Knights or Bishops? What AlphaZero Thinks

Sort:
IMKeto
Ubik42 wrote:
Keto you aren’t going to play the WC either, don’t try and tell me you won’t look at their games, though!

I was asked to but declined as it would have interfered with my gym time, bacon buying, and keto lifestyle.

To be honest with you...Watching chess bores me.  Oh i can do the occasionally look at games, but to just watch a game for hours?  No thank you.

For the Carlsen/Kajakian match a friend invited a bunch of us to his office to watch the games.  I told him to give my seat to someone else.

Boradin75
Uhohspaghettio1 wrote:

I feel like you are still framing it wrong. Alphazero would not be a match for traditional top chess engines like Stockfish because otherwise why would they refuse to ever compete? LC0 incorporates both traditional chess evaluation techniques and neural network techniques, it's a hybrid. 

Recently, as in July of 2020, Stockfish started to incorporate neural networks in it as well to help its evaluation function. However 1) it was already at the top in 2020 before it started to do this and 2) it is still predominantly a traditional chess engine that has been created by human intervention.   

https://en.wikipedia.org/wiki/Top_Chess_Engine_Championship

https://en.wikipedia.org/wiki/Stockfish_(chess)

So if OP wants to talk about something, he should be talking about these engines. AlphaZero sucks compared to these, that is my point.  

  

Please provide a source for these claims. 

I've studied the mathematics behind NNs and chess engines. I have yet to find an argument that deep dbase tree searches are preferable to the multiple NN layered approach that AlphaZero uses. In fact, that's what made all the headlines a few years back, when the deep learning communities embraced stochastic processes. Can you provide some material that backs your claim that a "traditional chess engine" would be better?

Ubik42
Yah Keto I mean look at analyzed games later. I rarely “watch” chess, though I will tune in to Carlsen/ Nepo since I am sure there will be top notch commentary.
IMKeto
Ubik42 wrote:
Yah Keto I mean look at analyzed games later. I rarely “watch” chess, though I will tune in to Carlsen/ Nepo since I am sure there will be top notch commentary.

Ill check in on it but i wont be staring at a screen for hours.  Half of the fun of watching online is all the would be "GM's" that come out of the wood work.  The last match with Fabi was a very good match(even with all the whining about the draws).  The game where the french engine "sessi" or whatever its called.  Showed that Fabi had a mate in 39.  NO human sees it, and no one comments on it until the engine shows it.  Then its all these comments: "OMG he missed the mate."  Yea...like any of you saw it.

IMKeto

If Svidler, and or Seirawan are commenting ill be more apt to tune in.  If "The engine junkie"(Maurice Ashley) is commenting and just repeating what engines show i wont.

Ubik42
Oh yeah. Anyone can sit there with an engine. “look he just lost .21 pawns with that move!!”
IMKeto
Ubik42 wrote:
Oh yeah. Anyone can sit there with an engine. “look he just lost .21 pawns with that move!!”

I know a guy that is a career USCF C player that gives his life to whatever the engine says.   The guy will run a position for days just to see what happens.  He is around my age (58) and hasn't improved, and you cant talk to him about how he studies. 

Stil1

Engine evals often change, farther down the line.

You can let an engine crunch a position for a full month, and it might spit out an eval of +0.23.

But then you play a few moves down the line, in the same line it evaluated, and suddenly the eval is now -0.16 ...

Personally, I only pay attention to the eval if it swings more than 1.0 in either direction ...

beebo10
It really depends on the game. In very closed games I definitely prefer knights than bishops as they can find a way to penetrate the lines and control some good squares. In open games definitely the bishops are much stronger as they control more squares with their longer range movements
haiaku
Ubik42 wrote:
It’s N for knight, not K!

Sorry! I wrote too fast and forgot the ambiguity between King and Knight 🤭. Changed it.

Arnaut10

Interesting