Author: Chris Whittington
Date: 07:44:52 01/01/98
Go up one level in this thread
On December 31, 1997 at 14:59:39, Thorsten Czub wrote: >I will not comment on Chris captain's log stardate 1997. > >>8. I know, from my own observations with tthe CSTal book (which is not >>'cooked' in any way, but is simply enormous, and just compiled from PGN >>files without any human weightings added), that sometimes it can exit >>book, after a long line, with totally won game. it can also exit totally >>lost. So I know that these high evalaution book exits can occur without >>any human intervention to the book lines. > >The shown phenomena is different. You are wrong. >Your experience has nothing to do with the phenomena seen with mchess. >The deep lines (I would call 27 moves or more deep) >have no opening-theory in a human-game-library. >I tested this ! You tested the contents of Sandro Necchi's brain ? You looked at his analysis notes for the Sicilian Najdorf Poisoned Pawn ? You were with him when he tested variations and sub-variations on several programs ? And ? >The games were not played by any human-opening-theory >nor any human-games from database. Big deal. And this is your opinion only, not fact. >And the answers played out fit to hiarcs6 answers. >Other programs would have played different and would have thrown out >mchess much earlier out of book. So what, do you want uniform behaviour from everything ? >Why are these lines in the book ? Because they got stored as data. >Who has played them WHERE when ? Doesn't matter, its just stored data. Leanr your way around it if you don't like it. >I can tell you, because this was my point: >Somewhere on an autoplayer mchess-experimental played them out against >hiarcs6 commercial and then the winning-games were merged/added into the >opening-book. > >THIS is the thing I am talking about. So what if autoplayer games are included in the book ? If the program found the moves anyway (as it must have done) whats wrong with including the moves as straight lookups, it saves time, no ? Maybe they save the losing games as well, so as to divert away from the stored moves. Why is this a problem for you ? > >I have not get any comment from you on the specific opening-line I >posted. >No comment was made about the opening line Mchess played with black >against CSTal in Paris. No comment from Marty about the posted line. So what. Are we supposed to be answering machines for everything ? > >Instead of calling other people whatever insults, you should concentrate >on the specific data mentioned. Thank you. I do but you don't notice. >When somebody gives data you COULD answer by showing other data or >comment on the data. Precisely. I posted a detailed alternative explanation is the post to which you just replied. Ed replied to it. Jeroen replied to it. You all have in common that you snipped the explanation and did not refer to it. Ok, think your own fixed stuff. You believe it so much that an alternative view is seen as suitable only to be censored. > You try to claim a campaign against Mchess. >This is wrong. I was just asking the questions. The questions never got answered. There was a pattern in the two rounds of attack: both at Mchess release time. Both times Rebel team members were heavily involved. This time Jeroen saw fit to keep on repeating stuff about Necchi. Is looking for patterns not allowed anymore ? Instead of answers, the messenger just got attacked instead. Instead of looking at the alternative explanation, the snipping tool got used. Instead of participatory debate, the censors got appealed to. >Mchess is a strong program but from my point of view the opening book is >somehow contraproductive ! >Although these lines WORK (!!) they suggest Mchess would not win the >games without the preparation. THIS is my critics. I am sure Mchess >would do a good job without cooked book. Therefore I cannot understand >it. I'll try again. There's some data posted here recently. Mchess SSDF game results, against Genius, I believe. Roughly speaking, since I do this from memory: mchess5 beats genius3 15-5 or so. genius4 and 5 beat mchess5 15-5 or so The generally accepted argument goes like this: we all know Mchess isn't 15-5 better than Genius, so the 15-5 result is cooked, due to the books. When Genius team has a chance to fix the lines, they turn the result round 5-15. Therefore Mchess is cooking books, Mchess is weaker than the SSDF results indicate. QED. And, btw, Mchess is 'cooking', Genius is 'fixing' Alternative argument goes like this: Once apon a time some programmer did a lot of work on his book - this gave him an advantage. The other computing programming teams started working on their books too, as a counter. This process developed and continued year after year with lines being cooked, fixed and cooked again. The data 15-5 turned to 5-15 indicates *both* sides engaging in the cooking-war or fixing-war or arms-race or whatever yoiu want to call it. Mchess first, then Genius, only nobody actually knows who began it, and nobody can ever stop it once it started. Note that this alternative explantion doesn't try and apportion blame onto any one programming team. Now my gripe is this: you can call this process whatever you like, call it evil book-cooking. Call it morally correct plugging holes, call it what you like; but DON'T claim one side is 'plugging holes' and the other side is 'cooking books'. They are both part and parcel of the same process. And if you ascribe moral correctness to one side, and moral incorrectness to the other you're merely taking sides in a war; and that's the point at which questions get asked as to why. Or, extracting the 'fact' that Mchess cooks its books based on the 15-5 data, and ommitting the 'fact' that it gets countered, and presumably counter-counter-countered in a never ending spiral is just a plain cheating argument. > >Jeroen is a human beeing. Sometimes I am the same opinion he has, and >sometimes I am different opinion (I remember when he worked for Mephisto >and we met us at the Nuernberger Spielwarenmesse I was often against his >opinion ! :-). >Your construction of a campaign is not amusing. There's a pattern that goes back to last year. If they don't like the idea that this pattern could just possibly be construed as a campaign, then they should avoid making the patterns, or else have a good explanation handy. I haven't seen it yetr. >Sometimes I agree with you. Sometimes with ed. >Sometimes I disagree. There is no money and no campaign behind this >phenomena. >It has something to do with the fact that human beeings have different >opinions. >The fact that human beeings have different opinions is not reason to >insult somebody ! > I can have a different opinion than my best friend >without calling them nasty things. Really. You think going for a censorship attempt, dismissing arguments with the words blah-blah-blah, and accusations of 'disgusting' came from me ? I think you see insults where there are arguments that you disagree with. Chris Whittington >I can live with this. I can even have >different point of views than my girl-friend has. Maybe in your family >it is different. >But in my family we try to have our own opinions. And not anybody is >taking anything personal.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.