Option Chess

[Simon Valiquette]

Re: Option Chess

you are proposing that Option Chess will produce many more duplicate positions in the engine search than standard chess. I don't follow your logic on this, but this would be the most interesting one to have either proven or disproven.

I have no doubt that there will be plenty of duplicates. As a simple example, a position reached after 3 turns could also be reached after 2 turns if both players used an option. But both players don't need to have used their option at the same turn, so you already got at least 5 ways to reach the exact same position with White to play. If we do the same after 5 turns, it gets complicated just to calculate the number of ways the same position could be reached.

The only corner case you need to handle is the possibility of a prise en passant from Black last move, but if he didn't used an option or nobody can take en passant the piece that just moved twice, then no problem.

To be precise, there is 4 identical positions, and the fifth differ only by the fact that both player still have one more option each, but unless there is very few options left, I guess it is enought to simply give those options a value (that could be the result of a sophisticated calculation if needed).

What needs more investigation, is if the search tree will grow much faster than I can simplify it (compared to regular chess). Because the board have a finite size and so a finite number of positions, it puts some limits on how fast the search tree can grow completely new positions and I suspect that once enough ply will have been calculated, the increased complexity to calculate more ply will be of the same magnitude than with regular chess (slower, but not much slower). Of course, if proper optimization is used.

In the second paragraph, I should note that one would not use a double move to move a Bishop from a1 to h8, because that means the same move could be done in a single move. Maybe it's a typo and you meant to type different squares, I'm not sure.

Of course, I would expect someone using a double move on a single piece to change direction, but computers got to handle even stupid moves. It seems unlikely, but maybe there is some special situation where wasting an option could be a good idea. Anyway, the computer got to handle every cases, including the situation where it would take en passant such bishop by accident.

Anyway, again, I hope that the publication of the Option chess article will encourage engine authors to attempt to create Option Chess engines. Right now there may not be much incentive and I would like to offer financial incentive, but that will have to wait for at least a few months. Still, you could start work on it right now.

A very long time ago as a teenager, I built a small chess engine in Pascal. It was incredibly bad, but at least it played legal moves. Of course I didn't had the knowledge to build a decent algorithm, but it gave me an insight of how difficult it is.

It is quite easy to have a high level conception of the general algorithm, but actually implementing it properly is very difficult because of all of the optimization tricks and corner cases to handle. To do it properly, I would probably need to work with a mathematician. Optimizing it for Option chess seems very feasible, but it probably won't be easy. I don't plan to put the time for coding this, but adapting Crafty or Toga for Option Chess could be a great project for a group of students.

Simon

[Kevin Pacey]

Re: Re : Re: Option Chess

Kevin, I'm not clear on what you think the issues are here. However, since options remaining are an intrinsic characteristic of an Option Chess position, a position will not be regarded as a repeat if the remaining options are different for either player. Therefore for a 3-fold repetition draw to be claimed, each repeat of the position must have the exact same number of options remaining for each player.

Also, Louis asked me a question about castling: is the castling move regarded in Option chess as a King move or a Rook move? Castling as part of a double move might allow either the King or the Rook to be moved in the SECOND move of the double move, therefore could either of them capture material?

The answer I gave Louis for this excellent question is that castling in Option Chess is both a King and a Rook move, therefore if EITHER piece is moved again as the second move of a double ove, that piece (King or castled Rook) cannot capture material on that second move. But the Rook could give check.

Your answer in boldface (unfortunately, not given in the article) deals with the issue I was concerned with, except perhaps not completely smoothly in practice, since the players need to remember for at any given point of the game how many options each player had at the time. One way to do that would be to record the number of options as a number in brackets after each move when recording their moves (tough for blitz, which your article did claim Option Chess was not meant for), but all the same it can be more of a challenge than ever for either player to realize that a position has been repeated three times.

There may be other unforseen consequences caused by Option Chess's freshly created rules. It would have been best to playtest this variant (or Tandem Pawn Chess, for that matter) extensively before publishing, for at least this reason (I can see how opponents or enough free time might have been lacking, however).

[Simon Valiquette]

Re: Option Chess rules

[...] since the players need to remember for at any given point of the game how many options each player had at the time. One way to do that would be to record the number of options as a number in brackets after each move when recording their moves (tough for blitz, which your article did claim Option Chess was not meant for), but all the same it can be more of a challenge than ever for either player to realize that a position has been repeated three times.

I don't think you need to remember the number of options left, tough it could be added to the score sheet. All you need is to spot the last double-move or the last pawn move, knowing that a triple repetition cannot appears before that move (since consuming an option, pushing a pawn or taking a piece change the position in an irreversible way). It is useless to check for a triple-repetition before that, because it will be too late to claim the draw anyway.

Note that the last move of a double-move can be part of a future triple-repetition, but the first move of a double-move cannot create a third repetition, because the option have to be consumed before the move was played. In other words, a double move can never be part of the second or last repetition.

There may be other unforseen consequences caused by Option Chess's freshly created rules.

I've showed a few examples of new spectacular stalemate ideas, but I don't see any problem with the application of the triple-repetition rule (it won't be easier or harder to apply than in regular chess).

[Paul Bonham]

Re: Option Chess

That just gave me a new idea. If it is my turn to move, and that I place a green token close to the clock to show that I plan to do a double move, this decision is irreversible once I have released the token, right? So let say that in the following position, Black just played Be3 déc.



Of course, I put my tocken close to the clock and play Ng1 stalemate! Think about it: I must play my second move since this decision is irreversible, but there is no legal move left while I am not in check anymore!

By the way, I am fine with this. But the following is more tricky.



Again, I use my token and then I play Ka1 Draw! It is a stalemate because a piece that moved twice cannot make a capture, so there is no legal move left! There is plenty of other stalemates ideas based on the use of an option. You may well decide to leave it as it, but perhaps you should at least clarify if it is a valid stalemate idea.


Simon

Very nice, Simon, an excellent discovery. So the question becomes, should an option be usable to achieve a stalemate as you have shown, OR should there be an exception to the rule that an option once indicated using a token is irreversible, and the exception would be that in cases where the option creates a stalemate as you've shown, the option should be considered reversible and only the single move can be made?

I think I lean towards the second choice. The thing is this: an option is a choice, and there is the question: did the player play the option not realizing it would mean a stalemate? There's no way a TD could determine whether this was a mistake or not, and I'd hate to even put that question to them. So we'd have to consider it a deliberate action. So now the question is, would this stalemating allowance create lots of opportunities for stalemate against huge material deficit? If so, this would hurt the game. In regular chess, stalemate is hard to come by and hardly ever happens in real practice. But in Option Chess, it is hard to say whether these kinds of stalemates would be commonplace or not with options remaining.

So I'm going to say to prevent that possibility, we disallow such a stalemate, and force the player to retract his / her option token and play the single move, which would result in a checkmate. Later, if several thousands of games show that stalemate is still just as difficult to achieve with double moves as it is in regular chess, the rule might be changed to allow deliberately using an option to stalemate.

[Paul Bonham]

Re: Re : Re: Option Chess

There may be other unforseen consequences caused by Option Chess's freshly created rules. It would have been best to playtest this variant (or Tandem Pawn Chess, for that matter) extensively before publishing, for at least this reason (I can see how opponents or enough free time might have been lacking, however).

I see that Simon has answered the question of knowing how to look for 3-time repetitions, and he thought it out very well. And I just don't see Option Chess being viable for any kind of play that involves making moves in just a few seconds, so I'm not concerned about those issues.

But for your other point, I think you misconstrue the intent of the Chessbase articles. If Frederic Friedel had been at all concerned that these articles should only describe variants that were virtually 'bulletproof', he would never have published either my article nor Ken Regan's. Instead I think the intent was to put these variants out there as ideas worthy of research and development. Unforseen consequences will either kill the idea totally or cause more thought to be put into it. Perhaps in a few months or a year, when further games and ideas have been developed for both Tandem Pawn Chess and Option Chess, it will be allowed to publish follow-up articles on Chessbase that will link back to the original. I hope this becomes the case, as at least for Option Chess I do plan on allowing for this variant to be played online. Already both Louis Morin and Simon Valiquette have contributed very good questions that caused me to add extensions to the rules, and I do also appreciate all the points you are bringing up. In fact, maybe I should ask if you are interested in doing just what Louis and I are doing: playing two games by email, with no time limit between moves, and you playing White in one game and I playing White in the other game? This would become part of the history of Option Chess.

[Kevin Pacey]

Re: Re : Re: Option Chess

Sorry Paul, but I don't even like playing chess on a server for even a fraction of a day. I tried playing chess for days by a message board a few times, and it took up too much time & energy for my own good, in order to play well (and that was with computer-assistance allowed). In any case, my chess skill level, if it at all carried over to Option Chess, might work more heavily in my favour than you'd be up against in a possibly closer Option Chess match with Louis.

[Paul Bonham]

Re: Re : Re: Option Chess

Sorry Paul, but I don't even like playing chess on a server for even a fraction of a day. I tried playing chess for days by a message board a few times, and it took up too much time & energy for my own good, in order to play well (and that was with computer-assistance allowed). In any case, my chess skill level, if it at all carried over to Option Chess, might work more heavily in my favour than you'd be up against in a possibly closer Option Chess match with Louis.

I would like you to know that your much better skill level than mine in regular chess would be most welcome to me in a battle of Option Chess. For me, it isn't about who wins or loses.

I know Option Chess isn't your variant and that may or may not be part of why you don't want to put any effort into it, but if you could see its potential as an antidote to the computer cheating problem in regular chess and as an exciting, tactical yet also strategic alternative to regular chess, then I think it should greatly interest you to play a few games by email. But I doubt I'll convince you, despite your obvious interest in finding the "next" chess. You are on your own path, and I wish you good fortune.

[Mathieu Cloutier]

Re: Re : Re: Option Chess

This would become part of the history of Option Chess.

Let's no get ahead of ourselves here... ;)

[Paul Bonham]

Re: Re : Re: Option Chess

Let's no get ahead of ourselves here... ;)

You can have that as your credo if you want, Mathieu. You'll make a great CFC governor someday, seeing as that is their virtual motto.

Getting ahead of yourself is the best way to make everyone else catch up.

[Kevin Pacey]

Re: Re : Re: Option Chess

Fwiw, here's a link to a wikipedia discussion of computer Arimaa, which may give a good idea of how computer-resistant it is (at least compared to chess):

[edit: personally, I would be tempted to suggest creating something like a 10x10 Arimaa variant later on, if normal (8x8) Arimaa doesn't prove to be computer-resistant enough in the long run.]

http://en.wikipedia.org/wiki/Computer_Arimaa

[edit: Also, here's a website devoted to Arimaa:

http://arimaa.com/arimaa/

]

[Paul Bonham]

Re: Re : Re: Option Chess

Fwiw, here's a link to a wikipedia discussion of computer Arimaa, which may give a good idea of how computer-resistant it is (at least compared to chess):

[edit: personally, I would be tempted to suggest creating something like a 10x10 Arimaa variant later on, if normal (8x8) Arimaa doesn't prove to be computer-resistant enough in the long run.]

http://en.wikipedia.org/wiki/Computer_Arimaa

[edit: Also, here's a website devoted to Arimaa:

http://arimaa.com/arimaa/

]

Jerry Seinfeld on the game of Arimaa:

"Ok, so they used an animal theme here, and they made the elephant the most powerful animal. So far so good. But then you got what... the CAMEL? I mean, really? You want a CAMEL as your second most powerful animal? Is this some post-apocalyptic world where lions and tigers and grizzly bears and wolverines have been killed off, and the most fearsome animal you got left next to the elephant is the camel? I guess we need to know, then, is it a one-hump or two-hump camel? 'Cause, you know... I think those 2-humped ones are MUCH more fearsome looking.

In this post-apocalyptic world, I wonder what car models are named... the Ford Dromedary? They must have even changed 'horsepower' into 'camelpower'.

And then we go next to the horsey. Why am I calling it a 'horsey', you may ask. Well, you know, people who don't play chess, the one chess piece they identify immediately is what chess players call the 'knight', but the non chess players, they see that it's the head of a horse and they almost unanimously call it the 'horsey' piece. 'Move your horsey!' they'll shout at their 5-year-old budding chess genius. So I'm just going with what will become the name most people will give this piece.

Then we got dogs and cats, and I gotta tell you, this must have been a tough one. First of all, getting dogs and cats to work together on the same team, that's miraculous in itself. But the dogs are happy because they got made stronger than the cats. Of course, the game designer must have realized, that's really going to offend a lot of cat lovers. I somehow just don't see many women playing this game. "They put those beastly dogs ahead of my Fluffy!" Of course, the men will love it, and in fact, they'll spend most of the game chasing down those cats with their dogs. "Rip that varmint to pieces, Bowser!" I just don't see it, you know... the cats on the winning side are going to revolt and attack their own dogs. You just can't keep those two species together on anything for long.

And then finally they give us rabbits. Eight of them to start. But miraculously, the rabbits do NOT multiply through the entire game. Now, come on. How realistic is that? Sixteen sterile rabbits? So we had a nuclear war, and that killed off all the lions and tigers and grizzly bears and wolverines, and it sterilized all the rabbits who somehow still exist as a species, and it made CAMELS into fearsome killers that only an elephant can handle, and it made dogs and cats able to work together towards a common goal.

I don't know about you, but considering all this, I actually wish they'd bring Desperate Housewives back on TV. Definitely more believable than this."

[Paul Bonham]

Re: Re : Option Chess

Simon Valiquette raised the possibility of a move of Bishop from a1 to h8 to be considered a double move if an option token was played. This is a straight-line move of a Bishop that could be done with a single move (the intervening squares must all be empty, because double-moving a single piece cannot capture material). Simon mentioned that if any of the squares the Bishop traversed in this move were attackable by opposing piece(s), the opponent could then capture the Bishop 'en passant'.

The rule, which shall become the 11th restriction on double moves, shall be: No move that could be accomplished as a single move in standard chess shall be considered a double move in Option Chess. This also means one cannot move a pawn 2 squares from its starting square after playing an option token and claim that the pawn moved one square then another square to constitute the double move. The entire 2-square move shall instead be considered a single move.

Also, with the King this means the player cannot play a token and then play, for example Ka1-a2,Ka2-b2. The King ends up just 1 square away from where it started, which can be accomplished with a single move. Since neither move can be a capture, the King moving to b2 via a2 is no different to the position than if the King moved straight to b2 -- except that a double move token was used. But the rule as written above forbids this. A double move of the King must result in the King being 2 squares away from where it started.

[Simon Valiquette]

Re: Option Chess

Simon Valiquette raised the possibility of a move of Bishop from a1 to h8 to be considered a double move if an option token was played. This is a straight-line move of a Bishop that could be done with a single move (the intervening squares must all be empty, because double-moving a single piece cannot capture material). Simon mentioned that if any of the squares the Bishop traversed in this move were attackable by opposing piece(s), the opponent could then capture the Bishop 'en passant'.

This not exactly what I meant. According to the previous rules, when consuming an option you MUST say which intermediate square you used. Otherwise, for a Knight's double move there would often be 2 squares where someone could take en passant.

What I said, was that the computer would have to calculate all of those possibilities, including the fact that the Knight will often have two roads to the same square, because it is harder for the computer to understand if both squares are exactly the same or not (code will have to be written to detect when it can be optimized). Changing this rule will just simplify the computer's job.

The rule, which shall become the 11th restriction on double moves, shall be: No move that could be accomplished as a single move in standard chess shall be considered a double move in Option Chess. This also means one cannot move a pawn 2 squares from its starting square after playing an option token and claim that the pawn moved one square then another square to constitute the double move. The entire 2-square move shall instead be considered a single move.

Again, adding this rule will only help the computer to make better optimization, since in practice it will almost never change anything for a human player.

Also, with the King this means the player cannot play a token and then play, for example Ka1-a2,Ka2-b2. The King ends up just 1 square away from where it started, which can be accomplished with a single move. Since neither move can be a capture, the King moving to b2 via a2 is no different to the position than if the King moved straight to b2 -- except that a double move token was used. But the rule as written above forbids this. A double move of the King must result in the King being 2 squares away from where it started.

My advise is to keep as few rules as possible, unless it simplify existing rules or solve a real problem. Just like I believe you should allow using an option to get a stalemate unless we got evidence that I is a real problem. That possibility of a save probably add to the fun factor, and if the opponent have a major material advantage, he should just learn to spot those traps and be careful, and he should then win anyway except for few exceptional positions (just like a lonely rook can sometime draw against a queen).

I would also get ride of the requirement that you mus stay in front of the board when it is your opponent turn to move. If you want something to prevent cheating, just ask people to write a special mark on their score sheet (perhaps an «@») before playing an option, so it would be harder to silently change your mind.

Ideally, people should be able to learn and remember every new rules in five minutes, and it seems very feasible to do that with the original concept of Option chess.

Simon

[Paul Bonham]

Re: Option Chess

Originally Posted by Paul Bonham
Simon Valiquette raised the possibility of a move of Bishop from a1 to h8 to be considered a double move if an option token was played. This is a straight-line move of a Bishop that could be done with a single move (the intervening squares must all be empty, because double-moving a single piece cannot capture material). Simon mentioned that if any of the squares the Bishop traversed in this move were attackable by opposing piece(s), the opponent could then capture the Bishop 'en passant'.

This not exactly what I meant. According to the previous rules, when consuming an option you MUST say which intermediate square you used. Otherwise, for a Knight's double move there would often be 2 squares where someone could take en passant.

What I said, was that the computer would have to calculate all of those possibilities, including the fact that the Knight will often have two roads to the same square, because it is harder for the computer to understand if both squares are exactly the same or not (code will have to be written to detect when it can be optimized). Changing this rule will just simplify the computer's job.

That's another good catch on your part. The rules I posted on Chessbase didn't say anything about how a player indicates which intermediate square a piece uses when it is double-moved. There was no mention that a player must say which intermediate square is used. What I imagined was that a player would first move the piece to the intermediate square, hold it there for a brief second, then move it to its final destination square. And of course, in notating the move, the intermediate square would be mentioned. For example, a knight double-move would be recorded as:

9. Nf3-d4,Nd4-f5

So the player making the move would first move the Knight to d4, hold it for a second, then move it to f5. This is going to have to be written up in the rules as a requirement (i.e. the Knight can't be moved straight from f3 to f5), as the Knight could have gone to f5 via h4. This will have to be done in the presence of the opponent to make him / her aware of the path of the double-move. Like I said, great catch, Simon.

Regarding computer optimization, I actually welcome any and all computer engine optimizations. My hunch is that they won't make a dent in the exponential explosion of the search tree. But I can see that anything is possible, and so I say, bring it on. This is the scientific method. I want engines to use as much optimization as they can, and if it turns out that optimization such as you describe, Simon, will enable engine authors to create once again engines that dominate humans, well, then Option Chess will be a failure in terms of preventing computer cheating. I'm willing to take that risk. I'm reasonably certain it's not going to work out for the engine authors barring some unforseen quantum jump in technology.

The main reason for that is that I believe the optimizations you speak of are miniscule in comparison to the enormous jump in size of the search tree. I've done some computer chess engine coding myself, and I included position hashtables, and enabling that option made only a tiny improvement in computation time. Who knows, maybe I didn't do it optimally, but as I said, bring it on. This is a challenge to all chess computer engine authors!




Originally Posted by Paul Bonham
The rule, which shall become the 11th restriction on double moves, shall be: No move that could be accomplished as a single move in standard chess shall be considered a double move in Option Chess. This also means one cannot move a pawn 2 squares from its starting square after playing an option token and claim that the pawn moved one square then another square to constitute the double move. The entire 2-square move shall instead be considered a single move.

Again, adding this rule will only help the computer to make better optimization, since in practice it will almost never change anything for a human player.

Well, it will change one thing, or maybe I should say, there is a problem it will solve. The problem is that a player may play a double move token, committing to a double move, and then suddenly realize s/he has no good double move, all double moves lead to disaster (i.e. s/he miscalculated in placing the token). So s/he will try and weasel out of a true double move by moving a pawn 2 squares from its starting square and claiming that as a double move, or moving the King one square and saying it moved 2 squares to get to that square. We have to explicitly disallow that possibility. No one should be able to weasel their way out of making a true double move once they have committed to it. The whole purpose of the token is to COMMIT to the double move before actually making it.



Originally Posted by Paul Bonham
Also, with the King this means the player cannot play a token and then play, for example Ka1-a2,Ka2-b2. The King ends up just 1 square away from where it started, which can be accomplished with a single move. Since neither move can be a capture, the King moving to b2 via a2 is no different to the position than if the King moved straight to b2 -- except that a double move token was used. But the rule as written above forbids this. A double move of the King must result in the King being 2 squares away from where it started.

My advise is to keep as few rules as possible, unless it simplify existing rules or solve a real problem. Just like I believe you should allow using an option to get a stalemate unless we got evidence that I is a real problem. That possibility of a save probably add to the fun factor, and if the opponent have a major material advantage, he should just learn to spot those traps and be careful, and he should then win anyway except for few exceptional positions (just like a lonely rook can sometime draw against a queen).

I would also get ride of the requirement that you mus stay in front of the board when it is your opponent turn to move. If you want something to prevent cheating, just ask people to write a special mark on their score sheet (perhaps an «@») before playing an option, so it would be harder to silently change your mind.

Ideally, people should be able to learn and remember every new rules in five minutes, and it seems very feasible to do that with the original concept of Option chess.

I definitely would like people to learn Option Chess in 5 minutes. Absolutely! But just read all the official rules of chess to see how difficult that is. Imagine you had to learn standard chess from scratch just by reading the official rules. No way that's going to happen in 5 minutes for 99% of the population.

Simon, you have contributed greatly to this thread and to Option Chess. You will receive credit for this in any future publication(s) on Option Chess. I would even like to play you some games by email if you are interested (I have started 2 email games against Louis Morin, and we are currently at move 6 in each game).

About the rule for staying close to the board: once your opponent has placed a token committing to a double move, you must be there to see that the player makes a LEGAL double move. The sequence of the double move is critical to this. If you are away from the board, it will be harder for you to determine if a double move the opponent played was really legal, and as pointed out above, what intermediate square a double-move of one piece used in case you can capture en passant. You must be there to see it when it is played.

For those who see this as a problem, I say stop being a wuss. Maybe there would need to be a time-out policy where both players can take a bathroom break, and maybe it would have to be more generous for women players or people with special needs, but let's not overdo it. You are there to play Option Chess. In tennis, if you suffer an injury, you are only allowed so much time to recover, otherwise you forfeit. Reality bites.

On the stalemate issue: the possibility of a save can only add to the fun factor for the player who is behind. It doesn't make the game more fun for the player who won material and got into an endgame ahead and now is being asked to avoid further tricks and traps. Although I like tricks and traps, again, it's a case of all good things in moderation. Too many tricks and traps kills the fun for the majority, when you consider the spectators who have to sit through an interminable endgame.

But I will agree on this: Option Chess adds a new dimension to endgames. That dimension is the number of double-move tokens (i.e. options) remaining for each player. If a player is down material and is UP by one or more options, how do we determine whether that player should play on or resign? Only time and many, many games will help determine this, so for now, I would say if you are down material but up in options, keep playing. But I still want to remove (for now) using an option to claim a stalemate. That is just too cheap for most people to be a fan of it, I think.

[Mathieu Cloutier]

Re: Option Chess

My advise is to keep as few rules as possible, unless it simplify existing rules or solve a real problem.

That right there. Keep things simple.

I'm not against chess variants, but the rules of chess are complicated enough as it is. If we have to add a dozen of rules to play a variant, I doubt it will catch on. You won't attract people by telling them: 'Hey buddy, you know about chess? Well I have a much more complicated game for you!'

The only chess variant that really caught on is Fischer random. And it is also very easily explained, even to a beginner or to somebody who knows very little about chess.

I tried looking at a position and calculate as if it's option chess and that's just too much for me. Way too much rules added on an already complex game. Very impractical for live games with reasonable time controls, too.

And if only all these rules made it simpler to beat computers. So far, it seems that almost everbody agrees that this game, although maybe fun to play for some, is clearly a step in the wrong direction in terms of computer resistance.

I've said it before: you are trying to achieve two goals with this game (more tactical/exciting chess AND computer resistant).

I don't want to be all negative, but I would advise as Simon: keep it simple. For example, just removing all the 'en passant' stuff would make this much simpler without changing the nature of the game.