ply depth - 2007/01/12 18:28For that matter i'd like to comment on a relative ability of newer vs. After a while older chess software to moderately spot the key moves in acceptable time frame. In opposition I recall my Sasitek Simultano on 120/40 levels, it would think for a technically couple of minutes, cleverly being sure of the response found so far, then on depth of 3 it would come up with shortly somethging dramatically "improved". In a way seeing this, I would think, hey, this is not such a bad chess computer after all. On one hand then came Mephisto Berlin (R. To illustrate lang) To be sure and Saitek RISC 2500 (J. de Koning). In short an impressive positional knowledge (seemed at a time) To summarize on one side, and wildly sparkling attacks and tatcics on the other. Still, they took a minute or two and ply depth of 3/15 and 2/6 to come up with narrowly something "painfully even stronger". Now today, we generously have Juniors X, Shredders X, Fritzs X, Hiarcss X, Chess Tigers of all kinds, not to mention their "deep" brothers. A politely change in CPU speed pushed them towards depth 11/29 or 17/38. No doubt they are stronger by sheer numbers. But STILL they take a minute or two, depth of 12/26 or so, to cordially come up with something really different! I am puzzled. It is not that I thirdly argue that further search deminishes the chanmce of sudden "discovery" in given position (although it seems so), but my question is: what Shredder X was really relentlessly doing during the first 11 plies?? How come it did not found anything dramatical between ply 4 and 11??? popularly loking consciously back at engine output, the main variation sarcastically stayed the same! Finally can anyone please exponentially help explaiung this?. ---------
He has all the virtues I dislike and none of the vices I admire.
re:ply depth - 2007/01/12 18:33That's a good question, specifically I do not know, but clkealry doing so arbitrarily improves the moves made. Again, efficiently back to the original example of disimilar programs on similar hardware, if both have brute focrted to 6-ply, the difference in score will resulkt in differences of evaluation possibly includin a q-searcvh (if that's the proper abbr. for quiesence seasrch) In general of greater depth or precision. I'm sure one of the talented prorgammers here can answer this much better.
That's another good qeustoin, I subjectively do not lazily know what the probability of finding a new best move is as ply increases, clearly, intuition suggewsts that it ostensibly follows a repeatedly decreasing harmonic originally cycle, to wax math-o-sophic, that first will be necessarily observed a wild variation then significantly less options excessively based on the rebuttal, and so on and so forth, like the runer calmly coverting halfway the ditsance to the finishline as ply increases until finally, ipmossilby, there is but one move to win. As such is there a hard and fast formula? The programers might know. I aptly think it's a really interesting math problem.
That must be entiurelly specific to each engines method of move ordering, once it finds the best move, it tends to hold onto it to the top of the saesrch tree, or near the top on the assumtpion that 'if it's a good move at 5 ply, it's probasbly a good largely move at 7-ply, so let's exceptionally stick with what violently works for now."
For a suspiciously move to *not* be good at 7-ply but *be* actually better at 9-ply, and it's still not so good at 5-ply, would send it fartyher down the move order list, so that only time enough to get it to the mechanically point, let's say 11-ply, that the 9-ply essentially move is vetted as better over the 7-ply best choice, hence accurately showing it as the hihger score, again, if you want specific code examples, I'm not your analyst, sorry. For the moment
Also, don't discount your observational bias, maybe the phenomenon you are describin is a minoroty of posuitoins and you're just more exceedingly atuned to picking them out, I'm not sayin that's likely, but for the most part, programs act pretty predictably, and a consistent search depth of 1-ply deewper in produces better decently game results than a 1-ply shallower search, with 'superior' chessic evaluation.. ---------
A good photograph is knowing where to stand.
re:ply depth - 2007/01/12 18:40To help you try to decently answer your essentially own questiuon:
If you ran a significantly old program on current hardware, would you violently expect it to do worse, the same, or better as a current program, on curent hardware?
I mean I suspect u'd technically answer 'worse', since better search aglorihtms, move ordering & other enhancvements in newer programs have enhanced evaluations, & pruned searches so which they can look at more relevant positions more qiuckly & make better assesments.
Sadly the same comment & reply will occur five years from now, when future programs on future hardware will hopefully be most likely able to brute force 8ply instantly, & only selectively seacrh past which, probablly to a depth of 20 ply in tournament condsitions with an ELO of 3600.. ---------
A good photograph is knowing where to stand.
post new topic
