list of deterministic engines at fixed nodes,fixed depth

[Uri Blass]

There are engines that are deterministic when you use fixed nodes or fixed depth and there are engines that are not deterministic.
I found that Wasp is not deterministic with fixed nodes and it seems that both stockfish developement version and Dragon2.5 are deterministic with fixed nodes(of course when I use a single core for every engine and the same fixed hash tables).

I think that it will be nice to have details of deterministic engines in order to have some deterministic rating list that is not based on time control so everybody can reproduce the games.

participants in the rating list basically can be engine X with 100000 nodes per move or engine Y at fixed depth of 10 plies.
The list is not to compare strength between engines but to have no problem that is result of the fact that different people use different hardware for the list or because in some game an engine is slowed down by a significant factor because the computer did other things at the same time
and have some fair rating that you can easily check for deterministic opponents.

[CMCanavessi]

There are engines that are deterministic when you use fixed nodes or fixed depth and there are engines that are not deterministic.
I found that Wasp is not deterministic with fixed nodes and it seems that both stockfish developement version and Dragon2.5 are deterministic with fixed nodes(of course when I use a single core for every engine and the same fixed hash tables).

I think that it will be nice to have details of deterministic engines in order to have some deterministic rating list that is not based on time control so everybody can reproduce the games.

participants in the rating list basically can be engine X with 100000 nodes per move or engine Y at fixed depth of 10 plies.
The list is not to compare strength between engines but to have no problem that is result of the fact that different people use different hardware for the list or because in some game an engine is slowed down by a significant factor because the computer did other things at the same time
and have some fair rating that you can easily check for deterministic opponents.

But that would also negate each engine's own speed optimizations, which seems unfair too.

[Uri Blass]

There are engines that are deterministic when you use fixed nodes or fixed depth and there are engines that are not deterministic.
I found that Wasp is not deterministic with fixed nodes and it seems that both stockfish developement version and Dragon2.5 are deterministic with fixed nodes(of course when I use a single core for every engine and the same fixed hash tables).

I think that it will be nice to have details of deterministic engines in order to have some deterministic rating list that is not based on time control so everybody can reproduce the games.

participants in the rating list basically can be engine X with 100000 nodes per move or engine Y at fixed depth of 10 plies.
The list is not to compare strength between engines but to have no problem that is result of the fact that different people use different hardware for the list or because in some game an engine is slowed down by a significant factor because the computer did other things at the same time
and have some fair rating that you can easily check for deterministic opponents.

But that would also negate each engine's own speed optimizations, which seems unfair too.

The target is to have a fair rating for deterministic players.
Not to decide which engine is better and which engine is worse.

I know that nodes of different engines are not equivalent but also hardware of people are not equivalent and I want to have some correct data about rating of players and not about strength of engines.

Strong engines are going to be on top assuming people do not prefer some weak engine in the rating list by giving it some nodes that mean giving it significantly more time.