There are only 3 ELO separating Loop, Naum and Fruit in positions 3-5 on the list
Alex's EGTB fix came too late because I was already more than half way into my testing. So I had to just continue with the original exe for consistency. Naum was not accessing tablebases during this testing.
Hermann 2.0 is underway now, and when that is done it may also be time to test another version of Movei.
Tony Thomas wrote:I wonder if it is that much stronger in normal chess. I guess I would buy it if it is indeed more than 50 points stronger than previous version.
It is unclear
based on the cegt the difference to naum2.1 is estimated to be 42.1 elo but it is only for the 2 cpu version.
Tony Thomas wrote:I wonder if it is that much stronger in normal chess. I guess I would buy it if it is indeed more than 50 points stronger than previous version.
It is unclear
based on the cegt the difference to naum2.1 is estimated to be 42.1 elo but it is only for the 2 cpu version.
CEGT are testing with generic opening books while Naum has an excellent own opening book,I don't trust their results
_No one can hit as hard as life.But it ain’t about how hard you can hit.It’s about how hard you can get hit and keep moving forward.How much you can take and keep moving forward….
The answer I think is that Naum 2.1 underperformed here compared to standard chess, plus it had no FRC specific code. Alex tells me that 2.2 does have some FRC specific code, and that may have made a difference.
One also has to wonder, if Naum 2.2 was able access tablebases, how much of a gain that would have been
Small correction. Naum 2.2 doesn't have any FRC specific code.
There is one evaluation change made for normal chess that I think is very good for FRC too, but I didn't confirm this in test matches.
Both Naum 2.1 and 2.2 played without tablebases, so comparison is still valid.
I think that maybe Naum 2.2 was able to win more games in middlegame, so it was less disadvantaged by the lack of EGTB access then 2.1
My test matches indicate that its main strength is in the middlegame.
I would guess that ELO difference between 2.1 and 2.2 is around 50 ELO.
I was slightly disappointed with Naum 2.1's performance originally, feeling it should have done better. Still, against this range of opponents, and playing FRC, that was the result, with a large number of games. So under these conditions, playing FRC, and with a large number of games, +74 ELO it is. I guess we will have 40/4 results for standard chess fairly soon, so then we will see what we get there
Dr.Wael Deeb wrote:
CEGT are testing with generic opening books while Naum has an excellent own opening book,I don't trust their results
We also test with generic opening books for "normal" chess.
Don't get offended Graham,you know my opinion regarding this subject,testing a wild variety of engines with a big variety of opening books will give a big number of combination engine-opening book and will result in wild estimation of the chess engine's strength tested....
A strong chess engine is a combination of own opening book and the engine itself
_No one can hit as hard as life.But it ain’t about how hard you can hit.It’s about how hard you can get hit and keep moving forward.How much you can take and keep moving forward….
Dr.Wael Deeb wrote:A strong chess engine is a combination of own opening book and the engine itself
Which is totally irrelevant for those who want to use an engine for analysis. Many people want to know how strong an engine is by itself, without book, and we are trying to provide such information.
I estimate that very few people use engine with own book. Most of people don't play with an engine set to full strength, but use it to assist their games or preparation. In those tasks people often use huge game databases in the opening, not an engine book tuned to particular engine's style.
