How powerful is the Queen compared to 2 Rooks ?

Discussion of computer chess matches and engine tournaments.

Moderators: hgm, Harvey Williamson, bob

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Chessqueen
Posts: 564
Joined: Wed Sep 05, 2018 12:16 am
Full name: Nancy M Pichardo

Re: How powerful is the Queen compared to 2 Rooks ?

Post by Chessqueen » Tue Feb 05, 2019 6:36 pm

hgm wrote:
Mon Feb 04, 2019 3:15 pm
Chessqueen wrote:
Mon Feb 04, 2019 2:23 pm
Thanks for making me realize that it is not how fast a machine with Komodo or SF10 can calculate, but according to what I think I read is that Alpha Zero has done thousand and thousands of training hours and analyzed billions of positions under those training and it is capable of retaining those positions in case it comes again in one of its future game. Therefore, LCO will be stronger than Komodo or SF10 very soon.
That is also completely wrong. AlphaZero cannot retain any positions. It just has a far more advanced (but far slower) evaluation than SF10 / Komodo. And not significantly more training games have been used for tuning it than have been used for tuning Komodo or SF10. After 4 hours of training AlphaZero had already reached the level of Stockfish. That was done with many machines in parallel, though. But that is what fishtest does too.
Alpha zero does NOT necessarily have a more advanced evaluation than Komodo as I will show you in this game analysis, it just has its own evaluations algorithm. I believe that a much stronger engine can be develop by using a combination of Neural Networks combined with either Alpha-beta Search or Pruning.
Now to show you that the evaluation of Leela or LC0 necessarily is not much better than Komodo both engines gave a different evaluation of the possible outcome of this game in particular, but at the end the game ended in a draw.
https://www.youtube.com/watch?v=AAnkucMsL0M

At the end of this position Leela evaluated its positions as 1.34 whereas Komodo evaluation was 0.00 here is the position in question =>
[D]8/8/6BB/2n5/3bp3/1kp2PP1/4K2P/8 b - - 0 1

Chessqueen
Posts: 564
Joined: Wed Sep 05, 2018 12:16 am
Full name: Nancy M Pichardo

Re: How powerful is the Queen compared to 2 Rooks ?

Post by Chessqueen » Tue Feb 05, 2019 7:13 pm

Chessqueen wrote:
Tue Feb 05, 2019 6:36 pm
hgm wrote:
Mon Feb 04, 2019 3:15 pm
Chessqueen wrote:
Mon Feb 04, 2019 2:23 pm
Thanks for making me realize that it is not how fast a machine with Komodo or SF10 can calculate, but according to what I think I read is that Alpha Zero has done thousand and thousands of training hours and analyzed billions of positions under those training and it is capable of retaining those positions in case it comes again in one of its future game. Therefore, LCO will be stronger than Komodo or SF10 very soon.
That is also completely wrong. AlphaZero cannot retain any positions. It just has a far more advanced (but far slower) evaluation than SF10 / Komodo. And not significantly more training games have been used for tuning it than have been used for tuning Komodo or SF10. After 4 hours of training AlphaZero had already reached the level of Stockfish. That was done with many machines in parallel, though. But that is what fishtest does too.
Alpha zero does NOT necessarily have a more advanced evaluation than Komodo as I will show you in this game analysis, it just has its own evaluations algorithm. I believe that a much stronger engine can be develop by using a combination of Neural Networks combined with either Alpha-beta Search or Pruning.
Now to show you that the evaluation of Leela or LC0 necessarily is not much better than Komodo both engines gave a different evaluation of the possible outcome of this game in particular, but at the end the game ended in a draw.
https://www.youtube.com/watch?v=AAnkucMsL0M

At the end of this position Leela evaluated its positions as 1.34 whereas Komodo evaluation was 0.00 here is the position in question =>
[D]8/8/6BB/2n5/3bp3/1kp2PP1/4K2P/8 b - - 0 1
Going back to who has a better evaluation Leela or Komodo? On this next position Leela or LC0 evaluate its position as 1.53 winning, whereas Komodo evaluate its position as 0.00 ==>
[D]8/8/8/2nB4/3b4/5PP1/4K2P/2k5 w - - 0 1
https://www.youtube.com/watch?v=AAnkucMsL0M

User avatar
Guenther
Posts: 2958
Joined: Wed Oct 01, 2008 4:33 am
Location: Regensburg, Germany
Full name: Guenther Simon
Contact:

Re: How powerful is the Queen compared to 2 Rooks ?

Post by Guenther » Tue Feb 05, 2019 7:50 pm

hgm wrote:
Tue Feb 05, 2019 2:49 pm
It seems to me that being fluent in even a hundred programming languages would not make up for the inability to trust others. So no, it is of no interest to me (or anyone else, for that matter) what programming languages they taught you.
It is very clear from a bunch of thousands posts already, that the pichy entity is only fluent in hundred ways of lying and fantasizing...
No idea why he is allowed to have at least 3 troll accounts already in the current forum?

search.php?author_id=8323&sr=posts
search.php?author_id=239&sr=posts
search.php?author_id=7080&sr=posts
https://www.stmintz.com/ccc/index.php?t ... d&search=1
Current foe list count : [91]
http://rwbc-chess.de/chronology.htm

Dann Corbit
Posts: 9901
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: How powerful is the Queen compared to 2 Rooks ?

Post by Dann Corbit » Tue Feb 05, 2019 7:54 pm

Chessqueen wrote:
Tue Feb 05, 2019 3:08 pm
hgm wrote:
Tue Feb 05, 2019 2:49 pm
Ah, so it is just old-fashioned paranoia? You know that the AlphaZero article says it started with nothing but the rules, but you refuse to believe it, because in your world view they are all lying bastards at Google/Deep Mind, and whatever they publish could not possibly be true? Is that it? And despite all the computer languages you master, you did not bother to check out the Leela Chess Zero code, to find this hidden initial knowledge you imagine to be there?

It seems to me that being fluent in even a hundred programming languages would not make up for the inability to trust others. So no, it is of no interest to me (or anyone else, for that matter) what programming languages they taught you.
You still do not get the point. When you stated about having knowledge, you are confusing knowledge with information about certain subject or news. If a person did not have time to read all the information about any given matter or subject it does NOT make that particular person, a Zilch or un knowledgeable, it just simply make that person un informed.
The LC0 project is using the same approach, if I am not mistaken.
Therefore we can conclude that the approach does in fact work. And very, very well at that.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.

User avatar
hgm
Posts: 23506
Joined: Fri Mar 10, 2006 9:06 am
Location: Amsterdam
Full name: H G Muller
Contact:

Re: How powerful is the Queen compared to 2 Rooks ?

Post by hgm » Tue Feb 05, 2019 8:46 pm

Chessqueen wrote:
Tue Feb 05, 2019 6:36 pm
Alpha zero does NOT necessarily have a more advanced evaluation than Komodo as I will show you in this game analysis, it just has its own evaluations algorithm. I believe that a much stronger engine can be develop by using a combination of Neural Networks combined with either Alpha-beta Search or Pruning.
Now to show you that the evaluation of Leela or LC0 necessarily is not much better than Komodo both engines gave a different evaluation of the possible outcome of this game in particular, but at the end the game ended in a draw.
https://www.youtube.com/watch?v=AAnkucMsL0M

At the end of this position Leela evaluated its positions as 1.34 whereas Komodo evaluation was 0.00 here is the position in question =>
[D]8/8/6BB/2n5/3bp3/1kp2PP1/4K2P/8 b - - 0 1
Well, as you seem bent on advertizing your ignorance and lack of understanding:

Evaluation = static evaluation. What you show here are search scores, not evaluations. 'More advanced', doesn't necessarily mean 'better', because that also depends on how well it is tuned, and not just on the number of characteristics that are taken into account. LC0 is still learning. And it certainly doesn't mean 'more accurate in every conceivable position'. It is already sufficient to be better on average to have a higher Elo and beat your opponent.

Any more embarrasing nonsense you want debunked? :roll:

Dann Corbit
Posts: 9901
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: How powerful is the Queen compared to 2 Rooks ?

Post by Dann Corbit » Tue Feb 05, 2019 9:05 pm

There is nothing wrong with being wrong.
See Ted's signature.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.

User avatar
hgm
Posts: 23506
Joined: Fri Mar 10, 2006 9:06 am
Location: Amsterdam
Full name: H G Muller
Contact:

Re: How powerful is the Queen compared to 2 Rooks ?

Post by hgm » Tue Feb 05, 2019 10:23 pm

"No, but being wrong and pigheadedly insisting that you are right might be a different matter altogether...

Dann Corbit
Posts: 9901
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: How powerful is the Queen compared to 2 Rooks ?

Post by Dann Corbit » Tue Feb 05, 2019 10:34 pm

Insisting that you are right when you are wrong is clear evidence that you do not understand the relevant information.
Of course, there is also such a thing as refusing to be taught.
However, it is not possible, so far as I can see, to tell which is which unless the offending party states clearly that they refuse to learn.
Therefore, I always try to give the benefit of the doubt.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.

Chessqueen
Posts: 564
Joined: Wed Sep 05, 2018 12:16 am
Full name: Nancy M Pichardo

Re: How powerful is the Queen compared to 2 Rooks ?

Post by Chessqueen » Wed Feb 06, 2019 4:41 am

hgm wrote:
Tue Feb 05, 2019 8:46 pm
Chessqueen wrote:
Tue Feb 05, 2019 6:36 pm
Alpha zero does NOT necessarily have a more advanced evaluation than Komodo as I will show you in this game analysis, it just has its own evaluations algorithm. I believe that a much stronger engine can be develop by using a combination of Neural Networks combined with either Alpha-beta Search or Pruning.
Now to show you that the evaluation of Leela or LC0 necessarily is not much better than Komodo both engines gave a different evaluation of the possible outcome of this game in particular, but at the end the game ended in a draw.
https://www.youtube.com/watch?v=AAnkucMsL0M

At the end of this position Leela evaluated its positions as 1.34 whereas Komodo evaluation was 0.00 here is the position in question =>
[D]8/8/6BB/2n5/3bp3/1kp2PP1/4K2P/8 b - - 0 1
Well, as you seem bent on advertizing your ignorance and lack of understanding:

Evaluation = static evaluation. What you show here are search scores, not evaluations. 'More advanced', doesn't necessarily mean 'better', because that also depends on how well it is tuned, and not just on the number of characteristics that are taken into account. LC0 is still learning. And it certainly doesn't mean 'more accurate in every conceivable position'. It is already sufficient to be better on average to have a higher Elo and beat your opponent.

Any more embarrasing nonsense you want debunked? :roll:
If it was Komodo search scores indicating that it had an advantage of 1.53 but failed to win, you would not have said it it okay that is my Baby that it is still learning it might need another 5,000,000 tp 10,000,000 games until it mature to be a grown up. You will simply say to Larry come up fix it! so you admit it it is NOT accurate in every conceivable position. :shock:

It is also embarrassing to misspell the word embarrassing :mrgreen:

User avatar
hgm
Posts: 23506
Joined: Fri Mar 10, 2006 9:06 am
Location: Amsterdam
Full name: H G Muller
Contact:

Re: How powerful is the Queen compared to 2 Rooks ?

Post by hgm » Wed Feb 06, 2019 9:31 am

Do you get a kick out of looking stupid, or what?

For one, there is nothing wrong with a score of +1.53 and failing to win: the draw margin in Chess is about 1.5 Pawn, and +1.53 thus means the position offers about a 50-50 chance on win / draw. I wouldn't be surprised if the 0.00 score of Komodo in reality came from using EGT (which LC0 doesn't, and then wasn't really Komodo's own evaluation of the position as well). Komodo's score is obviously wrong here: this is not a dead draw, and white has the better chances with imperfect play. So you'd better call up Larry and ask him to fix it. Even whether an engine will eventually be able to see how the game ends with perfect play is a search matter, and not an evaluation matter.

You also seem to mix up Alpha Zero, which was fully trained, and to which my earlier remarks referred, and LC0, which is still in training.

Post Reply