Rybka node speed: the truth?

[CThinker]

yes, that is clear and we know that, but if they said different it seems nothing to do with normaly computerchess, it is only like a fake node outputs or randomly calculations to hide something.

Yeah, it does make one think so.

I'm not big into conspiracies, so, I give Vas the benefit of the doubt. If that is how we wants to count nodes, then so be it.

Btw, for comparison, human chess players make very low node counts.

[bob]

Do you guys count all the nodes in the tree your engine explore, or only the terminal leaves? My engine counts them all. Maybe Rybka only counts the evaluated terminal leaves? (so the counter might be in eval() instead of the search routines).

All. That is the definition of a classic node in AI textbooks. Two nodes are connected by an arc. The nodes represent positions, the arcs represent moves. A move takes you from position A to position A'...

[bob]

But why does it matter if Rybka counts its nodes differently? Node counts only make sense for the program itself and cannot be compared with other programs.

The node count of Crafty cannot be compared with the node count of Stockfish.

The only comparison that makes sense is the node count of the same program on different machines. This gives an indication of the relative speed of machines. You can do this comparison with Rybka's node count on different machines.

The node counts _can_ be compared, and the result is quite useful/informative. For example, how many nodes does it take a program to find move X in a tactical position. That gives an indication of how well (or poorly) it is pruning/reducing non-important branches.

If a program consistently finds a correct move searching only 1/2 the nodes I search, that gives me something to look at. It might be that to find the solution in 1/2 the nodes, each node will be twice as expensive (or worse) but it does lead to investigation and perhaps a good discovery...