Let Dr Chess 1.30.10 participate in any tournament.

[Han]

Geschrieben von: / Posted by: Han at 11 July 2000 17:09:27:
Let Dr Chess participate in any tournament that you guys like to organise. No permission required! Just go to my page, and download it!
The only thing that I request for is to let me know the results of Dr Chess, and your personal comments!

[Christian]

Geschrieben von: / Posted by: Christian at 12 July 2000 12:48:23:
Als Antwort auf: / As an answer to: Let Dr Chess 1.30.10 participate in any tournament. geschrieben von: / posted by: Han at 11 July 2000 17:09:27:

Let Dr Chess participate in any tournament that you guys like to organise. No permission required! Just go to my page, and download it!
The only thing that I request for is to let me know the results of Dr Chess, and your personal comments!

I just watched a few games of dr chess.
The time allocation is exactly like tscp 152.
Do have used some code of this engine?
BTW: result was 3:3
Christian

[Dann Corbit]

Geschrieben von: / Posted by: Dann Corbit at 12 July 2000 18:32:17:
Als Antwort auf: / As an answer to: Re: Let Dr Chess 1.30.10 participate in any tournament. geschrieben von: / posted by: Christian at 12 July 2000 12:48:23:

Let Dr Chess participate in any tournament that you guys like to organise. No permission required! Just go to my page, and download it!
The only thing that I request for is to let me know the results of Dr Chess, and your personal comments!

I just watched a few games of dr chess.
The time allocation is exactly like tscp 152.
Do have used some code of this engine?
BTW: result was 3:3

That's the general idea behind TSCP -- learn how to write chess programs. The algorithm in TSCP's time control is incredibly simple. I suggested it to Tom, but he already knew all about it [decided it was a good idea] and implemented it in a few minutes!
Despite it's simplicity, I think it is probably superior to what a lot of other programs do at fixed time controls. You will find that it very rarely loses on time.
Against what engine?


My FTP site

[Dann Corbit]

Geschrieben von: / Posted by: Dann Corbit at 12 July 2000 18:36:50:
Als Antwort auf: / As an answer to: Re: Let Dr Chess 1.30.10 participate in any tournament. geschrieben von: / posted by: Dann Corbit at 12 July 2000 18:32:17:
One more note on TSCP:
If you add a piece list, it will add 200 ELO to the strength and triple the NPS the engine can produce for sparse positions.
Add better draw detection, and you would have another huge boost in ELO.
With the tiny amount of code Tom has produced, I think the results are pretty incredible.


My FTP site

[Mogens Larsen]

Geschrieben von: / Posted by: Mogens Larsen at 12 July 2000 19:13:21:
Als Antwort auf: / As an answer to: Re: Let Dr Chess 1.30.10 participate in any tournament. geschrieben von: / posted by: Dann Corbit at 12 July 2000 18:32:17:

Against what engine?

A wild and reckless guess would be Mint.
Best wishes...
Mogens

[Christian Söderström]

Geschrieben von: / Posted by: Christian Söderström at 12 July 2000 19:30:17:
Als Antwort auf: / As an answer to: Re: Let Dr Chess 1.30.10 participate in any tournament. geschrieben von: / posted by: Mogens Larsen at 12 July 2000 19:13:21:

Against what engine?

A wild and reckless guess would be Mint.
Best wishes...
Mogens

Sure, blame the Swedish guy You guys in purple are all the same.
Actually, thats another Christian, or maybe an imposter pretending to be me (trying to draw some cheap glory from my fame). I believe he was talking about TSCP, I'm sure Mint would have done far worse than 3-3.
- Christian Söderström

[Mogens Larsen]

Geschrieben von: / Posted by: Mogens Larsen at 12 July 2000 20:04:27:
Als Antwort auf: / As an answer to: Re: Let Dr Chess 1.30.10 participate in any tournament. geschrieben von: / posted by: Christian Söderström at 12 July 2000 19:30:17:

Against what engine?

A wild and reckless guess would be Mint.

Sure, blame the Swedish guy You guys in purple are all the same.
Actually, thats another Christian, or maybe an imposter pretending to be me (trying to draw some cheap glory from my fame). I believe he was talking about TSCP, I'm sure Mint would have done far worse than 3-3.
- Christian Söderström

It was worth a shot since I couldn't recall anyone else with that first name. But it could be Christian Koch off course.
Best wishes...
Mogens

[Han]

Geschrieben von: / Posted by: Han at 13 July 2000 09:46:54:
Als Antwort auf: / As an answer to: Re: Let Dr Chess 1.30.10 participate in any tournament. geschrieben von: / posted by: Dann Corbit at 12 July 2000 18:36:50:

One more note on TSCP:
If you add a piece list, it will add 200 ELO to the strength and triple the NPS the engine can produce for sparse positions.
Add better draw detection, and you would have another huge boost in ELO.
With the tiny amount of code Tom has produced, I think the results are pretty incredible.

To Dann:
What is this piece list? Can you give me an example that is as simple as TSCP?
So that it is easier to digest. Why not you write a draw detection for the tscp151 on your ftp? I will then look at the code there If that is not possible, then just give me a simple example on that too, well the program just loves to repeat positions, and I find it hard to stop it:-)
Best Wishes
Han

[Dann Corbit]

Geschrieben von: / Posted by: Dann Corbit at 13 July 2000 09:55:38:
Als Antwort auf: / As an answer to: Re: Let Dr Chess 1.30.10 participate in any tournament. To Dann geschrieben von: / posted by: Han at 13 July 2000 09:46:54:

What is this piece list? Can you give me an example that is as simple as TSCP?
So that it is easier to digest. Why not you write a draw detection for the tscp151 on your ftp? I will then look at the code there If that is not possible, then just give me a simple example on that too, well the program just loves to repeat positions, and I find it hard to stop it:-)

A piece list is...
Err...
Um...
A list of pieces. Usually a list of black pieces and a separate list of white pieces. Often, the pawns are in the list too, or even separate lists for pawns.
Don't forget that ten bishops or knights are possible and also 9 queens.
Sometimes, the king square is also stored separately.
What's it for?
You don't have to iterate over each square of the board looking for what is there and what is not. You instantly know where all the pieces are. Makes evalutation go a lot better sometimes.
That's Tom's program. I mess around with it and make custom changes but I keep them to myself (except any [if any] Tom wants to add that I suggest). If Tom wants to make such changes and post them it is up to him.
Making wholesale changes to someone else's program and posting them back to the net is a bit hokey, unless it is only minor tweaky things {to my way of thinking}.
I actually have had a few exchanges with him on these notions, but I think he just wants to keep the code simple.
Lots of programs have piece lists (most of them, actually). If you want to see one, just download a few programs and nose around.


my ftp site

[Han]

Geschrieben von: / Posted by: Han at 13 July 2000 16:20:03:
Als Antwort auf: / As an answer to: Re: Let Dr Chess 1.30.10 participate in any tournament. To Dann geschrieben von: / posted by: Dann Corbit at 13 July 2000 09:55:38:

What is this piece list? Can you give me an example that is as simple as TSCP?
So that it is easier to digest. Why not you write a draw detection for the tscp151 on your ftp? I will then look at the code there If that is not possible, then just give me a simple example on that too, well the program just loves to repeat positions, and I find it hard to stop it:-)

A piece list is...
Err...
Um...
A list of pieces. Usually a list of black pieces and a separate list of white pieces. Often, the pawns are in the list too, or even separate lists for pawns.
Don't forget that ten bishops or knights are possible and also 9 queens.
Sometimes, the king square is also stored separately.
What's it for?
You don't have to iterate over each square of the board looking for what is there and what is not. You instantly know where all the pieces are. Makes evalutation go a lot better sometimes.
That's Tom's program. I mess around with it and make custom changes but I keep them to myself (except any [if any] Tom wants to add that I suggest). If Tom wants to make such changes and post them it is up to him.
Making wholesale changes to someone else's program and posting them back to the net is a bit hokey, unless it is only minor tweaky things {to my way of thinking}.
I actually have had a few exchanges with him on these notions, but I think he just wants to keep the code simple.
Lots of programs have piece lists (most of them, actually). If you want to see one, just download a few programs and nose around.

Dann:
If you can't post the modified tscp up on the net because it is against your principles, why not mail me a copy of the one with both piece list and draw detection implemented
That would be very helpful to me, as I do not have to spend time looking for piece lists and figure out how to implement it.
With regards to your program, why not just make it moderately simple and not use so many threads :-} So that we don't have to wait so long :-]

[Dann Corbit]

Geschrieben von: / Posted by: Dann Corbit at 14 July 2000 07:17:08:
Als Antwort auf: / As an answer to: To Dann: geschrieben von: / posted by: Han at 13 July 2000 16:20:03:
[snip]

If you can't post the modified tscp up on the net because it is against your principles, why not mail me a copy of the one with both piece list and draw detection implemented
That would be very helpful to me, as I do not have to spend time looking for piece lists and figure out how to implement it.
With regards to your program, why not just make it moderately simple and not use so many threads :-} So that we don't have to wait so long :-]

Why? The world abounds with free chess source code. I have (maybe) 50-60 chess source codes. Lots of them have pondering. If I send you some pondering code, in what way will you benefit? It's only Dann's code. If you don't know how to implement it yourself you will not have gained anything. If you just want the fastest program, buy a copy of Deep Junior, Chess Tiger, Fritz, Rebel and ChessMaster. Each of those programs will beat the pants off of anything you can write in a year's time. If you want a program to be yours, then you have to figure it out for yourself. If you read my other postings you will see that I feel strongly about this. Even with revealing the source code from some author, I won't do it. I have many source codes that I have been given to inspect or code review or try to speed up or whatever, but have not been given the original author's permission to release it so I don't.
I disagree. Looking for piece lists and learning how to implement it is *the only* benefit you can gain. If you just want something fast, then buy it. If you want really fast source code to play with, UAB is only a few clicks away. Phalanx is world class, and a paultry few lines. His searching code is sheer genius. I get all tingly reading it.
People will probably have to wait even longer than for Ferret. I'm one of those horrible perfectionists. And if it is not the best program in the world but only second best after ten years of effort, it may still stay unknown.



My FTP site