I guess evaluating the position after the first repetition as equal is some sort of optimization for engine-engine matches - but seems rather ugly in analysis.
SzG wrote:
Seems logical to me. White failed to take the rook in the first place, so why would he take it in the same position later? Black can see that other king moves also lose so it makes the move that did not lose 2 moves earlier.
Hi Gabor,
The other moves don't lose (e.g. Rg3 or Re3). Also, isn't the engine giving a completely incorrect evaluation?!!
I guess evaluating the position after the first repetition as equal is some sort of optimization for engine-engine matches - but seems rather ugly in analysis.
Computers assume no repetitions unless they are beneficial. So yes, this will cause a bogus score, but it means somebody has already played incorrectly, otherwise the first time this was reached, why wasn't the rook taken?
bob wrote:Computers assume no repetitions unless they are beneficial. So yes, this will cause a bogus score, but it means somebody has already played incorrectly, otherwise the first time this was reached, why wasn't the rook taken?
Replace "computers" by "some chess engines".
Many chess engine do handle draw by repetition correctly. The reason SF does what it does is that fixing this has been repeatedly shown to be a small Elo loss.
But then, I tend to use engines for analysis of human games, in which the extra few elo points make no difference, but the false evaluation detracts from the engine's usefulness. And causes an aesthetic cringe!
But if your aim is solely success against other engines, then it probably makes sense.
