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 ]

Ratings & Comments

LatestLater Reverse Order EarlierEarliest
Catapults of Troy. Large variant with a river, catapults, archers, and trojan horses! (8x11, Cells: 88) [All Comments] [Add Comment or Rating]
💡📝Gary Gifford wrote on Tue, Sep 25, 2018 01:55 AM UTC:

Thank you for your detailed review, Kevin... it is appreciated. I've not played this game in many years, but I do not recall the play being slow. Antoine Fourrier created a Zillions version of this and if you play that I think you will find the game moves along pretty well... also, if you play another human I think you will see it moves along well.  I did find one minor bug in the Zillions app concerning the Ram...not sure if it still exists after all these years. If I ever  update  the rules it would be to delay movement of the Traojan Horse 3 moves. I have reasons for this.... but will not go into them at this time.  Thanks again for the comment.


Six Fortresses. Capture Fortress to releases free pieces to drop. Imposter Kings will complicate checkmates. (9x9, Cells: 81) [All Comments] [Add Comment or Rating]
💡📝Gary Gifford wrote on Tue, Sep 25, 2018 01:39 AM UTC:

I created Six Fortresses 12 years ago and no longer recall the game... there is no working  link to the setup.... If I can find the game in old notes I will see if I can update this.... otherwise, it seems to be a lost game.


PHP 7[Subject Thread] [Add Response]
🕸Fergus Duniho wrote on Mon, Sep 24, 2018 09:02 PM UTC:

There was a tab character after the closing PHP tag. Redirecting works only if the script doesn't display any text. I deleted it, and it works now, though I had to refresh the page to see the results change in the menu.


Greg Strong wrote on Mon, Sep 24, 2018 04:31 PM UTC:

If I attempt to un-favorite a game from the heart menu on the game's page I get the following error: "The page isn’t redirecting properly"


Chess with Different Armies. Betza's classic variant where white and black play with different sets of pieces. (Recognized!)[All Comments] [Add Comment or Rating]
H. G. Muller wrote on Mon, Sep 24, 2018 12:57 PM UTC:

Indeed, I addressed this this inconsistency in a follow-up comment, at the time. Paper-Sciccors-Rock situations are very uncommon with piece values; usually the empirically determined value of a piece is highly independent on what you play it against. (Except for extreme situations such as 3 Queens vs 7 Knights; it has to be a mix of pieces of different value.) I guess that this is why 'piece value' is a useful concept in the first place.

It could be that the Clobberers are composed such that they can better exploit the most important weakness of the Nutters, namely that they cannot quickly pull back. The Clobberers have only one major piece, but they have several combinations of two minor pieces that together can force checkmate (through repetitive checking) on an unprotected King. As Kings tend to stay on the back rank until the late end-game, it is rather tempting for a naive Nutters player to abandon its King while aggressively attacking (possibly gaining significant material), to discover that a counter-strike expedition of two pieces will unescapably kill its King. I don't think any of the other armies has the ability to inflict mate with such a small force. (In FIDE there is the pair of Rooks, but that already fails when there is a Pawn to shelter behind, while the FAD can jump.)

So it seems it is more important for the Nutters to have some strategic knowledge (which Fairy-Max utterly lacks), namely that it should always keep a 'sweeper' piece near the back rank to defend its King against sudden break throughs. That the opponent also doesn't know that this is a weakness, and won't intentionally lure the Nutter pieces forward (e.g. by forcing them to make a forward distant recapture) only partly compensates this ignorance, as it will happen enough that the Nutters will just accidentally (unforced) move their pieces ahead. This is actually statistically likely, as the Nutter pieces in general have more forward than backward moves. So they tend to 'drift' forward when they would wander around aimlessly. In an engine with the required strategic knowledge the Nutters should do even better, though, and they were already one of the strongest. So if the method has a systematic error here, it is in the wrong direction.

The reason I was not so worried about 'disadvantaging' the Nutters by denying them the opportunity to promote to Queen / Marhall / Archbishop is that they already seem to have too strong an army despite this 'handicap'. Also, the disadvantage for the Clobberers that they cannot do better than Archbishop is not nearly as large as what Betza thought, as the Archbishop is unexpectedly strong. And also with this handicap, Clobberers seem stronger than FIDE. So I thought it entirely acceptable to limit promotion to each army's own super-piece.

If it would help, it would not be a bad idea to limit the Rookie's promotion choice to { Fibnif, Short Rook, Half Duck }, or perhaps even just to Fibnif. But such promotion limitations seem actually ineffective in altering the strength of an army.


Aurelian Florea wrote on Mon, Sep 24, 2018 11:52 AM UTC:

I wanted to notice something in HG's old results.

I see the CC and the NN are balanced against each other but the CC behaves quite worse versus the RR (10%). So it seems same small but not insignificant rock-paper-scisors effect is taking place. This could be due to a possible need (I mentioned a long time ago in the different context) of a concept of multidimensional piece values. But it is probably more than that if any such thing is possible? The NN are a "pressure" army as they have more forward moves and the RR are usually slower as they can't turn a corner that easily. This seems a reduction of the weakness of the NN. On the other hand CC has the strategic weakness that it can be twice impaired by te lack of an counterpart of the other color bound piece.. RR can profit more easily from that as because of it slowness weakness it is a more strategic army herself. The NN don't have time for such debates. They need to "act" so they can't profit out of it.

Such lines of reasoning are most likely usefull but I can't pinpoint why I'm a bit uncomfortable with the idea of studying it exactly here. Maybe the game dimensions.

 


H. G. Muller wrote on Mon, Sep 24, 2018 10:09 AM UTC:

I looked up my old comment in this topic ( https://www.chessvariants.com/index/listcomments.php?id=31222 ). There I report that the Rookies were actually strongest of all. So RR >~ NN > CC > FF. This was based on the scores in 400-game matches between each pair of armies with Fairy-Max. In a comment just before it Fergus had arrived to the same conclusion based on ChessV (but with far fewer games). Fairy-Max randomizes the first 4 opening moves of each player, which should be enough to not have significant duplication of games. (I did not actually check for duplicate games.)

My experience with this kind of materialy-imbalanced testing is that the result is not very sensitive to the piece values used by the engine. E.g. if you give one player an Archbishop instead of a Queen, and assign it a value of 900 (where Q=950), the side with the Queen will score about 62%. If you then repeat the test with A=1000), the player with Queen will still score around 62%. The reason is likely that, as long as the values are different, 1-for-1 trading is not frequent, because there is always one player that thinks it is to his disadvantage, and will avoid it. And it does not matter much which player this is. The imbalance is therefore long-lived, and you measure the relative effectivity of the imbalanced pieces for doing (or helping to do) damage to the common pieces. Which is pretty much independent of how the computer values them, as they will mostly not be traded directly for other material (2-for-1 trades are also pretty rare).

So as long as both players share the misconception on actual value, the programmed value doesn't seem to be very critical. Of course it should not be totally off; if you set the value of a Queen below that of a Pawn, it will indeed get worth as much as a Pawn, because it will be immediately traded for one. There is just no way the other player could shield all its Pawns from Queen attack, before the Queen gets to see it can force a more profitable trade. If you assign reversed values to pieces that differ very much in power, the strong one will probably succeed to force it being traded for the weak one, which it mistakenly considers profitable.

Before I did the test with complete armies, I did similar tests on all individual pieces in the armies to determine their value. E.g. use FIDE as context, and then replace Rooks of one side by (say) HFD on the other, to see whether HFD is better or worse than an orthodox Rook (and by how much). In such tests I always make sure they are self-consistent, i.e. performed with the programmed value equal to the value suggested by the eventual score. If my initial guestimate of the value was wrong in this respect, I just repeat the test with the value suggested by the outcome of the flawed test. Which then usually does not significantly alter the outcome. This should have made the individual piece values more or less OK, so that the play during the whole-army tests must have been realistic, and thus must have made the sampling of what the pieces can do representative.

I am a bit surprised about the low score imbalance you get for CC-FF. My old results table says +9% for this (meaning the match score averaged over both colors was 59% in favor of the Clobberers). You get only 52.6%. What is the Pawn-odds score for ChessV (i.e. when you use equal armies, except that one of the players gets f2 or f7 deleted)?

 

These are the piece values Fairy-Max is using (Pawn = 100):

FIDE
Knight           325
Bishop           350
Rook             500
Queen            950

Clobberers
Waffle           320
Fad              480
Bede             530
Archbishop       875

Nutters
Fibnif           310
Charging Knight  400
Charging Rook    485
Colonel          935

Rookies
Woody Rook       310
Short Rook       400
Half Duck        480
Marshall         935

 


Kevin Pacey wrote on Mon, Sep 24, 2018 06:20 AM UTC:

I'd rate the RR army and the FF army about the same, though the mobilitity of a Q may give an edge to the latter. The CC army would seem to be at least slightly better than either (based on material valuations) IMO, but noting, perhaps very significantly, that it seems it could often be really awkward to develop both of the two waffles (i.e. not just one) that a player has with any speed, especially with the Black pieces (the edge pawns being unprotected at the start doesn't help either, especially vs. a FIDE army, with its Q). The NN army seems clearly the best army of the 4 in theory to me (based on material valuations), except I've yet to play with it, rather than against it. Thus I find myself pretty much agreeing with H.G.'s assessment of the 4 armies' relative strengths, though perhaps for many differing reasons.

[Edit: It seems my memory of the relative strength of the 4 armies that I estimated long ago was rather off. In any case, here's my current tentative estimates of each army's relative strength (based on material valuations alone):]

FIDE:

Knight 3.5 approx. Bishop 3.5 Rook 5.5 Queen 10 Army (w/o Ps/K) 35

Clobberers:

Waffle 3.125 Fad 4.75 Bede 5.125 Cardinal 8 Army (w/o Ps/K) 34

Nutters:

Fibnif 3.125 Charging Knight 4 Charging Rook 4.9375 Colonel 8.9375 Army (w/o Ps/K) 33.06

Rookies:

Woody Rook 3.125 Short Rook 3.54 Half Duck 4.915 Marshall 10 Army (w/o Ps/K) 33.16


Home page of The Chess Variant Pages. Homepage of The Chess Variant Pages.[All Comments] [Add Comment or Rating]
🕸📝Fergus Duniho wrote on Sun, Sep 23, 2018 05:07 PM UTC:

Just testing to make sure nothing is broken after updating scripts for entering comments.


PHP 7[Subject Thread] [Add Response]
Greg Strong wrote on Sat, Sep 22, 2018 11:43 PM UTC:

Ah, got it.  Thanks!


🕸Fergus Duniho wrote on Sat, Sep 22, 2018 11:19 PM UTC:

This game has its rules described in a minirules file, not in the briefrules value. Minirules files precede the use of the briefrules field. I added the latter, because only editors could upload minirules files. Go to the /play/pbm/minirules directory to edit this file.


Greg Strong wrote on Sat, Sep 22, 2018 10:17 PM UTC:

Here's an example:

/play/pbm/play.php?game%3DOpulent+Chess%26settings%3DAlfaerie

The box for specifying HTML rules is empty, at least for me


Home page of The Chess Variant Pages. Homepage of The Chess Variant Pages.[All Comments] [Add Comment or Rating]
Fergus Duniho wrote on Sat, Sep 22, 2018 09:46 PM UTC:

The new CAPTCHA displays two words that are related to Chess variants, and it is case insensitive, because some capital and lowercase letters look the same.


Fergus Duniho wrote on Sat, Sep 22, 2018 09:42 PM UTC:

Anonymous sign in has been broken ever since reCAPTCHA made the code I was using obsolete. Instead of updating to the new version and waiting for it to break down eventually, I wrote my own CAPTCHA processing code. It consists of three parts. There is a script that displays a graphic image of a CAPTCHA, a script that displays the form for logging in with a CAPTCHA, and the script for checking whether the text matches the CAPTCHA. This comment is a test of this.


Chess with Different Armies. Betza's classic variant where white and black play with different sets of pieces. (Recognized!)[All Comments] [Add Comment or Rating]
Aurelian Florea wrote on Sat, Sep 22, 2018 03:00 AM UTC:

I think the NN despite their many weaknesses they have a wonderfull middle game. That should probably always do it!

@Greg The Bede thingie seems a good idea for me and it does sound more natural for a rook!

About the RR, I find the FDH quite akward and it sould be around rook level. It probably is not (but the R4 definetly compensates for it).

Anyway very nice effort on mister Betza part in an era when computers were much weaker :)!


PHP 7[Subject Thread] [Add Response]
🕸Fergus Duniho wrote on Sat, Sep 22, 2018 02:10 AM UTC:

In the test I did, the text I entered in the field showed up in the appropriate place. Can you give me an example preset where this is not working right?


Greg Strong wrote on Sat, Sep 22, 2018 12:09 AM UTC:

Thank you for fixing so quickly!

Sorry to say I've found another problem ...  If you edit a Game Courier preset, even if it has HTML rules defined, the text-box that should show you the rules is empty.


Chess with Different Armies. Betza's classic variant where white and black play with different sets of pieces. (Recognized!)[All Comments] [Add Comment or Rating]
Greg Strong wrote on Fri, Sep 21, 2018 10:50 PM UTC:

BTW, I always interpreted the promotion rule in CWdA as that you could only promote to pieces of your own army. This seemed logical to me; promoting to pieces of the opponent army strikes me as unnatural and ugly. But now I believe this is not what Betza originally meant

The rule is definitely that you may promote to a piece in either army.  I know I've seen that explicitly stated, with reasoning, on one of Betza's pages here.  (There are quite a few auxiliary pages about CwDA, armies, and piece values by Betza on the Chess Variant Pages, some of which might not be in obvious locations.  This is something I've been meaning to look into cleaning up.)

But the Colonel has a major shortcoming: it cannot move backwards fast. All native Nutter pieces actually have that problem. So if it comes down to a promotion race, and they can only promote to their own pieces, the Nutters are toast.

Yes, this is exactly right.  It would make the act of pawn promotion for the Nutters wildly inferior to every other army if the they could only promote to their own pieces.

NN > CC > FF > RR

This is certainly not how I would rank them.  I would do almost the opposite.  I'm not sure whether FF or NN is better and I'm not sure if CC or RR is better, but I am confident that both CC and RR are stronger than both FF and NN.

I've started studying this formally.  I was going to wait until I had completed more testing to get into it, but since the subject is being discussed, I'll explain what I have so far, which is some solid testing of the FF vs. CC match-up.

I started by coming up with a number of different, but balanced, opening positions.  I have 20 different opening positions with FF as white and CC as black and another 20 positions the other way around.  These positions are roughly 8 moves deep into the game.  I have also programmed ChessV with the ability to play a list of games one after the other in "batch mode" and record the results.  I then played each of these 40 different positions 10 different times with slightly different time controls and settings and compiled the results:

CC beats FF: 178
FF beats CC: 157
draw: 65

So it certainly appears that the Clobberers are stronger than the FIDEs (which is what I expected.)  It is possible, of course, that these results are not perfect.  Chess programs can only "see" so deeply and then they must evaluate the position, and they do that with parameters we supply - for example, we tell the program how much the pieces are worth.  There are lots and lots of parameters, and while I believe what I have provided are very reasonble, they are almost certainly not perfect.  One important evaluation parameter for this match-up that we don't know - what should the color-bound bonus/penalties be?  In Chess, having both Bishops is worth half a Pawn.  But the Clobberers have two color-bound piece types.  What if you lose both of the pieces on one square color but still have both on the other?  This should trigger a large penalty, since the opponent can avoid both of those peices by occupying the other color, but how large a penalty?  More testing is needed to continue to refine the accuracy of the evaluation parameters ...

The bottom line is that this result should not be considered 100% conclusive, but the difference is large enough that it is almost certain that the CCs are at least somewhat stronger than the FFs.  I will post all my test positions and results sometime soon, but until I release a new version of ChessV with the batch mode people won't really be able to reproduce.

Now, next question.  If we accept this result, and we believe the armies should be closely balanced, what to do?  Obviously the FIDEs shouldn't be messed with, so the Clobberers would need to be weakened.  My thought is to weaken the Cleric (BD) by making the Dabbabah move a "lame" leap - only allow leaping to the second square if the first square is empty (BnD).  I plan to test that change and see what the results are like.

the Nutters seemed to have not much trouble beating FIDE

Interesting, this is not what I would expect.  I think this is the next match-up I'll start testing.  The challenging part of the testing is computing a number of different but balanced opening positions ...  Stay tuned.

 


H. G. Muller wrote on Fri, Sep 21, 2018 09:24 PM UTC:

OK, I see. I don't recall exactly anymore where the RR fitted in; it was one of the last armies I implemented in Fairy-Max (because initially it had trouble doing the R4). BTW, I always interpreted the promotion rule in CWdA as that you could only promote to pieces of your own army. This seemed logical to me; promoting to pieces of the opponent army strikes me as unnatural and ugly. But now I believe this is not what Betza originally meant; he was afraid having different promotion rules would cause the Pawns to have different values, which could easily disturb the balance, as you have so many of those.

But I once made an attempt to weaken an army by allowing Pawns to only promote to a Commoner, rather than a Queen-class piece, and it did not seem to affect the strength of the army at all! At first this seemed very strange, but then I realized that in practice you almost never allow a promoted piece to survive: you sacrifice a minor for the Pawn while you still can, or dedicate a minor to prevent it reaching the promotion square. By that reasoning the value of a Pawn is not so much determined by what it promotes to (as long as that is at least a minor), but more by what the opponent can use to prevent the promotion. This would mean that an army where the weakest piece is stronger (compensated by some of its stronger pieces being relatively weak) causes the Pawns of its opponent to be worth more. Having 7 pieces each worth 4.5 (vs FIDE with 4x3 + 2x5 + 9.5) would pose a real problem w.r.t. stopping opponent passers.

Anyway, since Fairy-Max does not support under-promotion, I had to appoint a unique promotion piece, and chose the most-valuable piece of each army. For the Nutters this was the Colonel. But the Colonel has a major shortcoming: it cannot move backwards fast. All native Nutter pieces actually have that problem. So if it comes down to a promotion race, and they can only promote to their own pieces, the Nutters are toast. Even when they promote 2 or 3 moves earlier, there is no way for them to prevent that the opponent will promote as well. Even worse, the freshly obtained Colonel might not even be able to connect with its own King fast enough, and get lost through a fork on King and Colonel. Especially when the opponent can promote to Queen. The Colonel is completely defenseles against slider or night attacks from behind.

Despite this, the Nutters seemed to have not much trouble beating FIDE. The average superiority of there pieces makes that they hardly ever get into an equal Pawn ending.


Aurelian Florea wrote on Fri, Sep 21, 2018 05:55 PM UTC:

By NN I meant Nutty Nights. Sorry for the confusion. And I was not thinking about this. By definition the charging knight is a major piece. So is the charging rook, that should be obvious, and along with the colonel this means 3 major pieces. Although the colonel is probably weaker than the ordinary king or than the marshal!

The way I see it and I remember you commenting about this on wikipedia, HG (and us having this discussion a while ago) the order of the armies is NN>CC>FF>RR. But they are close. I like this game for the diversity though. In a private talk Vitya Markov has said that he think the RR are the stronger. I had never made this experiments, though.


H. G. Muller wrote on Fri, Sep 21, 2018 04:40 PM UTC:

I am not sure what you mean by that.

It stands to reason that the Charging Knight has mating potential on 8x8: the Gold General already has that, and the Charging Knight has 4 Knight moves instead of the single Wazir step of the Gold. (After you flip it, which should not matter in an otherwise 8-fold -symmetric context.) So it is bound to be a much stronger piece. I see that in Fairy-Max' implementation of CWdA I valued it as 400 cP (where R=500 and N=325).


Unicode Fonts with Chess Piece Images. Images of how the Unicode Chess piece characters are displayed in different fonts.[All Comments] [Add Comment or Rating]
Garth Wallace wrote on Fri, Sep 21, 2018 05:11 AM UTC:

A few more for you:

Chess Merida Unicode is Marroquín's Chess Merida font remapped to the official Unicode code points (glyphs that are not in Unicode, e.g. the pieces on dark squares, are in the Private Use Area). It doesn't contain any glyphs that aren't in the original, so it's strictly for chess only.

Everson Mono is a monospace font with extensive Unicode coverage. The chess symbols are piece outlines/silhouettes rather than the more traditional printed forms, and can be a bit hard to distinguish, particularly at small point sizes.

Fairfax is a 6×12 pixel bitmap font with extensive Unicode coverage and, like Nishiki-teki, an extensive Private Use Area as well (though it does not include variant pieces). The standard chess pieces can be hard to make out, since they're limited to 72-pixel blocks, but points for trying, and it's nice to know that if for some reason you need to use figurine algebraic notation in Klingon text on a console command line, there is a font for you. Unfortunately, Fairfax's non-bitmap "sister" font, Constructium, contains no chess pieces at all.


PHP 7[Subject Thread] [Add Response]
🕸Fergus Duniho wrote on Fri, Sep 21, 2018 02:24 AM UTC:

That's now fixed. The condition for exiting with this message began with empty($SESSION) ||, which would always be true, since the variable name should have an underscore in it, and $SESSION doesn't exist. 


Greg Strong wrote on Fri, Sep 21, 2018 12:22 AM UTC:

The Review Submissions doesn't seem to be working for me.  I can see the list of submissions awaiting review, but when I try to view one I see "If you are the author or an editor, you must sign in to view this submission" although I am signed in.


Unicode Fonts with Chess Piece Images. Images of how the Unicode Chess piece characters are displayed in different fonts.[All Comments] [Add Comment or Rating]
🕸📝Fergus Duniho wrote on Thu, Sep 20, 2018 11:14 PM UTC:

The western style pieces you mention are not circular or wedge shaped, and they are missing some of the images I can see on the page for the font, such as the circular arrow piece, but at least they do include images that can be used for Shogi and Xiangqi, and it would be possible to add these images to the appropriate shapes to make piece sets out of them.


25 comments displayed

LatestLater Reverse Order EarlierEarliest

Permalink to the exact comments currently displayed.