Peak Chess
Moderator: Ras
-
- Posts: 12512
- Joined: Thu Mar 09, 2006 12:57 am
- Location: Birmingham UK
- Full name: Graham Laight
Peak Chess
The discussions about singularity and peak AI in the AI forum have prompted me to run a poll on peak chess. How will Elo improvement in chess end?
Human chess is partly about tactics and strategy, but mostly about memory
-
- Posts: 1025
- Joined: Sun Jul 25, 2010 10:07 pm
- Location: the Netherlands
- Full name: Jef Kaan
Re: Peak Chess
1) and 5) don't have to be so different.
but with G algorithm you probably mean, no search. with is imo implausible
(agreeing with Uri Blass (polsting to you some time ago ).
but with G algorithm you probably mean, no search. with is imo implausible
(agreeing with Uri Blass (polsting to you some time ago ).
-
- Posts: 12512
- Joined: Thu Mar 09, 2006 12:57 am
- Location: Birmingham UK
- Full name: Graham Laight
Re: Peak Chess
If a game had just one piece that moved in just one way, the algorithm to accurately evaluate the position would be easy. Each piece added is like adding an extra dimension to the puzzle.. To make matters worse, the queen has two ways of moving, and a pawn has 4 ways, turning into 5 ways when it promotes.
It's not surprising that humans have never managed to find the quick algorithm that doesn't require a game tree. Worse still, for as long as I can remember, everyone has been asking how to improve existing methods instead of asking how to find the universal rules that always apply.
Human chess is partly about tactics and strategy, but mostly about memory
-
- Posts: 1025
- Joined: Sun Jul 25, 2010 10:07 pm
- Location: the Netherlands
- Full name: Jef Kaan
Re: Peak Chess
also then, it's a matter of choice, inherent to the game, and thus a decision tree.If a game had just one piece that moved in just one way
As for your example, take just one pawn on the 7th rank, can indeed move only
one way, but then has a choice to promote to queen, rook, bishop, or knight.
In most situations, promoting to queen is the best, for obvious reasons.
It's also how human chess players(*) decide on a move, define candidates (tree),
evaluate (at end of tree), compare, and decide.
(*) well, except players like Karpov maybe

Ps in some way, and the eval via an advanced neural has some built in
(positional) evals built in whereby tactical factors (eg attacking chances)
are being are taken into account, but in itself this is not accurate
for tactical positions. The minimax theorem by Von Neumann, in itself not
really such a revolutionary thought, is a fundamental concept based
two player games with perfect information like chess, checkers etc
is based on a decision tree, of course.
If you think you can find something better i won't attack you (like those
who didn't understanding my reasoning to 'prove' chess is a draw) but i like
to see more information, examples, etc. Ofcourse brute force tactical
calculation isn't all there is in chess, and it also has limits, but nevertheless,
some search is imo obligatory, in -almost- all situations/positions.
-
- Posts: 12512
- Joined: Thu Mar 09, 2006 12:57 am
- Location: Birmingham UK
- Full name: Graham Laight
Re: Peak Chess
As a human, the more you know about chess, the lower the percentage of positions you have to search in.
If you take the pieces and put them on the board in a random but legal way, the GM will say, "This position doesn't make sense". I have personally witnessed this with two GMs, one of whom was Nigel Short. Both GMs used almost exactly the same words.
When you have the God algorithm, if it's simple enough for a human to learn, and you have mentally applied it correctly to thousands of positions before, then every position will make sense.
Human chess is partly about tactics and strategy, but mostly about memory
-
- Posts: 268
- Joined: Wed Jun 16, 2021 2:08 am
- Location: Berlin
- Full name: Jost Triller
Re: Peak Chess
I'm surprised so many vote for the first option. I don't think humanity will continue to exist forever.
-
- Posts: 12512
- Joined: Thu Mar 09, 2006 12:57 am
- Location: Birmingham UK
- Full name: Graham Laight
Re: Peak Chess
Hmmmm... any possibility you might be overthinking this a tad?

Human chess is partly about tactics and strategy, but mostly about memory
-
- Posts: 268
- Joined: Wed Jun 16, 2021 2:08 am
- Location: Berlin
- Full name: Jost Triller
Re: Peak Chess
No. No such possibility.
-
- Posts: 12512
- Joined: Thu Mar 09, 2006 12:57 am
- Location: Birmingham UK
- Full name: Graham Laight
Re: Peak Chess
Notice to all persons working on peak chess: your deadline is 500 million years. At this time, the expansion of the sun will heat up earth to a level which will make it uninhabitable (the end of its "Goldilocks" era).
Human chess is partly about tactics and strategy, but mostly about memory
-
- Posts: 268
- Joined: Wed Jun 16, 2021 2:08 am
- Location: Berlin
- Full name: Jost Triller
Re: Peak Chess
Given my track record, I'll probably start in 499999999 years and 360 days then.towforce wrote: ↑Sun Dec 15, 2024 11:42 pm
Notice to all persons working on peak chess: your deadline is 500 million years. At this time, the expansion of the sun will heat up earth to a level which will make it uninhabitable (the end of its "Goldilocks" era).