NAUM 4.2

[naum]

It all depends on the choice of opening lines.
In my testing Naum is beating Rybka in the first half of Noomen test set (e4 openings). Then Rybka turns the tide in the other half (d4 and c4).

Results against the Stockfish are opposite. Here Naum is equal in score in the first half of Noomen and then pulls up in the second half.

It would be interesting to see statistics for openings in your test games.

[naum]

To be more precise.
Naum is doing very good against Rybka in e4,e5 and e4,c5 openings.
Then it gets crushed in Caro-Kann It's so bad in e4,c6 that I have it disabled in my tournament book for years now.

[Werner]

Hi Alex,
I send you the games: Naum - Rybka as I am not sure how to make it. perhaps this way?
Naum - Rybka
A 2-2
B 7,5-6,5
C 9,5-4,5
D 5,5-4,5
E 4-4

PS: with other openings interim result Naum 4.2 x64 4CPU - Rybka 3 x64 2CPU = 10-14
I think other CEGT Testers will make some more games till next Sunday.

Werner

[naum]

Many people think that using Ippo code/ideas (whatever) is an easy way to improve the engine.
Trust me, it's not.
For instance, it took me 6 months to make history pruning work in Naum (long time ago when Fruit introduced it first).
Ideas/code from other engines almost never work at first in your own. You have to put A LOT of effort to make them work. And I am not sure how much time I will have to spend on it.
And one more thing. I guess I'm stupid enough to admit that I will look and use Ippo ideas. But I you can safely bet your life on that there is no commercial author who will not do that.

[Marc MP]

I prefer this kind of modesty (aka. I may learn from some else program), than the opposite (aka. it is all from me!).

Alex: Don't forget to thank the guy who invented the wheel in your readme!

[Martin Thoresen]

Alex,

I am really impressed by your professionalism and nice attitude.

Also, good luck with Naum in my upcoming long time control swiss tournament: the Cutting Edge Swiss Open!

Best Regards,
Martin

[naum]

Based on the games you sent me against Rybka, Naum did almost all the damage in e4,e5 games (6 wins and 2 losses).
This is consistent with my tests, but it will be more interesting to see openings score in the less-lucky match

[Uri Blass]

I bought Naum and tried to download the rar file and had problems.
If I click download I only get back to the same email with no download and
If I try to do save target as then it save it as a mail with smaller size and does not save the right file(it seems to save the mail that I get and not the file)

It seems that the problem is not with the programmer because even if I send to my self .rar file I cannot download it.

I did not download .rar file that I got by email in the last months but in the past I had no problem to do it.

I thought that the problem may be with my computer but I tried to download it from a different computer with no success.

I wonder if the problem is with gmail and if people can download .rar file today with gmail.

I found that
I can download .txt file with no problem with gmail so downloading naum as .txt file and renaming it can solve the problem(in case that Alex Naumov send me a renamed naum to .txt that I need to change to .rar) but I wonder if it is possible to change my setting to allow downloading .rar file directly.

[mcostalba]

Many people think that using Ippo code/ideas (whatever) is an easy way to improve the engine.
Trust me, it's not.

Agreed 1000% !!!!!!

I know for sure

And one more thing. I guess I'm stupid enough to admit that I will look and use Ippo ideas. But I you can safely bet your life on that there is no commercial author who will not do that.

I already admired you a lot, and because you have openly said you will look in Ippo I admire you even more: this open and transparent statement makes you a great guy !

Of course everybody with economic interests in chess engines had _already_ looked in Ippo sources, if they didn't they have no professional attitude toward their job. If their engine still struggles to gain ELO it means that just blindly copy ideas doesn't work, and we knew this already

[govert]

With risk of hijacking the thread, what _are_ the most interesting ippo* ideas?

Maybe this already has been discussed in another thread?