Now u answer.
Nodes
Sort:
The number of nodes analysed per second is a little like the number of positions calculated per second. It's not quite the same because modern engines use clever programming shortcuts to increase their speed. Anyway, Engines like Rybka 4 report analysing far fewer nodes per second than engines like Fritz 14. There are two key factors to weigh up which are how aggressively the engine prunes the tree of variations and how much chess 'knowledge' the engine is applying to each position.
So, all other things being equal more nodes are better, but all things are not equal, Engines with lower node counts compensate for this by having more chess knowledge and by pruning the tree of variations more aggressively to reach the same depth.
Things become more interesting with programs like Stockfish which combine light and finely tuned chess knowledge with high node counts and incredibly aggressive tree pruning.
In addition to what chyss said, one huge factor is that there are many different ways of defining node count. The only way you even have a chance of comparing node count between two engines is if the programmers are defining node count in exactly the same way, and that usually isn't the case. And the programmers usually don't bother to tell the user what his definition of node count is, and even if he did, we might not understand what it all means anyway.
EscherehcsE is absolutely right.
So engines explore nodes and I noticed that some engines explore more nodes than others. I've heard that the number of nodes explored doesn't really matter, but I'm curious. Do they really not matter? Does the more-node-exploring-engine produce better results if given more time? Or is that engine uselessly exploring bad nodes? Why don't all engines explore the same amount of nodes when run on the same computer?