Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Reimagining Chess with AlphaZero

cacm.acm.org

Post reply on HN
Print view View original

31–40 of 116 posts

Re: Reimagining Chess with AlphaZero

#31

Post by otabdeveloper4 »

monsieurbanana wrote: post #15
gazookgaboot wrote: post #12

There is only one way to fix chess and it's called Shogi (A chess style game played in Japan). Shogi is similar to the bughouse chess variant in that captured pieces return to the board on the opponent's side. This prevents draw by insufficient material which is the real bane of Grandmaster level chess. Of course this isn't exactly chess anymore but like any design decision there are no perfect solutions.

There is only one way to fix chess and it's called Go. Of course, this isn't exactly chess anymore but like any design decision there are no perf- Wait a minute, Go is perfect. If we forget about ko rules.

Respectfully disagree.

Re: Reimagining Chess with AlphaZero

#32

Post by adgjlsfhk1 »

snovv_crash wrote: post #29
adgjlsfhk1 wrote: post #26

Earlier quoted context omitted.

the draw problem isn't entirely human. computers draw something like 80% of games even with openings specifically chosen to reduced the number of draws.

But it's that because we optimize the algorithms under the same ruleset? I imagine if a draw was counted as a loss to both sides, there would be a lot more aggressive plays, since just stonewalling a skilled opponent would be the same as losing.

lc0 has had bets trained where draws are a loss. it doesn't change much

Re: Reimagining Chess with AlphaZero

#33

Post by beardyw »

weedfroglozenge wrote: post #27
discardable_dan wrote: post #18

Earlier quoted context omitted.

I can't tell if this is satire. Draws by insufficient material almost never occur in Grandmaster chess. I have not seen a single grandmaster game end this way in a year of watching top-level chess games daily. If you inspect all of the dozens of draws in this world's Tata Steel, which included ~300 GM-level games, you won't find a single that ends this way.

Because they are ending the game before it gets to that stage. Similar to seeing check mates 3, 4, 5 moves in advance, they can see with perfect play that the reduction in pieces will end in an end game with insufficient material.

Exactly. I find it hard to see it myself, but I often play out games like that with an engine and very often the pieces just peter out to a draw.

Re: Reimagining Chess with AlphaZero

#34

Post by BearsAreCool »

mlyle wrote: post #20
discardable_dan wrote: post #18

Earlier quoted context omitted.

I can't tell if this is satire. Draws by insufficient material almost never occur in Grandmaster chess. I have not seen a single grandmaster game end this way in a year of watching top-level chess games daily. If you inspect all of the dozens of draws in this world's Tata Steel, which included ~300 GM-level games, you won't find a single that ends this way.

In addition to what my sibling comment says-- there's a whole lot of draws by agreement because the position is one that everyone knows becomes draws by insufficient material.

While this is slightly nitpicky, from what I gather many of those draws are because neither side can make progress against the opponent. These are positions where in order for the game to move "forward" both sides would need to individually choose to weaken their position, leaving them playing non-moves until triggering threefold repetition or agreeing to a draw.

Re: Reimagining Chess with AlphaZero

#37

Post by mlyle »

BearsAreCool wrote: post #34
mlyle wrote: post #20

Earlier quoted context omitted.

In addition to what my sibling comment says-- there's a whole lot of draws by agreement because the position is one that everyone knows becomes draws by insufficient material.

While this is slightly nitpicky, from what I gather many of those draws are because neither side can make progress against the opponent. These are positions where in order for the game to move "forward" both sides would need to individually choose to weaken their position, leaving them playing non-moves until triggering threefold repetition or agreeing to a draw.

> These are positions where in order for the game to move "forward" both sides would need to individually choose to weaken their position, leaving them playing non-moves until triggering threefold repetition or agreeing to a draw.

Sure, but there's also a whole lot of draws where insufficient material is inevitable. E.g. pawn vs nothing, where the king cannot reach the pawn to defend.

And there's also draws where the only move "forward" doesn't weaken the position, but lead towards an end of insufficient material. E.g. most situations pawn vs. bishop.

Re: Reimagining Chess with AlphaZero

#38

Post by Andrew_nenakhov »

joelthelion wrote: post #36
Andrew_nenakhov wrote: post #11

Ah, chess is fine as it is. It's just we meek humans are no longer the ones who play it best.

There is one problem: there are too many draws at top human level.

I'm in top 1% of rapid chess players on lichess, and I hardly ever see a draw. If top 100 or 500 players on the whole planet have this problem, well, they can try some other game, probably, I don't mind.

Re: Reimagining Chess with AlphaZero

#39

Post by travelhead »

beardyw wrote: post #35

It would be interesting to see if there is anything which would let black start on equal terms.

Yes! White moves first, then black makes two moves in a row, followed by regular back and forth moves after.

This would theoretically allow black to start on equal terms.

Re: Reimagining Chess with AlphaZero

#40

Post by londons_explore »

riwsky wrote: post #5
pests wrote: post #3

Great read. Interesting variations to the rules. Here's an annotated quote: "The nine changes considered in this study are listed in Table 1. No-castling and No-castling (10) involve a full and partial [not allowed during first 10 turns / 20 plies] restriction on the castling rule. Pawn-one-square, Semi-torpedo [can move 2 squares on 2ed and 3rd ranks], Torpedo [can move 2 squares], Pawn-back [up to your 2ed rank, do…

Software rendering of these rules might lag a bit—but if you have dedicated hardware, it should be possible to make this work.

Help me understand... Why would I need dedicated hardware to render chess moves that are slightly different to regular chess moves...?

Your comment makes no sense...

Post reply on HN