Botvinnik Markov revisited

[Don]

What is the state of the art on Botvinnik Markov extensions?

I'm experimenting with them a bit now and I researched the idea on the chessprogramming wiki and other places and saw a lot of enthusiasm for the idea, but I'm not aware that anyone is currently using it.

I don't remember seeing it in Stockfish for example, but some of the old posts indicated that Tord was getting some good results based on a test match and he seemed fairly enthusiastic about it.

So what's the deal?

[zamar]

What is the state of the art on Botvinnik Markov extensions?

I don't remember seeing it in Stockfish for example, ...

So what's the deal?

The deal is that I'm still so inexperienced in computer chess that I had never heard of this idea before. It's likely some Glaurung 1.x stuff which for some reason never entered Glaurung 2.x. Likely Tord knows better...

[Don]

What is the state of the art on Botvinnik Markov extensions?

I don't remember seeing it in Stockfish for example, ...

So what's the deal?

The deal is that I'm still so inexperienced in computer chess that I had never heard of this idea before. It's likely some Glaurung 1.x stuff which for some reason never entered Glaurung 2.x. Likely Tord knows better...

I can only assume then, that Tord eventually rejected the idea or perhaps the threatMove stuff stockfish does effectively replaces it.

[Dann Corbit]

What is the state of the art on Botvinnik Markov extensions?

I'm experimenting with them a bit now and I researched the idea on the chessprogramming wiki and other places and saw a lot of enthusiasm for the idea, but I'm not aware that anyone is currently using it.

I don't remember seeing it in Stockfish for example, but some of the old posts indicated that Tord was getting some good results based on a test match and he seemed fairly enthusiastic about it.

So what's the deal?

Smarthink puts a lot of energy into extensions and it works well for him.

[Don]

What is the state of the art on Botvinnik Markov extensions?

I'm experimenting with them a bit now and I researched the idea on the chessprogramming wiki and other places and saw a lot of enthusiasm for the idea, but I'm not aware that anyone is currently using it.

I don't remember seeing it in Stockfish for example, but some of the old posts indicated that Tord was getting some good results based on a test match and he seemed fairly enthusiastic about it.

So what's the deal?

Smarthink puts a lot of energy into extensions and it works well for him.

Extensions in general, or this specific extension?

I'm running a test which indicates that it is worth a few elo.

[Dann Corbit]

What is the state of the art on Botvinnik Markov extensions?

I'm experimenting with them a bit now and I researched the idea on the chessprogramming wiki and other places and saw a lot of enthusiasm for the idea, but I'm not aware that anyone is currently using it.

I don't remember seeing it in Stockfish for example, but some of the old posts indicated that Tord was getting some good results based on a test match and he seemed fairly enthusiastic about it.

So what's the deal?

Smarthink puts a lot of energy into extensions and it works well for him.

Extensions in general, or this specific extension?

I'm running a test which indicates that it is worth a few elo.

I have seen the code and so I think it is not proper for me to answer with any details because it is private code. Perhaps S.Markov will chime in eventually.

[Don]

What is the state of the art on Botvinnik Markov extensions?

I'm experimenting with them a bit now and I researched the idea on the chessprogramming wiki and other places and saw a lot of enthusiasm for the idea, but I'm not aware that anyone is currently using it.

I don't remember seeing it in Stockfish for example, but some of the old posts indicated that Tord was getting some good results based on a test match and he seemed fairly enthusiastic about it.

So what's the deal?

Smarthink puts a lot of energy into extensions and it works well for him.

Extensions in general, or this specific extension?

I'm running a test which indicates that it is worth a few elo.

I have seen the code and so I think it is not proper for me to answer with any details because it is private code. Perhaps S.Markov will chime in eventually.

I'm not asking for details. Are you saying that he put a lot of energy into this specific extension or just extensions in general?

[Dann Corbit]

What is the state of the art on Botvinnik Markov extensions?

I'm experimenting with them a bit now and I researched the idea on the chessprogramming wiki and other places and saw a lot of enthusiasm for the idea, but I'm not aware that anyone is currently using it.

I don't remember seeing it in Stockfish for example, but some of the old posts indicated that Tord was getting some good results based on a test match and he seemed fairly enthusiastic about it.

So what's the deal?

Smarthink puts a lot of energy into extensions and it works well for him.

Extensions in general, or this specific extension?

I'm running a test which indicates that it is worth a few elo.

I have seen the code and so I think it is not proper for me to answer with any details because it is private code. Perhaps S.Markov will chime in eventually.

I'm not asking for details. Are you saying that he put a lot of energy into this specific extension or just extensions in general?

Let me put it this way:
No program I have ever seen puts nearly so much effort into extensions (e.g. do you know any programs that have seven distinct categories for extensions?).
If we are to interpret the Botvinik Markov extension as using null move to detect threats and backing up two plies on the move stack then he definitely does that.
http://www.stmintz.com/ccc/index.php?id=318839

Code: Select all

if(SameTarget(ThreatMove[Ply],ThreatMove[Ply-2])) extend_flag=true;

Of course, there is a lot more to it than that.

For specifics about effectiveness or implementation details, I think you will need to get S.Markov to comment.

[Don]

What is the state of the art on Botvinnik Markov extensions?

I'm experimenting with them a bit now and I researched the idea on the chessprogramming wiki and other places and saw a lot of enthusiasm for the idea, but I'm not aware that anyone is currently using it.

I don't remember seeing it in Stockfish for example, but some of the old posts indicated that Tord was getting some good results based on a test match and he seemed fairly enthusiastic about it.

So what's the deal?

Smarthink puts a lot of energy into extensions and it works well for him.

Extensions in general, or this specific extension?

I'm running a test which indicates that it is worth a few elo.

I have seen the code and so I think it is not proper for me to answer with any details because it is private code. Perhaps S.Markov will chime in eventually.

I'm not asking for details. Are you saying that he put a lot of energy into this specific extension or just extensions in general?

Let me put it this way:
No program I have ever seen puts nearly so much effort into extensions (e.g. do you know any programs that have seven distinct categories for extensions?).
If we are to interpret the Botvinik Markov extension as using null move to detect threats and backing up two plies on the move stack then he definitely does that.
http://www.stmintz.com/ccc/index.php?id=318839

Code: Select all

if(SameTarget(ThreatMove[Ply],ThreatMove[Ply-2])) extend_flag=true;

Of course, there is a lot more to it than that.

For specifics about effectiveness or implementation details, I think you will need to get S.Markov to comment.

I'm a firm believer that there is more than one way to skin a cat.

[Dann Corbit]

What is the state of the art on Botvinnik Markov extensions?

I'm experimenting with them a bit now and I researched the idea on the chessprogramming wiki and other places and saw a lot of enthusiasm for the idea, but I'm not aware that anyone is currently using it.

I don't remember seeing it in Stockfish for example, but some of the old posts indicated that Tord was getting some good results based on a test match and he seemed fairly enthusiastic about it.

So what's the deal?

Smarthink puts a lot of energy into extensions and it works well for him.

Extensions in general, or this specific extension?

I'm running a test which indicates that it is worth a few elo.

I have seen the code and so I think it is not proper for me to answer with any details because it is private code. Perhaps S.Markov will chime in eventually.

I'm not asking for details. Are you saying that he put a lot of energy into this specific extension or just extensions in general?

Let me put it this way:
No program I have ever seen puts nearly so much effort into extensions (e.g. do you know any programs that have seven distinct categories for extensions?).
If we are to interpret the Botvinik Markov extension as using null move to detect threats and backing up two plies on the move stack then he definitely does that.
http://www.stmintz.com/ccc/index.php?id=318839

Code: Select all

if(SameTarget(ThreatMove[Ply],ThreatMove[Ply-2])) extend_flag=true;

Of course, there is a lot more to it than that.

For specifics about effectiveness or implementation details, I think you will need to get S.Markov to comment.

I'm a firm believer that there is more than one way to skin a cat.

As far as that goes -- some position x has 9 legal moves:
Joe examines the 9 possible moves and decides to extend move 4 by one ply.

Fred examines the 9 possible moves and decides to decrease all of them by one ply except move 4.

Essentially, they did exactly the same thing via opposite method.

Poor kitty.