Oscar L wrote:Fritz 11 has gained around 60-100 elo points in the different rating lists when you compare it with Fritz 10. AFAIK Fritz 11 has only been tested with default settings in these lists, so we can not be sure of the impact of null move parameter. I suppose that if the engine author decided that it was good, must be because it allowed to gain elo points in the internal testing before the release, and as you say it is because is looking deeper. Rybka was too far away!
If you noticed that Fritz 11 missed a lot of positions with zugzwang, sure it is because of nullmove parameter. Again, it would be quite interesting that you run your test with the option uncheched to compare the results, but of course it is up to you
In the wikipedia there is a good explanation, I think, about this, but perhaps Uri or other engine developer can comment something more:
Fritz is using null move pruning for a long time.
I doubt if testing of not using null move was done in the last years.
All top programs use null move pruning because they care about games and not about test positions.
Fritz10 is also using null move pruning and I believe that all Fritz versions from Fritz3 do it.
The reason that Fritz11 cannot solve zugzwangs that fritz10 solves is probably different than adding null move pruning(for example it is possible that Fritz11 is simply more aggresive in null move pruning and when Fritz10 have some exceptions to the rule then fritz10 have less exceptions).
In other words not using null move pruning is not going to help Fritz11 and I believe that telling fritz not to use null move pruning only in specific cases in the endgame is impossible.
There is a parameter now on Fritz 11, in fact it is nearly the only one, so you can check and uncheck it and run tests positions or games with or without null move. This option was not available in fritz 10, I do not know in previous versions.
You must be right when you say that fritz has used null move from a long time, and now it is too aggresive, this must be the reason that explains that the option is available now. I do not know if unchecking it restore Fritz 10 null move exceptions or disable null move completely.
These games between engines in endgames are very interesting but in this case more parameters are considered: speed of calculation, gestion of time and use of tablebases.
Tout le monde peut se tromper. Souvent les mots en -tion c'est les mêmes en français et en anglais: emotion, superstition, evolution, proportion, provocation...
My conclusion is Fritz 11 is weaker without null move. 64,3 points without null move and 66,1 with null move.
Stopping null move option is not a good idea unless you think there is a zugzwang in the position.
Surprising and disappointing results. You have demonstrated that fritz 11 is not always better than fritz 10, and indeed it is a quite different engine.
So if f11 with null move unchecked, gained points solving zugzwang positions but in the end got less points, I suppose that means that did not solve even more positions that were found before, can you post these unsolved positions?
A few positions unsolved by Fritz 11 (in 5 min) with null move off
[d]7k/8/pP1N4/3P1p2/5PpP/3r4/7P/6K1 w - - 0 1
1.Nf7 1-0
[d]6k1/5p2/6p1/1P1R3p/5P2/8/5KPP/1r6 w - - 0 1
1.Ke3 1-0
[d]7K/2p3Pb/4pk2/2p4P/8/4P3/8/4n3 w - - 0 1
1.e4 1-0
[d]5q2/n2P1k2/2b5/8/8/3N4/4BK2/6Q1 w - - 0 1
1.Qg5 1-0
[d]8/4B1pp/8/pp1PppP1/b3p3/P7/P3KP1P/2k5 w - - 0 1
1.g6 1-0
[d]3Q4/5pk1/4q1p1/4P2p/7P/4K1P1/5P2/8 w - - 0 1
1.Qf6 1-0
[d]6k1/5p1p/2R3p1/3Pp3/1P6/4r3/6PP/6K1 w - - 0 1
1.Rc8 1-0
Null move off decrease the pruning so Fritz 11 has to calculate too many positions. It must be why these positions are unsolved (in 5 min). I'm too lazy to try longer times.
Tout le monde peut se tromper. Souvent les mots en -tion c'est les mêmes en français et en anglais: emotion, superstition, evolution, proportion, provocation...