now hear that the analogous-but-diagonal Ferz is considered more valuable, despite being colorbound.  Why would that be?  They have equal "firepower".  A little thought produces the answer.  The Wazir is devalued by its lack of mobility, especially on a board crowded with pawns (and others).   A Ferz can easily slip through pawn formations (which depend on diagonals), but a Wazir cannot.

These same considerations will impact a Rook when compared to a Bishop.  With that in mind, it seems likely that the colorbound nature of the Bishop does affect its value, but this probably does little more than balance the pawn-bound limitations of the Rook.

This also explains the phenomena you discuss.  A rook/knight combo breaks the rook's pawn-bound status, and is probably more valuable than a knight+rook as separate pieces.  A bishop/knight combo breaks the bishops color-bound status, and is probably more valuable than a knight+bishop as separate pieces.

But a queen already breaks both these limitations, and a Knight's powers probably don't break them much further than they are already broken.  I therefore still doubt very much that an Amazon is worth more than the combined value of Queen + Knight.  

The pawn-synergy factor brings me back to my original point.  The value of a piece will depend on what else is on the board and the synergy or lack thereof between them.  "Sac Chess" throws us all for a loop by radically altering the other pieces on the board.  It's guesswork, and the value of a piece as measured in one context (such as a close approximation of FIDE Chess), will not necessarily apply here.

H. G. Muller wrote on Thu, Dec 17, 2015 09:14 AM UTC:
> <i>I'm not sure how you measured an Archbishop's relative point value for every case that you mentioned. ...</i> <p> The values were indeed measured by play-testing through self-play of computer programs. To measure the value of, say, an Archbishop, I set it up opening positions where one side has the Archbishop instead of a combination of other material expected to be similar in value (like Q, R+B, R+N+P, 2B+N, R+R). For any particular material imbalance the back-rank pieces are shuffled to promote game diversity. I then play several hundred games for each imbalance, to record the score. This is rarely exactly 50%, and then I handicap the winning side by deleting one of its Pawns, and run the test again. This calibrated which fraction of a Pawn the excess score corresponds to. E.g. Q vs A might end in a 62% victory for the Q, and if Q vs A+P then ends in a 54% victory for the A+P, I know the P apparently was worth 16%, so that the 62% Q vs A advantage corresponds to 0.75 Pawn. <p> I tried this with two different computer programs, the virtually knowledgeless Fairy-Max, and the 400 Elo stronger Joker80. The results are in general the same (after conversion to Pawn units), and also independent of the time control. (I tried from 40moves/min to 40 moves/10min.) Typically they also are quite consistent: if two material combinations X and Y exactly balance each other (i.e. score 50%), then a combination Z usually scores the same against X and Y. <p> The results furthermore reproduce the common lore about the value of orthodox Chess pieces. E.g. if I delete one side's Knights, and the other side's Bishops, the side that still has the Bishop pair wins (say) by 68%, and after receiving additional Pawn odds, loses by 68%. Showing that the B-pair is worth half a Pawn. Deleting only one N and one B gives a balanced 50% score, showing that lone Bishop and Knight are on the average equivalent. This is exactly what Larry Kaufman has found by statistical analysis of millions of GM games. <p> BTW, John Whelan's claims are at odds with the facts. Combining pieces in general makes the compound more valuable than the sum of components (if they had no common moves, of course). For short-range leapers this is summarized by the empirical formula for the value of a (symmetrical) piece with N move targets: 1.1*(30 + N*5/8)*N (in centi-Pawn). The quadratic term in this causes the synergy value. Also, slamming extra short-range moves on a slider, like upgrading the Bishop to Missionary, increases the value by about 2 Pawns, while a piece with only the 4 extra moves (the Wazir) proves hardly worth more than a single Pawn (~1.25), and then only if you start in in a favorable place (open file). <p> That the Bishop is not just a Rook devaluated by its color boinding (a claim which John already retracted, I believe) can be easily seen from the fact that a pair of 'augmented Bishops', which move as Bishop but have an extra backward non-capture move that allows them to switch colors, are hardly superior to a pair of ordinary Bishops, (like ~1/3 of a Pawn for the pair). And that the small difference there is is very close to the advantage you would get by putting this extra backward non-capture on the pair of Knights. So the main effect of this extra move is just increased tactical agility. Of the augmented Bishop does not involve a pair bonus, though. So one could postulate that the negative effects of color binding are almost completely masked when you have a pair of the piece on opposite colors, and are only felt when you have a single copy of the piece.

💡📝Kevin Pacey wrote on Thu, Dec 17, 2015 01:15 AM UTC:
To H.G.:

I'm not sure how you measured an Archbishop's relative point value for every case that you mentioned. Did you always base your measurements on, say, the outcomes of a large number of games that were play-tested which involved a variant using that piece? If so, the results might at least somewhat depend on the average strength of the players involved, even if all/many were computer playing engines, though I would assume you took that into account if that was always your method. Hopefully the method can be described briefly, if you are happy to do that.

To John:

It looks like you have a valid point about saying an Amazon = Q + N is fundamentally different than objecting that a Q is greater than a R + B, in comparing by analogy. I hadn't thought about a B's limitations when a piece by itself.

[edit: though a vital follow up question could be: are 2 Queens only = 2 Rooks + 2 (different coloured) B's, or are 2 Q's in fact = 2 Rs + 2 (different coloured) Bs + 2 Pawns? Looking at your last post before this one of mine, I assume you now would put the 2Qs value as closer to the latter.]

[edit: Another follow up question could be: are 2 Amazons = 2 Qs + 2 Knights? The reason I thought of asking is that either Amazon might (using just its bishop and/or rook type powers) be able to double attack the 2 enemy knights, or else attack one enemy knight and another enemy piece. In fact, a single Amazon could do the latter in case of having a Queen and Knight for it, so again the real question becomes: is an Amazon really just = Q + N (or even less)? To strengthen my doubts a little more, I would note a single Amazon can also use its knight type power to double attack an enemy queen and another enemy piece, say even another queen, though such a queen might often have a good chance to move and guard the other piece that is being double attacked by the Amazon. As an aside, fwiw long ago I read in some chess book that it is the great mobility and double attacking capability of a queen that makes it such a powerful chess piece, and a knight of course is awesome at forking.
]


In any case, I hope it's not too objectionable to anyone if I leave my own estimates for Sac Chess pieces (which I admitted were tentative) the way they are for at least a little while longer. I can edit my submission to change my estimates after I think about it more if necessary, and perhaps have even play-tested variants using Amazons or Archbishops myself at some point. 

[edit: I had forgotten that I've already played a quite small number of games of Seirawan Chess (8x8 chessboard variant, with first rank drops of Hawks [aka Archbishops] and Elephants [aka Chancellors] in the opening phase) and it seemed to me based just on these games (two of them with a fellow chess master) that the Chancellor seemed at least as dangerous a piece as the Archbishop, especially after the opening phase, if nothing else. Also, I wonder if there really has been enough human experience playing with such fairy chess pieces yet for even a very strong human chess player to really know how to defend (when necessary) against their unusual movement capabilities.]

John Whelan wrote on Thu, Dec 17, 2015 01:14 AM UTC:
On consideration, it seems I must revise my position that the one-color only limitations of the Bishop seriously affect its value.

The Rook controls 14 squares on an empty 8x8 board.  The Bishop controls 13 at most (when in the center of the board) but more often only 7 (when at the sides).  The average for the whole board seems to be control of 8.75 squares, though in practice that number will be higher as players will tend to position their bishops to best advantage (i.e., they will tend to avoid sides and corners).

Divide both by a similar factor (2.8) and you get 5 for a Rook, and 3.125 for a Bishop (or perhaps higher in practice for the reasons stated), which is not too different from their relative values as found in practice.

🕸Fergus Duniho wrote on Thu, Dec 17, 2015 12:57 AM UTC:
Although you might say a Queen combines the powers of a Rook and two Bishops, this is not precisely true. At most, from any given position, a Queen may move as a Rook or as the Bishop on that color. If it's on a Black square, it cannot move diagonally along the White squares, or vice versa. So, for any given move, a Queen's possible moves are limited to those that can be made by a Rook or by one Bishop. This puts a Queen in a grey area between having the power of two pieces and the power of three pieces.

🕸Fergus Duniho wrote on Thu, Dec 17, 2015 12:41 AM UTC:
It's a fair point to say that the Queen combines the powers of a Rook and both Bishops. But this might not make your case, because the lesser value of a Queen to a Rook + two Bishops might be due to the effect of diminishing returns when you start to combine more than two pieces or to the similarity of one Bishop's move to the other. To make your case, the Marshall is the most unambiguous example. It combines the powers of a Rook and Knight, neither of which were originally colorbound, and neither of which have moves similar to the other. In some tests I just ran on Zillions-of-Games, I had one side use a Marshall in place of its Rook and Bishop. I ran two tests so that each side could use the Marshall in place of its Rook and Bishop, and in each test, the side with the Marshall won.

John Whelan wrote on Thu, Dec 17, 2015 12:22 AM UTC:
>  If the Queen is your example, you must compare the Queen to the value of a Rook plus a Bishop, not to the value of two Rooks.

Okay.  But haven't I already done that?

A Queen does not really have just the power of a Rook and Bishop.  It has the power of a Rook and both Bishops.  It can move like a Rook, and in addition, can move diagonally on black diagonals, and diagonally on white diagonals.  

As to "firepower" (the number of squares it controls at one time), the firepower of the Queen is nearly equal to that of two Rooks.  Or to put it another way, it has the firepower of a Rook and a Bishop (both of which can control an almost equal number of squares), but does not suffer those limitations that restrict the Bishop to one half of the board (which limitation reduces the Bishop's value compared to the Rook, despite it's almost-equal "firepower").

🕸Fergus Duniho wrote on Thu, Dec 17, 2015 12:07 AM UTC:
Let me examine what you said in more detail. You said,

> a single piece that combines the powers of two lesser pieces would be worth less than those two pieces.

Then as evidence for this you gave an example of something else,

> the fact that a Queen controls twice as many squares as a Rook (and as many squares as 2 Rooks), but is worth less than two Rooks.

If the Queen is your example, you must compare the Queen to the value of a Rook plus a Bishop, not to the value of two Rooks. Because you did not, there is no logical connection between your evidence and your conclusion.

John Whelan wrote on Wed, Dec 16, 2015 11:39 PM UTC:
>  You're doing your math wrong. A Queen at 9 points is worth more than 
>  a Rook (5 points) + a Bishop (3.25 points).

There's nothing wrong with my math.  If firepower were the only issue, Rooks and Bishops would be worth the same (5).  A Bishop is worth less because of its limitations.  A Queen is worth more than Rook and Bishop because it gains the Bishop's firepower without suffering from its limitations.

Imagine a piece that had the power to move like a Rook and, in addition, the power to move diagonally but ONLY along black diagonals.  This piece would come closer to combining the powers of a Rook and (black-square) Bishop.  And its value would be less than the combined value of those two pieces. 

A Queen does not really combine the power of one Rook and one Bishop.  It might be closer to the truth to say it combines the powers of a Rook and both Bishops.

🕸Fergus Duniho wrote on Wed, Dec 16, 2015 11:25 PM UTC:
> But I think that generally, a singe piece that combines the powers of two lesser pieces would be worth less than those two pieces. As evidence of this, I offer the fact that a Queen controls twice as many squares as a Rook (and as many squares as 2 Rooks), but is worth less than two Rooks.

You're doing your math wrong. A Queen at 9 points is worth more than a Rook (5 points) + a Bishop (3.25 points).

🕸Fergus Duniho wrote on Wed, Dec 16, 2015 11:22 PM UTC:
Test comment. Intended for deletion.

I can't delete this unless I do so as an editor.

John Whelan wrote on Wed, Dec 16, 2015 11:22 PM UTC:
I don't think one can be too dogmatic about the relative value of pieces across all contexts.

It is very useful in Chess to throw lesser pieces onto the front lines for profitable exchanges, while using the Big Guns for backup.  Whether a Big Gun is worth more to you than 2 lesser value pieces will probably depend on the balance between Big Guns and lesser pieces that you already have.

But I think that generally, a singe piece that combines the powers of two lesser pieces would be worth less than those two pieces.  As evidence of this, I offer the fact that a Queen controls twice as many squares as a Rook (and as many squares as 2 Rooks), but is worth less than two Rooks.

A Bishop has the same firepower as a Rook, in terms of the number of squares it controls.  The reason a Bishop is worth less than a Rook is that a Bishop is confined to half the board, and the Rook is not.  A Queen gains the Bishop's extra firepower without suffering from its limitations.  This is the reason why a Queen is generally worth more than a Rook and Bishop combined.

This logic does not apply to the Amazon.  Neither the Knight nor the Queen is limited in the way the Bishop is; so a piece that combines their powers should be worth less than the value of Knight+Queen, not more.

I agree that 10x10 board size will tend to increase the value of unlimited range pieces.  On the other hand, the SAC chess board is rather crowded, and this may tend to decrease the value of ranged pieces (while increasing the relative value of the knight's unblockable movement). SAC chess starts with 60% of the board occupied, as opposed to 50% for FIDE chess.

Derek Nalls wrote on Wed, Dec 16, 2015 10:59 PM UTC:
Editors: We need the ability to delete one of our own comments for whatever appropriate reason.

Derek Nalls wrote on Wed, Dec 16, 2015 10:33 PM UTC:
You requested a third party opinion. I have playtested Muller's relative piece values in CRC and found them to be extremely reliable. In fact, I was so intrigued by my verification of his correct (yet surprisingly high) value for the archbishop that I revised and expanded my own work to drive deeper into the underlying geometric & arithmetic foundations in a somewhat successful attempt to gain a theoretical understanding as to why.

H. G. Muller wrote on Wed, Dec 16, 2015 06:57 PM UTC:
Oh yes, you can read many things on the net. Especially nonsense is extremely abundant... <p> The point is that I don't have to <i>estimate</i> the value of most of these pieces: I have <i>measured</i> them. And guessing, no matter how educated, which is the only source of information behind anything you wil read about values of unorthodox elsewhere on the net, is just no substitue for accurate measurement. <p> Fact is that if you play the FIDE start position where for one side you substituted the Queen for an Archbishop between equally strong opponent's, the Queen wins less often than when it had been granted Pawn odds. If the side with the Queen gets its f-Pawn deleted, the side with the Archbishop will beat him more often than not. <p> Another observarion is that in an end-game with many Pawns a single Archbishop beats Rook + Knight + extra Pawn more often than not.

💡📝Kevin Pacey wrote on Wed, Dec 16, 2015 06:27 PM UTC:
Thanks for your kind words. I'll add just a bit more to what I wrote:

I forgot to note that I've read somewhere on the net that it is the Chancellor that is approximately worth a Queen, while the Archbishop is worth about what I gave its value for. Perhaps someone else may chime in if you still do not concur.

The Amazon, I also read, has been estimated as low as 11 or 12 points, but perhaps this does not take into account it being on a larger board. In any case, I found it hard to believe that a queen and knight acting seperately are as effective as an Amazon in general. Compare that by analogy to asserting that a queen would be worth only a rook plus bishop on a standard chess board. Still, as I wrote I am new to fairy chess.

H. G. Muller wrote on Wed, Dec 16, 2015 06:19 PM UTC:
Well, keep up the good work with the blood and such. And just keep this message for later. <p> As a non tech-savvy person you would be an ideal tester for the diagram wizard. When you are also new to Chess variants, however, you might not be familiar with the 'Betza notation' for how pieces moves, and the diagram wizard expects you to specify the moves in this notation. But for your pieces in Sac Chess, which are simple compounds of orthodox Chess pieces, it would just be a matter of combining the letters. E.g. the move of your Judge would be written as KN (or NK), that of the Archbishop as BN, etc. It only gets complex when the pieces are not symmetric and only move in a sub-set of the directions, or when they capture different from how they move, or must jump others in order to move, etc.

💡📝Kevin Pacey wrote on Wed, Dec 16, 2015 05:48 PM UTC:
Sorry, I haven't played much around with modern engines. I'm not tech-savvy at all. I joined The Variants Page in order to give my ideas for chess variants at the least some more exposure on the net, though I haven't lots of free time these days to try to promote any of them vigourously, if I wished to.

I can try out The Variant Pages diagram generator at some point to update my submissions as requested, e.g. for Sac Chess, though I would note I'm virtually new to the chess variant world as far as conventions or terminology goes. Today I'm still recovering from giving blood for the first time last evening (partly for my own health benefit), so I'm going to try to take it easy for a while.

Take care, Kevin

H. G. Muller wrote on Wed, Dec 16, 2015 10:45 AM UTC:
Some of your piece values are off, especially Archbishop, which is about C - 0.25 P = Q - 0.75 P (so 9.25 on your scale). The Amazon seems to be worth only Q+N, so 13 on your scale. <p> As a concequence, the opening line you suggest seems suicidal for black, as bad as being piece vs Pawn behind. I doubt any development advantage would make up for such a huge material deficit. <p> As to the computer resistance: can you beat the Sjaak II engine in this game?

H. G. Muller wrote on Tue, Dec 15, 2015 10:23 PM UTC:
Yes, or the interactive diagram generator. I am looking for feedback on the design wizard for the interactive diagrams anyway. Is everything sufficiently clear, and easy to operate?

100 comments displayed

LatestLater Reverse Order EarlierEarliest

Permalink to the exact comments currently displayed.