B strategy is the future.

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Harvey Williamson, bob

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
User avatar
MikeGL
Posts: 873
Joined: Thu Sep 01, 2011 12:49 pm

Re: B strategy is the future.

Post by MikeGL » Thu Nov 08, 2018 5:18 am

mclane wrote:
Sun Nov 04, 2018 3:47 pm
The race is not anymore about strength. The race is about b strategy and c strategy.
Komodo, Houdini and Stockfish showed that a strategy works. Or ab strategy.

But the future is not about this. It is about b strategy (no brute force)
and about c strategy (Computing like the humans do, with patterns).

Strength is not the target anymore. But PLANS.

Chess engines need to PLAN chess.
But I think there are no shortcuts. Engines needs to be a heavy brute forcer to play positionally and strategically.
Was it GM Joel Benjamin or another chess writer who claimed (during G.Kasparov-DB2 match) that Deep Blue 2
played strong positional chess almost similar to Capablanca due to its billions of nodes per second of search?

There was even the controversial h5!! move of Deep Blue 2 as black wherein kasparov complained because
that h5!! is too positional to be played by DB2 nor by any computers.
It turns out later that the then strong engine and fast brute forcer Junior 5.0 was able to find that h5! move being protested by kasparov.

1997 Kasparov-DB2 Game 5
After 11.Nd2 of Kasparov, DB2 played 11...h5! which made Kasparov suspicious because
this is a very strong positional move according to Kasparov.
Turns out Junior 5.0 at that time was also able to replicate 11...h5 of DB2 in less than 1 hr.
[d]r2qk2r/pp3ppp/2p1pn2/4n3/1b6/3P2PP/PPPN1PB1/R1BQK2R b KQkq - 4 11
I told my wife that a husband is like a fine wine; he gets better with age. The next day, she locked me in the cellar.

Uri Blass
Posts: 8530
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: B strategy is the future.

Post by Uri Blass » Thu Nov 08, 2018 5:45 am

mclane wrote:
Mon Nov 05, 2018 10:13 pm
I know NO human being playing chess doing brute force. What a nonsense.
Of course in blitz it is not possible but at long time control I try to look at all the options that I have in the next move and all options that the opponent has in reply to my move as part of my thinking.

User avatar
mclane
Posts: 15957
Joined: Thu Mar 09, 2006 5:40 pm
Location: US of Europe, germany
Full name: Thorsten Czub
Contact:

Re: B strategy is the future.

Post by mclane » Thu Nov 08, 2018 9:17 am

Chess engines using heavily brute force search are not playing chess.
They emulate it via search tree.

What we need is an engine that knows about chess and chooses those branches and lines
That makes sense using a chess knowledge,
Like humans are doing it,
What seems like a fairy tale today may be reality tomorrow.
Here we have a fairy tale of the day after tomorrow....

Uri Blass
Posts: 8530
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: B strategy is the future.

Post by Uri Blass » Thu Nov 08, 2018 10:12 am

mclane wrote:
Thu Nov 08, 2018 9:17 am
Chess engines using heavily brute force search are not playing chess.
They emulate it via search tree.

What we need is an engine that knows about chess and chooses those branches and lines
That makes sense using a chess knowledge,
Like humans are doing it,
stockfish and other top programs already do a lot of pruning of lines that do not make sense otherwise they could not get depths above 12 in the opening.

Chess knowledge of humans may be misleading and they may prune also good lines.
Of course at fast time control humans have no better choice but when the time control becomes longer it is better to use brute force at least in the first ply in order not to miss something trivial.

Computers may use do brute force in more plies because they are faster so I think that a better strategy for computers should be use alphabeta with no pruning(let say in 1% of their time) and later they can extend or prune the lines that they calculated based on chess knowledge.

Note that I believe that computers can have more knowledge than humans so they can allow themselves to do more pruning.
beginners who start to play chess do not know patterns so brute force(with check extensions) is the only chance for them to find patterns that they do not know.




imagine human who do not know the pattern.

How can he find mate in 2?(and the first human who found it did not know the pattern)
He must do brute force search otherwise he is going to say Qg8+ is not logical because white lose the queen and never find Qg8+
Of course every human who is not a beginner know this pattern but there are still many patterns that humans (at level of 1800 or 2000) do not know.

If the time control is long enough they should be able to find patterns that they do not know and believe the first move of them is bad
by brute force.

User avatar
mclane
Posts: 15957
Joined: Thu Mar 09, 2006 5:40 pm
Location: US of Europe, germany
Full name: Thorsten Czub
Contact:

Re: B strategy is the future.

Post by mclane » Thu Nov 08, 2018 10:50 am

By relying on brute force you do not allow the chess engines to grow and get more knowledge.
IMO b strategy is the right decision because it completely gives up the idea that alpha beta or A strategy would allow to play chess.
Ok it looks like chess. But it isn’t.

Stockfish can compute 12 or 26 plies.
And it will not be able to play chess.
It can beat the human world chess champion.
But still cannot play chess,

When I have 20 keys,
And the door is locked,
I would not say that a computer trying out all 20 keys until it gets open, is intelligent.

Same counts for chess.
A machine can mazybe try out all 20 keys much faster then any human being could do.
But concluding the machine is intelligent is way wrong.

It is not.

It’s faster.
But not clever.

A human looks to the key.
To the lock,
And then decides which keys could fit.

This is the way humans play chess. You look to the board position and see ideas and plans.
And then you try them out in your brain.
But you do NOT look for unimportant moves.
A machine does so. Therefore even when beating the world chess champion,
A machine with A strategy is not playing chess, it solves to beat the human because it is faster.
But it is not really understanding chess at all.

It’s only faster with the keys,

So IMO we have to teach the engines which keys could fit.
What seems like a fairy tale today may be reality tomorrow.
Here we have a fairy tale of the day after tomorrow....

Post Reply