Human Chess is a chess variant where playing the top engine move is forbidden
humanchess.abcd.party
humanchess.abcd.party
My game quality is measured in how many times I say "fuck!" right after moving a piece. A very good game for me is about a two-fuck game.
Most of them it's more like four or five "oh my god I hope they don't see that thing I spotted the second I took my hand off" moments—and that's just the ones I notice before they're exploited. I'm sure I make tons of moves that anyone half-decent would call blunders but that simply go unnoticed by both players at the board.
I'm so very bad at spotting diagonal attacks, especially. Anyone who can open up their bishops then play for time will eventually see me put my queen in some dumbshit situation that lets them take it free or cheap in a single move, for instance, not even any multi-move planning required.
You basically want to "blunder" into giving your opponent long chains of captures while avoiding any positions that allow your opponent to hang a piece.
While I enjoy the conversation ideas like these create, I'm often left wondering why Fischer random isn't more popular.
Edit: I didn't read the fine print. First moves are exempt.
However, the thought you had, and similar ones, are very much the intentional side-effect of the rules. The only way to win the game, as stated, is forcing your opponent to make the top engine move. Or, of course, correctly claim that your opponent made such a move (even though it wasn't forced). Or, having your opponent make the incorrect claim about your move.
So, it isn't necessarily "playing good chess". Though, I must say, I'm not qualified to have any good idea of what it would mean to be good at this game. It definitely helps to be good at chess, and have a good command of what are the correct engine moves. Especially since you lose if you incorrectly claim a position and opponent move was "the top engine line".
I suppose most would reduce this to leaving the opponent to only one legal move. In which case, the problem is is trivial. But, after move 2? You need to know most opening lines, and probably play intentionally bad in many situations.
Hm, this is cooler the more I think about it.
Imagine intentionally setting up material sacrifice with a resulting choice of multiple moves for the opponent to capitalize. If you can correctly evaluate the best computer move, you have a strong advantage. If it is not obvious, then the opponent might not dare to gamble the challenge.
Has Hikaru tried this in one of his streams? I'm certain he would have a blast.
The only thing about this that doesn't "spark joy" are the ambiguous practical implementations.
- Which engine? This matters a lot.
- How do you define the computational cut-off? CPU-minutes? Move depth? Etc. Not necessarily a simple problem.
- The rule "When multiple moves have the top score, they are all top moves". Needs a specific score evaluation delta for grouping "top moves".
All of these could rather simply be resolved if it isn't very important... might even add some uncertainty to it, for fun. Like, say: 1. Stockfish 15. 2. Allow the computer whatever resources it has available, 1 minute, and play some drum roll sample. 3. Pawn-evlauation of 0.05.
Human players are dealing with both human limitations and human limitations which really changes the game. So a grandmaster can for example benefit from playing a slightly weaker but less well known opening that they have recently studied in depth with the assumption that their opponent hasn’t done the same.
A further variant for fairness purposes: let each player bring their preferred chess engine, instead of arguing about the choice of only one; and have each player run both engines for mutual anti-cheating verification. Then either the two engines agree on the best move (likely case if they are both strong) or all moves that either engine considers better than the other engine's best move (at least 2, usually not too many) can be interdicted.
Especially in this game, the humans will be trying to play badly.
Hikaru Nakamura is a top GM that does commentary streams — and frequently comments things like “but what do I know about chess?” when responding to AI suggestions. It’ll suggest a weird move that seems to leave a mess on the board while insisting that everything is great. And obviously if you actually tried to play against Stockfish from that position, it would crush you.
Looks like 400 ELO, but hey — what do we know about chess?
There's been times when the chess.com calls my move a mistake, and suggests something nonsense that leads to a guaranteed mate in 15, when I got a (non-guaranteed but real) mate in 5 from my move, because I know I'm playing against a human. The engine move is more guaranteed, but very illogical unless your brain can do the equivalent of the 30 turns of minmax.
A minimally modified engine lookup wins here.
There's a whole bunch of openings that ensure it for white, this game is rigged even more than playing the best move, even if you do enforce a random opening.
So essentially this converts most mates in two into mates in one, but some become ties by repetition.
It is an interesting variation on chess given the current state of tournament play, yet it isn't really a solution to the cheating problem since it is effectively a new game with a new end-state. But you are probably right about there needing to be some sort of threshold. While there the rules of the variation says that any move with equal scores is considered equivalent, I would imagine the players would need a very intimate knowledge of how the engine scores moves in certain scenarios.
My immediate first reaction was also that it would be interesting to have a variant that is the same except you are allowed to checkmate, except then I realized the recursive nature of how board positions are evaluated makes that problematic. For instance, if there's a mate in 2, the first move of the mate in two is now certainly the "best move". Creeping up on a checkmate without ever making the "best move" until the very last one might actually be harder than the win condition based on strangling the opponent described in the current rules.
It just seems like you're changing the objective of the game entirely to the point where it's only slightly related to chess.
Therefor, such an engine can only hang in computation - being unable to produce a top move because if it were to make a suggestion then the actual best move changes to avoid it. Since the engine is unable to produce a move - there is no top engine move which makes every move legal.
A normal game of chess is played while the engine locks up on the sideline
So you end up in this scenario where both players are taking one another's pieces while leaving their queens en prise the whole time. Is it a draw, or is there some clever way to break this loop?
For instance after 1.e3 e6 2.Qh5, White threatens Qxf7+ which would force black to play the top engine move. Then 2...g6 3.Qxg6 is one idea -- but there are two recaptures, fxg6 and hxg6, and only one of them can be the top engine move (hxg6, I'm guessing). So 3...fxg6 probably refutes this idea. But are you sure enough as white to try to claim a win if black goes hxg6?
And after say 1.e3 e6 2.Qh5 g6 3.Qxh7 (avoiding that line and going for material), not only does black not have to care about their rook (white can't take it, it would be the best move), black actually has 3...Qh4 winning -- he threatens 4...Qxf2+, white can't play 4.Qxh4 as that's the best engine move, and white's queen is threatened twice, so black will be able to take it -- provided he checked this line before the game to know which piece to take with.
Edit: it doesn't actually win, white has 4.Qf5 to defend f2... what a strange game.
Edit 2: once a piece is _en prise_ somewhere, the game can otherwise become somewhat normal as taking it would be the best move and so would moving it to a safe spot - so other moves can be played as usual. But would they be good?
We can argue if a forced move ends the game, or just allows it.
This would (more) move the game forwards in the basic historical rule-set.
- giz
Second, if you ever hang your queen two ways at once - one of them could be a less optimal take (-5 is not as good as -8)
Third, whoever is a move ahead in a race of taking pieces will be the first to run out of weak pieces to take. Their available move pool is shrinking faster. Not sure how it would play out, black would need to cater to it by removing defenders and hanging pieces of their own, etc.
That said, first move advantage does seem strong still due to how forcing a queen can be. An example would be 1.e4..e5 2.Qh5..d5 3.Qxf7#
It is an interesting theoretical question whether we can have Aleph Zero Human Chess where Human(Human(...(Chess))) is applied infinitely, approaching Aleph Zero trainings of AlphaZero, or we get a redundant variant after some application where further application of Human() no longer produces a new variant.
I think it’s possible that at least 1 position (and probably a lot more) will have more than 1 optimal solution. If the engine is not fully deterministic, then it’s possible instead of a normal cycle, there are a fixed set of strategies at each n that form a cycle, but no single sequence that repeats.
The other thing that would compound this is that no current chess engine solves the game fully. There would be even more positions that have multiple “optimal” solutions if the engine only looks ahead to bounded x.
> The other thing that would compound this is that no current chess engine solves the game fully. There would be even more positions that have multiple “optimal” solutions if the engine only looks ahead to bounded x.
I'm not sure this is an obstacle; we're explicitly excluding the computer's preferred move, rather than the necessarily optimal move, after all. You could easily play human chess with the engine from (e.g.) Battle Chess, which honestly is sort of an interesting idea in its own right.
In fact, the opponent cannot play the best move to escape a mate, so a bunch of the games would become forced surrenders.
Moreover if you hang your queen in more than one way, your opponent can still take it in whichever way the computer evaluates as worse. Which is often easy to guess. The weird part of this will come from the fact that accuracy of engines diverges very rapidly off of the critical path. Once you're down a queen, you're basically free to play however you like.
There are a lot of dynamics here.
Next: AI that can play Human Chess.
After: Human^2 Chess: you can't play what the AI above would play.
etc
I wonder if this creates distinctly new games at each level, or if it's just nonsense one level down.
https://www.chess.com/blog/AcceleratedPog/bongcloud-opening-...
(I believe a chess engine could play human^2 chess exactly like it plays regular chess. A human couldn't because a human doesn't know what moves the chess engine would pick.)
AI can't win at human chess, because any move that it attempts to make is by definition the top move choice of an engine, and so causes immediate defeat.
In Veto Chess you get one chance per game to veto your opponent's last move, and force them to make a different one.
This shares with Human Chess the property that you can win by checking the king such that the response is forced.
It may also serve as a handicap system in games between players of widely different strength, where only the weaker player gets the veto.
(“Too Awesome to Use” on TV Tropes. Link omitted - you’re welcome).
But then, I’m a terrible chess player.
It would be pretty neat between players of similar skill level though, then I could see the hoarding taking place.
As the saying goes, "the threat is stronger than the execution".
<Zach Braff voice> How did we get here?
See for example several books by Elwyn Berlekamp.
One outcome of this work was Berlekamp (IIRC) solving a small class of endgame problem that has eluded professional (full-time) go players for literally hundreds of years.
This will lead to extremely cagey games where no one ever dares make the game sharp and imbalanced.
When multiple moves have the top score, they are all top moves,> When multiple moves have the top score, they are all top moves, even if visual markers (like move arrows) suggest the engine prefers one over the other.
Since all moves that checkmate the opponent will have the same score (M1 or -M1) they'll all be illegal.
In this variant the attacking piece will effectively be protected by how bad its move was. Creates some interesting incentives - the only way to checkmate is a move that is normally not optimal and has only one way out of check.
Two players compete to win a chess game, where you only win if you work tirelessly to play perfect chess moves on every turn (as determined by benevolent artificial superintelligence). The loser is tortured in a virtual reality simulation.
So both players need to stop playing immediately after the first move and start working towards making strong AI happen, or they will be VR tortured forever.
My bet (uninformed, very novice at chess) is that it's likely there's guaranteed setups that would always catch an engine.
You need a different engine that's focussed on not only avoiding conventionally top moves itself, but also forcing its opponent into them.
It's interesting that this brings another degree of indirection to victory conditions.
If you never played chess before, you'd assume the goal is to take the opponent's king. But as we know, making a move that would allow your opponent to take the king is forbidden, so the goal of normal chess is force your opponent into a position where you could take the king next turn (checkmate).
This variant takes this another step further: Now any move which could result in checkmate (or check with only one exit) is forbidden, and the goal is to force the opponent in a position where any next move would result in checkmate or check.
Just make the move disallowed. You'll need the computer to be paying attention at all times, but nonetheless it would be more enjoyable to play.
Anytime you can checkmate is going to be the computer’s recommended move.
But I assume the top-level commenter meant make winning exactly like conventional chess - just neither party can use the top moves to get there. You could even start from move 1 instead of 2 too, take the best openings off the table.
How? Top engine move changes with evaluation time. Longer the wait, better the move.
I think that's like saying you can't play scrabble because the dictionary changes over time. You specify the engine and wait time before you start a game.
Waiting for a specific duraion may yield different results depending on CPU usage or other variables.
So you just choose in advance what settings to run with and the stopping condition. And then it doesn't matter that if you had run it with different settings, you may have gotten a different answer.
I don't think there's a real fix for the issue, unless someone effectively solves chess someday. Otherwise your win/loss is fundamentally based on the imperfect evaluation of a particular engine.
If it's really just the nondeterminism that bothers you (which is fair enough, preferences vary), there's engines that either are deterministic or can be made so with settings.
Engines can be configured to limit search to a certain depth, which will produce a result after every branch has reached the limit or been pruned. That process will vary in time but be deterministic.
Recent neural based engines tend to not be deterministic, especially if ran multi-threaded.
Even though you could modify an engine to evaluate each of it's moves against the selected "top engine" move to avoid them, there is no clear route to success since there's going to be a lot of overlap between human and computer for more obvious moves... So you'd need some kind of tunable difficulty threshold above which it avoids the best solution.
Even then, your difficulty setting is a gamble on whether your opponent will call your bluff.
In the opposite case, because of the same overlap, false positives are going to be a combination of frustrating and flattering.
I find this is the case in most of the online FPS I have played, the knowledge that cheating is possible combined with the disbelief of the ceiling on human ability makes a huge number of people think you are cheating even if your ability is merely above average. There are also confusing overlaps between cheating behaviour and pros on FPS when trying to evaluate replays e.g wallhackers (especially pro wallhackers) and pros sometimes look very similar, because the pros are attempting to track through the walls in their mind... if they get lucky, a replay makes them look super suspicious and hard to distinguish in a single case. There are going to be a ton of games like this where the cheating behaviour are close or identical to the top pros.
I give them out as gifts too. Everyone loves a chess set.
I've played and lost enough games against engines that I would say I've learnt some of the "best moves" (as suggested by the engine, when analysing why I lost) in almost every "usual" scenario for my "usual" openings all the way to maybe the fourth move. There are a lot of variations, but even past the fourth move I still remember some engine suggestions based on my own errors.
But if you're in check, do the rules say you need to try to escape it? Or can you take the opportunity to capture the queen, thumbing your nose at the false threat?
> When multiple moves have the top score, they are all top moves, even if visual markers (like move arrows) suggest the engine prefers one over the other.
This would be the case for two check mates.
In rapid chess, if you overlook check or put your own King in check, your opponent can claim the win.
Now, what happens in Human Chess, I don't know. Because I would assume that the best move would be to capture the King and win the game. Assuming that you can't capture the King because it's the recommended move, this does seem like something you could exploit to some degree.
Some set-up considerations: 1) Actions must be discrete, or at least binned for restriction, 2) The number of times to restrict is limited by the size of the action space
I would imagine for CartPole, the balancing would become more wobbly, while still somewhat successfully balancing. But in more complicated environments, it could result in much more different behaviors because the states visited (and trajectories) could be different.
So every move you make (I'll be watching you) could end up being the top move. Even if you run the chess engine yourself to decide what not to play, you're still at risk of bad luck because you happen to run the chess engine on a faster or slower machine than the person checking for the top move, and they diverge.
EDIT: Never mind, contesting and getting it wrong causes you to lose - that's the downside.
The page doesn't say, but it's cheating to use an engine yourself to decide what move to make (or to decide whether to contest)
I have a low chess rank (900 on chess.com), yet on an okay game 25% of my moves will be "top engine moves": https://i.imgur.com/TGaDtzr.png
I could even easily find games where I had 50% of top engine moves.
It's really not exceptional. Often the top engine move is the only good move and that only good move is pretty obvious.
You just described all of chess.
The Fisher 960 variant tries to do this, but it can be very different from regular chess, and some of the positions are unbalanced.
I think we can use the fact that engines know when a position is even. There must be millions of even positions in the first 10 or so moves. Pick one of those randomly, and start the game.
One way to adjust would be to have everyone in a round play the same position, and calibrate. As in if black wins 90% of the games, a white win counts for more.
To make top level chess more interesting I have a handful of ideas that work in tandem.
1. Change the scoring and rating systems so that a win is worth more than two draws. E.g a win is 3 points for the winner, draw is 1 point to each player. Game theoretically this should favour players that play for a win and avoid easy draws. But also modifying the rating system is crucial, otherwise we'll get the same drawmeisters dominating the rating list.
2. Change the repetition rule to be similar to xiangqi(Chinese chess) where repetitions are illegal and don't lead to a draw. This eliminates most of the lowest effort draws right out of the gate.
3. Make the game sharper and more complex. The easiest way to do this is just to remove the concept of castling altogether. Former world champion Kramnik has advocated this, and computer analyses of the ruleset is promising. King safety is suddenly a hard problem to solve in most openings and the game becomes much, much sharper.
4(optional). add more pieces. The best way is Seirawan-chess, a modification of Capablanca chess that adds a knight-bishop(hawk) and knight-rook(elephant) without changing the board geometry and starting position.
Norway chess also has the spectaculary stupid idea that if a game is drawn, the players play an armageddon(white gets more time, black wins with a draw) blitz game, and the winner gets half a point extra, so 1.5 to 1. This just ruins it to me. A draw should still be a draw, sometimes the players were just equal and not all draws are lazy. And this makes drawing more attractive again because if you win the armageddon you still get half a victory worth of points. And decided by a blitz game in a classical tournament.
Bishop, Rook, Knight, King, Knight, Rook, Queen, Bishop. Here's the post: https://www.reddit.com/r/chess/comments/yeregq/fischer_rando...
Maybe this should be used as the starting point? Traditional openings would usually give an advantage to White.
I feel like this would be a funny novelty for a YouTube video. Maybe we can get some YouTube grand-masters like Nakamura to tolerate a few games for the schadenfreude.
When they make the move you claim that’s the top engine move, and you’d be correct.
> If a player only has one move available, that move will always be the top engine move, which loses the game.
And you can say, "oh well but the boards are roughly the same" and that's sort of true, but it doesn't really solve the problem of the people memorizing the tree. It just changes the shape of the tree. It goes from looking like a pine to a maple.
If you make the the 960 times as wide, people will only be able to memorize 1/960th as deep.
So not much at all.
In three plies (one sided moves) of chess, there are over eight thousand possible games. By making the game 960 times as wide before you start you're not meaningfully changing the impact of memorization on the outcome of the game. You're just changing how deep you memorize the various trees.
You can say, "well then, mission accomplished!" but the reality is that most of the tree memorization goes pretty deep at the highest levels before a new game is found because you're in effect following the games before it or you're blundering, or, at best, gambling if you've found something kinda unexpected and interesting.
Put another way, any given top rated chess player has a finite set of possible game memorizations. Introducing a mere 960 new configurations at move 0 is only trading 2 to 4 plies worth of depth to the game. It's more complicated than that, because board positions can be essentially forced and board positions can overlap between pre-configurations, etc. But that's the essence of my argument.
You're not meaningfully changing the impact of memorization on the outcome of the game, even if one thousandth sounds like a lot, it isn't really when dealing with permutations.
> Starting from move 2, players can claim their opponent’s last move was a top engine move. This immediately ends the game.
> Claims are settled by asking the chess engine to evaluate the position before the contested move. If the chess engine suggests the opponent’s move, the claimant wins the game. If not, their opponent wins instead.
So your suggestion would be a completely different game.
Once the opponent has made her move, she can place the chicken on any other empty field.
> Checkmating loses the game, as it is always the top engine move. Rather than aiming for checkmate, players seek to force their opponent to make a top engine move. If a player only has one move available, that move will always be the top engine move, which loses the game.
In other words, whatever the rules, you can have an engine that will try to do the best according to the rules.
I think you overestimate how much you would have to defer to this external engine that would have to say which move is "top".
Every move you advance you chessboard situation by one move only and that move is already part of the tree calculated previously. You don't need to search through massively more new moves because, assuming sane players, the move each player makes is one of the very few top moves previously considered by the engine.
Mastering that kind of new threats does not seem easy IMO, and in fact could well be mastered by... computers ;-)
I've heard it said that the best parodies are almost as good as the things they parody (and a sign that the comedians in question both love and understand the thing they are making a parody of). It could be argued that this chess variant is a really good "parody" in that sense, but encoded in the rules of the game itself.
When you have mate-in 1 it's impossible to have anything else recommended by the computer. Flip of a coin for which one is on top when you have two?
Pfff... I do this all the time /s
That's making my head hurt.
https://en.wikipedia.org/wiki/Fischer_random_chess#Praising_...