Hi all,
does move sorting matter (much) for zero window searches?
I'm asking because I'm thinking about using MTD-f, an algorithm which uses zero window searches (which yield either an upper or a lower bound) to close in on the real value (until the bounds meet).
My guess would be, that it does matter, but by far not as much?
Also another question: Since zero window searches are like blazingly fast, wouldn't it be an idea to guess the value straight away with a [guess-1, guess+1] search and run N guesses in parallel? Or would this in sum always search more than MTD-f or Alpha Beta?
Best regards,
Tobi
