The Next Big Thing in Computer Chess?

[jhellis3]

Well, IMO he had a point back then. His post was about Intel's Larrabee GPU project:

Why are you trying to explain things to me as if I wasn't there? I am the one who posted it. I witnessed the conversation at the time; I participated in it. I know what the thread was about, FFS. The only point is that he was completely wrong; he allowed what he wanted to be true to influence his perception of reality. And the CPU being "close" to the GPU is orthogonal to the discussion in the post.

[smatovic]

Well, IMO he had a point back then. His post was about Intel's Larrabee GPU project:

Why are you trying to explain things to me as if I wasn't there?

Hmm, cos this is a public forum, and others might not get the context?

I am the one who posted it. I witnessed the conversation at the time; I participated in it. I know what the thread was about, FFS. The only point is that he was completely wrong; he allowed what he wanted to be true to influence his perception of reality. And the CPU being "close" to the GPU is orthogonal to the discussion in the post.

Well, IMO he had a point back then.

--
Srdja

[jhellis3]

Hmm, cos this is a public forum, and others might not get the context?

I suppose if they didn't click the link and weren't able to read. But then, if they can't read...

Well, IMO he had a point back then.

Anyone who thinks this is a wrong as now as Nick was then.

[smatovic]

Hmm, cos this is a public forum, and others might not get the context?

I suppose if they didn't click the link and weren't able to read. But then, if they can't read...

...it is an 12 years old thread about an dead hardware architecture, and your quoted post was on page 48, from 61.

https://forum.beyond3d.com/threads/22-n ... 31/page-48

--
Srdja

[jhellis3]

Ok.

[jdart]

It is quite rare that games are won with some magic move that requires super-deep search to find. I am very impressed by how fast Stockfish solves tactical problems, and how few nodes it takes to zero in on the correct solution. If you just set it to search with a fairly low node limit, it would still be a formidable opponent for humans and weaker engines. If you get to the right move fast, it does not matter so much what your branching factor is.

[towforce]

For many optimisation problems, there are algorithms that get you close to the optimal solution, but don't guarantee hitting it. These algorithms usually require a lot less memory and CPU time than algorithms that are guaranteed to hit the optimal solution.

So... is it possible to create an algorithm for chess that doesn't need much CPU time or memory, and usually gets you close to the "correct" solution without constructing a game tree (so ply-1) or otherwise requiring large amounts of time or memory?

IMHO the evidence suggests that such algorithms are likely to be possible.

There is an old proof (which has been peer reviewed) that, for games like chess, as the board gets bigger, the game tree explodes exponentially - but that doesn't disqualify the possibility of a relatively simple algorithm that gets you close to the optimal solution most of the time: such algorithms exist in other domains with exponential growth in complexity.

"You’re more likely to unlock a big leap in performance by trying differently than by trying harder. You might be able to work 10% harder, but a different approach might work 10x better. Remain focused on the core problem, but explore a new line of attack. Persistence is not just about effort, but also strategy. Don’t merely try harder, try differently" - James Clear

[smatovic]

....

Point is, to find such an algorithm you will need a lot of CPU time and a lot of memory, doubt you can do it in your mind alone, I guess you tried

--
Srdja

[towforce]

Point is, to find such an algorithm you will need a lot of CPU time and a lot of memory, doubt you can do it in your mind alone, I guess you tried

If there is a "good" solution to chess in linear space, it would probably be relatively easy to find. A good solution is likely to be non-linear, though.

I came up with "straight lines will work" - but they will only work for parts of the model which are close together - and it seems likely that some parts of the model would be far apart.

Your answer (quoted above) prompted me to have an idea - something I'm calling "compatible lines". I will reinvestigate...

[smatovic]

Larry Kaufman mentioned in another post 97% draws between SF 16 and SF 15 with 2"+1' TC in a 620 games match with standard opening and 2 threads, he estimated 99% draws for Rapid TC. How much Elo is still to gain on CCRL Blitz? Time's running, finish line in sight.
--
Srdja

Engines need to use selective search in order to cause the opponent to fall to designed traps and this is not the way they work today.

If you want to find what is possible to achieve you need to build anti-stockfish engine when the target is beating stockfish when you get more time.

Anti-stockfish with white is going to work in the following way:
For white search every possible legal move but for black do not search every possible legal move but simply calculate the move that stockfish is going to play and prune the rest of the moves.

After part of your time(for example half of the target time but maybe different percentage is optimal) you do not calculate stockfish's moves because calculating stockfish's move is too expensive but you remember the stockfish's moves that you already calculated in order to prune the rest of the moves in your search.

In this way the engine may prefer lines when stockfish does mistakes so there is a bigger probability to win and the question is what is the percentage of wins that you can get against stockfish 2''+1' TC in this way(I guess clearly more than using unequal time control).

So you basically suggest to give one player more time and outsearch the other engine, this is based on the assumption that there is still something to find via the outsearch approach.....the effective branching factor of ~1.25 of modern SF indicates that there is not much to search anymore nowadays, or alike.

--
Srdja