Comments/Ratings for a Single Item
We list at other article the reformists Alexandre 1820's, Bird 1870's, Lasker 1910's, Capablanca 1920's, Maura 1960's, and Fischer 1990's. None of them can be said to have succeeded in their advocacy, but what did they do anyway? In particular, Lasker, why is second world champion on the list? It is easy to locate Alexandre as forerunner of Fischer in randomizing starting positions. With some conviction Bird and Capablanca of course reinvent Carrera for their times. Maura's Modern reaches somewhat across Latin America. What about Dr. Emanuel Lasker, mathematician, friend of Einstein? When Capablanca defeated Lasker to become third Champion, Capablanca tossed around one of his first ideas for reform. It was simply to reverse Bishop and Knight. Lasker then and earlier advocated scoring wins differently by type. ''In order to prevent the decay of chess by the frequent occurrence of drawn games finer nuances of difference of execution must show themselves in the result, and stalemates should be considered and counted in the estimating of scores for tournament purposes, wins by themselves to count less than enforced mates.'' --Lasker's idea summarized by Reti (Source: Richard Reti, 'Modern Ideas in Chess') The ironies are that some GMs, but not variantists, might know of Lasker's scoring proposals, and that today that is the extent of debate within OrthoChess circles, how to reward points differently -- the same topic Lasker brought up 90 years ago.
Okay, thanks that works, and should be able to substitute for Chessboard Math in the code with Game Design or other one next time. 19.February.2008 here are the Falcon associations with Seven days, Seven Wonders, Falcon being the Pyramid, and the other mnemonic 'sevens'. 19.April.2008 is review of Mark Thompson's ''Defining the Abstract,'' the same author of 2002 Tetrahedral Chess.
If you click on the 'Edit' button to the right of David's comment, you can see and copy but not change what he actually typed.
George, the best approach to clicking on Next 25 item(s) that I have figured out is to create links like these: skipfirst=25, skipfirst=50, skipfirst=75. Sorry, but the posts displayed on these pages default to one-line descriptions, in spite of my attempts to tinker with the HTML code in the links.
As the newer posts pile up, a search for paulo will take you down to this indexing post. EDIT: expanding on Joe's follow-up comment, you can simply replace ChessboardMath in the link's A href='http:// expression with the name of any other thread.
[Maybe Paulowich could explain again how to read Comments 26+ in threads, which is readily done in articles.] Mark Thompson's Tetrahedral Chess is on our recognized list at Chessboard Math. Here's a variant 3-D board also with 84 squares, now called Pyramid (another different 3d has name Pyramidal already). One block, or cell, sits centrally atop 3x3 cells, then 3x3 above 5x5, then 7x7. Four levels, or layers, each of 1, 9, 25, 49 cells respectively (=84). Connectivity is easier to visualize than Tetrahedral, and there are all the usual orthogonal, diagonal and triagonal directions of Raumschach (125 cubes, 1907). At the one-cell top level, a Rook has only one direction to move through, first 3x3, then 5x5, and in three steps to the very center of the bottom 7x7. Raumschach King at any corner has 7 cells to which to move, whereas Pyramid King, also omnidirectional, would have only four from the lower corners.
Wikipedia: ''In May 2006 a record-shattering 517 move endgame was announced. Mark Bourzutchky found it using a program written by Yakov Konoval. Black's first move is 1 ...Rd7+ and White wins the Rook in 517 moves.'' Black R-b7; B-b3; K-f4; N-g5. White Kd2; Qh1; Nh2. They are still finding these things after 512 years. What's the rush?
| Who would waste time on Centaur(BN) and Champion(RN) anymore? No one | is interested. Knight was not meant to be compounded but must always | stand alone. I am in total disagreement. The Archbishop (BN) is one of the most elegant and agile pieces ever designed. It is simply marvelous to see it in action, dazziling the opponent. To do justice to its play, this piece should be renamed 'Dancer'.
Who would waste time on Centaur(BN) and Champion(RN) anymore? No one is interested. Knight was not meant to be compounded but must always stand alone. We will call attention to where we prove the inefficacy of those two under this Chessboard Math and Game Design. Please check tomorrow, and you will relieve addiction to Capablanca misadventures Chancellor and Archbishop, whatever they may be called in this or that embodiment. We announced solemnly and theatrically their demise and RIP in January, venerable Centaur and Champion, and sure enough the next day Bobby Fischer died -- after the fact. Check it out.
(3) How many pieces at most does Chameleon capture in one move within Ultima? in Rococo instead? Why the difference? (4) What is the minimum Fool's Mate at Alice Chess? (5) Consider practical piece values the way Betza does without particularly computer aid. First, here compare Rococo Cannon Pawns and Centennial Quadra-Pawns, neither able to promote. Of course, Cannon Pawn benefits from larger board in general, compared to Quadra-Pawn. At 10x10 Cannon Ps. are superior, whilst 8x8 Quadra-Ps. Where would be the cross-over point, 8x9, 8x10, 9x9, 9x10? Which of the two have higher piece value and better winning chances on those intermediate sizes, other things being equal? // Now the upcoming decade of the teens, after these aughts, will probably not see us get so sophisticated as Barton's: (Z) ''By how many seconds a year does proper time in Singapore drop behind a hypothetical reference clock fixed relative to but far away from the sun, (a) due to the orbital motion of the earth; (b) due to its orbital motion jointly with its rotation? (Orbital speed is u = 3 x 10^4 m/s, and the speed of rotation at equator is u' = 460 m/s. Pretend that the axis of rotation is perpendicular to the orbital plane.)''
What is a chess piece but a vector anyway? We want CV problems of the calibre of good Physics questions, like ''(W) A seagull sits on the ground. Wind-velocity is v. How high can the gull rise without doing any work? (X) A siren is fixed at the origin. Wind is blowing at w = 100 km/h from north. Determine (i) the group speed and (ii) the phase speed of sound going north, south, and east. (Y) Neutral pi-mesons (mass m) in flight at speed cB (with respect to laboratory) decay into two photons. Calculate the energy of the photons emitted at a given angle to the flight path.'' --all from Gabriel Barton 'Relativity Principle' (1999). For Chess we can devise rough counterparts. (1) A Falcon sits on the board centrally. How many moves are possible without re-crossing any of its paths on 8x8? 8x10? (2) A Springer(N) is fixed at an opening position. How many moves are maximally possible without crossing any of its paths on 8x8? 8x10? 10x10? [N.b.: these are not classic Tours, which permit route-crossing.] (3) more Problems in follow-up
There are 10^32 or so configurations of Chess pieces on 8x8. Tom Standage writes ''Computers are unquestionably the modern descendants of automata: they are 'self-moving machines' in the sense that they blindly follow a preordained series of instructions, but rather than moving physical parts, computers move information. Just like automata before them, computers operate at intersection between science, commerce and entertainment.'' We are comparing automata from 17th, 18th and 19th centuries -- ''The Conflagration of Moscow,'' ''The Slack-Rope Dancers,'' Chess player ''The Turk'' -- with modern computers. In 1937 Alan Turing published ''On Computable Numbers.'' ''The chess machine is an ideal one to start with for several reasons. The problem is sharply defined, both in the allowed operations and ultimate goal. It is neither so simple as to be trivial or too difficult for satisfactory solution. And such a machine could be pitted against human opponent, giving clear measure of the machine's ability in this kind of reasoning,'' writes Claude Shannon in 1950 ''A Chess-playing Machine.'' All of Turing, John von Neumann, and Oskar Morgenstein were also thinking before, during, and after World War II about the possibility of programming computers to play chess. [Source: Tom Standage 'The Turk' 2002]
One thing that makes chess easier to program for is the standard opening. With a standard setup, there becomes the possibility of opening book, which severely limits the search space of the computer. Go starts with the most options and the game simplifies as it approaches the end. In fact computer can play near flawless Go, if starting from near end of mid-games, yet starting from the first move, computers can only reach low amature dan level. Arimaa is hard bacause of high branching factor by employing multiple moves and weak pieces. Shogi is relatively hard because of the drop rule which increases branching factor as well.
Machines beat human beings at things such as speed and power. I don't see it as a problem if an AI can beat humans at Ortho-Chess. What DOES matter in this area is that Chess doesn't get so optimized in its play that it leads to excessive amounts of draws, and failure of players to creatively beat their opponents, limiting the drama. Also, if a game is feeling overly played out, it begins to lose the community it is supposed to serve. If people raise up the game as some sort of infalliable god, and refuse to look at how it can better serve, then the game has a problem.
Gary Kasparov in promotion for his 1990's Computer matches repeatedly represents himself as ''mankind's last stand'' against Computer. Then he lost to Deep Blue in 1996 and claimed there was at least one move that was not recognizable ''computer move,'' whatever that means. I think ''Chess Variants'' biggest problems are twofold, one the same Computer dominance problem of OrthoChess. There must be solution for it, or all these games will continue obvious decline. Problem Two, the other one is the quality problem, how to determine good games. Who decides? I have said within game conversations to different individuals over years, there are ''prolificists'' (having more than 15 CVs) whose every CV I personally would be ashamed to put my byline on, had they been my own idea or ''invention.'' Yet these games keep pouring out and get published. And the more self-promotion, or outspokenness, the more attention for many, many atrocious CVs. There is serious divide between two opposing camps, not explainable away by debating points. Embarassingly, there is frequently not even common language for evaluation. One prolificist recently indicates complete ignorance of the difference between compound piece and multi-path piece -- concepts at opposite poles from each other. Same problem of prolificism blends into the sheer number of ''inventable'' creations possible, no one really addresses. The Betza Piece Values VI article, recently commented, suggests so many quadrillion -- get that 10^15 and more theoretically workable -- separate pieces, by commenter Levi Aho's calculation, not to mention games-rules' sets. Somehow those without stake in own inventions must start winnowing some categories, and maybe some actual Rules-sets would emerge. Lately Hutnik indicates some intention of the sort, but on side touts Calvinball with ever-changing infinity of Rules-sets.
On [2008-06-04] Joe Joyce wrote:
I lose to Zillions because I tend to attempt to match its speed. While I am beginning to look at 2 possible initial moves, it's already 11 plies down. I'd like a game where a 2 minutes per move time limit was an equal handicap to both me and the computer.
According to the quote in my previous comment, the [Strength] bar in Zillions appears to tick off search depths from 1 ply to 11 ply. Setting this feature at, say, 5 ply should cut Zillions down a lot (5 ply = three moves for the computer and two moves for its opponent).
Back in the Big-board CV:s thread, I also had trouble when clicking on Next 25 item(s). I figured out how to make links like these: skipfirst=25, skipfirst=50. Here is a little known computer trick:
Zillions of Games Discussion Boards
Desired Features for Zillions of Games
Jeff Mallett (Jeffm) Posted on Tuesday, April 24, 2001 - 6:21 pm:
>...would it be simple to have an option to search >to a fixed depth (and then apply quiescence or >whatever search extensions ZoG usually uses) You can do this now. Go to the Computer Opponent dialog and... * Set the search time to infinite * Set the variety to none * Set the strength according to the fixed depth you want (the minimum setting is a depth of 1). This will allow you to compare as you want.
Actually computers are far more sophisticated than merely 'adding machines'. IN fact the computer algorithms that play chess are not brute force. The brute force ones are the ones all GMs and Ims can easily defeat. Computers see many strategical advantages such as doubled pawns, isolated pawns etc _ these are all built in - Computers will choose moves based on above IF there are no branches that will give them an even greater advantage. In fact computers make better decisions by valuing material over positions a bit more than humans . Humans tend to make more unsound sacrifices. Computers don't do so (though they can be programmed to) I think the problem with making a computer play these games is to develop the algorithm which is a human endeavor. the computer is a machine that can handle and process logic that we program. Once an algorithm is developed to prune the unnecessary branches for Go and Arimaa then computers will easily dominate. Perhaps the problem with these games is that there is not enough theory yet to develop a suitable algorithm. What is been forgotten here are the brilliant programmers who contributed to the current chess machines we see now. So no breakthrough in computer technology is needed at all, just more human minds translating the strategy/tactics needed to win into programming. Pattern recognition is not a problem for computers but this is a vague notion at best. Humans tend to go with a 'feel' for something. This 'feel' cannot be translated logically. The computer needs something more tangible. I think winning patterns can be programmed into Go, but the Masters must be willing to GIVE UP THEIR secrets! Exactly how much literature is out there for Go and especially Arimaa ? I think Go is the next challenge of computer programmers. Arimaa is simply not popular enough to be taken seriously by computer programmers.
Basically, the computer in front of you is a complex adding machine. It doesn't think nor recognize patterns the way a human does. Yes, we've made the adding machines complex enough that they can do things like play music and movies, and even play Chess well. But we haven't been able to have it so computers can, for example, translate from one language to another without the translation being so bad it's just about not readable.
Nor have we been able to get a computer to play a game with a high branching factor, like Go or Arimaa well. Computers play Chess very differently from humans; they just look at all of the possible moves, using 'alpha-beta' pruning to determine which moves are and are not looking at. They don't recognize patterns; they just see possible future moves and how much material they have.
A computer needs to evaluate millions of possible positions to play as well as a human who only looks at dozens of possible positions. Computers aren't able to really see a given position to evaluate how good it is; they only play as well as we do because they basically brute force through just about every possible chess move so many moves down.
Games like Go and Arimaa are good because brute force just doesn't work with these games. In order to have a computer play these games well, we will have to make a true AI breakthrough. Which will probably have consequences far beyond just having a computer playing some abstract game really well.
Actually games can make good contribution to computer science in pushing it to create a good theory of practical complexity. Currently there's only a good theory of worst case complexity and a passable theory of complexity of approximating within certain percent of best or worst case. But practical complexity has to be estimated without really necessarily knowing the worst case. The practical significance is with such a theory computers can have a better feel for strategy, instead of either only planning for the worst case, or using more or less blind (actually guided) search.
P.S. - Arimaa has a nice web-site devoted to it (even has an animated tutorial with music); and has World Championships for humans, and another World Championship for computers (thus encouraging programmers to create a winner). I can see where this game would be difficult to program, after all, do the human programmers even know what is the best strategy/tactic in a given position?
Anyway, time is on the side of the computers.
So, lots of pieces, large board to give those subtle gradations of positional value, multi-move turns with a mechanism to spread each turn's moves across the gameboard, scalable... what games are examples of this [besides Gas Hogs ;-) ]? Go is, I believe, still an example of a game that humans play better than computers. The weak spot in the idea that computers can play any game better than humans, with the right algorithm, is the algorithm. I would guess that idea is not proved, and suspect it may not be able to be proved [Godel is the mystic name I invoke here, for the obvious reason]. But that is speculation. What is, to the best of my knowledge, true, is that computers don't play all games equally well now. [Otherwise, we wouldn't need generals or CEOs, except on gameboards.] So, by providing 'difficult' games for computers, we may encourage better AIs in the future. :-D 'Gaming has always driven computer design' is to a considerable extent, a truism. So let's maybe help it along in a slightly different way, by providing games that need new algorithms. [Edit] I see Ji is ahead of me. One point he made I only thought of is the amount of time a computer needs to come up with a good move. I lose to Zillions because I tend to attempt to match its speed. While I am beginning to look at 2 possible initial moves, it's already 11 plies down. I'd like a game where a 2 minutes per move time limit was an equal handicap to both me and the computer.
Given enough time a well programmed computer will at least draw humans on any finite game, that I certainly agree, but there still much utility in creating games that at this point computers are woefully bad at. At this moment at least computers think in many ways very differently than humans. By throwing widely varying situations for computers to master, we develop a fuller theory of cognition. I think humans should always strive to beat the computer so as to improve both us and the computer. It took over a decade to solve checkers completely, humans are not over at abstract games yet.
Assuming a computer is in good working order and that it has a program for the game in question, then if it cannot play the game well, it is only because it is lacking something in its code. With refined codes near optimization - the programs will defeat the humans. If a human cannot accept that, then he (or she) can simply play other humans to have a fair brain-to-brain playing field.
25 comments displayed
Permalink to the exact comments currently displayed.