Re: First post (and FailHigh question!)
Posted: Wed Jun 19, 2013 8:57 am
Hi Chan,
You get a hash hit when you are searching THAT position. And you get and backup the score of infi- 5. That's all. But in the parnet node, you will adjust the score to infi-5-12, infi-17, and that score is still above beta, so another cut-off occurs, and if a cut-off occurs you must store that position/score/move in the hash. In that way you will store a inf-17 score. And next time you will encounter this position again, a probe in THAT position will result in a hash hit of a score of infi-17 and you'll backup immediately that score as well, and the parent will adjust the score to infi-17-15,infi-32, and as it is still above beta, another cut-off will occurr, causing another store in the TT with a score of infi-32 and so on... Got it?Chan Rasjid wrote:Hello Natale,
If at the next iteration at a node C which has the same position as B (which is also that of A), then again the node will not be searched as there will be an exact TT hit. It will still be probing the same TT entry which B got and it would be the same TT entry that was hashed when node A was search: infi - 5. So the retrieved TT exact best-score that C has will be infi - 5 - 15, or infi - 20; So search immediately returns the score infi - 20. This contradicts your score of infi - 32.xmas79 wrote:Hi Chan,Chan, you're not thinking fourth dimensionally! It's easy to understand those mate scores as well: what if at next iteration (say depth 16) you hit the same position at a node C you already had at node B while searching ply 15? Since you got a cut-off early and stored a score of infi-17 now you will get that score-15, infi-17-15, infi-32! And you will store that score as well... And the more next iteration you'll get another cutoff getting a score of infi-32-ply_i_will_hit_this_position_again. In this way at ply 20 I can have a MATE76 score...Chan Rasjid wrote:There is something which I still cannot understand with Natale's output. From what I now understand, at depth 20, the +mate score that search may have is limited to infi - 19 - 19, or infi - 38; ie MATE19. So I don't understand why the output has scores like MATE76.
The contradiction arises from the contradiction of what you said earlier with what is stated in this quote. In an earlier reply, you mentioned that your search returns straight from any TT cutoff without rehashing the same position. But now, what you say is "you got a cut-off early and stored a score of infi-17". Storing infi - 17 in the TT for node B is wrong as it is mate-in-3 from node B; ie infi - 5.
My analysis here may also be wrong just as how I traveled to the moon and back
Best Regards,
Rasjid.