Author: Dirk Frickenschmidt
Date: 15:37:52 05/04/98
Go up one level in this thread
On May 04, 1998 at 11:22:28, Enrique Irazoqui wrote: Thanks Enrique, this is plain facts without too much emotions! I completely agree with you. Just because I doubt anything will change without new ideas, I want to add one. The idea just came to my mind, and I have not yet have long thoughts about it, so look at it 'cum grano salis': How about a) all programmers requiring standard hardware for their programs as you proposed below (now perhaps still 200MMX with 64Mb, in future perhaps PII-300 or something; whatever: common for all) b) all programmers providing an autoplayer which will have a kind of time stamp in it for normal users, thus only allowing autoplay after let's say a two months period after installing. Only exeption: SSDF-version. Thus the SSDF testers could have their autoplay testgames for a certain period in which *no* company could use the autoplayer for outbooking tests. After finally having the common version work plus testing time (so after about half a year) outbooking procedures would not be of much use any longer, because soon after (about once a year) a new version of the program would appear, with new (at first not common) autoplayer. This might save enough time to avoid extreme forms of outbooking and still aloow us coomon folks to autoplay after two months: having fun and proving if the rating was ok all in all. This is just *one* *possible* example of new ideas. So nobody should be angry telling me why *this* one might not work or not be useful, better just make more proposals leading away from the present situation which is a dead end... Kind regards from Dirk >On May 04, 1998 at 08:12:23, Jouni Uski wrote: > >>From different SSDF posts I get the feeling, that probably SSDF will >>stop their ratinglist service soon! And persons like Ossi Weiner and Ed >>Schröder are speeding up this by their childish attacks against SSDF! >>Additionally also some SSDF person have lost their computer chess >>interest (I am myself as interested as 15 years ago!). >>Let's hope that SSDF still manages at least two additional years! And >>they need our all support... > >If it is true that some new programs will not be available to the SSDF >with auto232 capabilities, then the SSDF list will be in deep trouble. >Without Genius, Rebel, Nimzo and Shredder, it won't be very meaningful. > >It all started February 22nd, when Fritz 5 became first with a 55 points >lead over the second. Ever since, we heard: > >1 - The non available autoplayer Fritz 5 cheated, and that's why manual >and autoplayed results were claimed to be different. Some demanded that >autoplayed games would not be counted for rating purposes. > >2 - The engine in the autoplayer version of Fritz 5 is not the same as >the commercially available. > >3 - The autoplayer Fritz 5 works only with the Powerbook and on machines >with at least 64MB ram. Other programs have no requirements of this >kind, and therefore testing conditions are unfair. > >4 - The fact that ChessBase can tune their engines and books against the >opposition without allowing the other programmers to do the same, gives >ChessBase an obvious advantage. > >As a consequence, some programmers will refuse to deliver auto232 >versions of their new programs to the SSDF. > >It has been proven by Ed Schröder that the autoplayer in Fritz 5 is >clean. By replaying the games posted by Tony Hedlund it is easy to >verify that the engine tested by the SSDF is identical to the commercial >Fritz 5. From all the people that claimed otherwise and accused >ChessBase of cheating and the SSDF of going along with it, I would have >expected an honest apology as public and enthusiastic as their prior and >false accusations. It never happened and I very much doubt it will. In >fact, they still demand that games autoplayed by Fritz 5 are not >counted. > >As for point 3), it is the first time a program has this kind of >requirements. I find them quite sensible and I think other programmers >should also make sure their programs are not tested on substandard >hardware. Right now, this gives some advantage to Frizt 5, but in my >opinion the way to go is not shooting CB from making sure their program >is properly tested. > >Point 4) is, I think, the troublemaker, the only real one. I see two >ways to solve it: >a - ChessBase releases the autoplayer Fritz 5. >b - Other programmers follow the example and release auto232 versions >only for the SSDF. > >I think a) is a much better option because not only solves the problem, >but it allows all of us to autoplay all programs. Nevertheless, once >manufacturers made a point it seems metaphysically impossible for them >to modify it. I don't know if this is a reflection of the secretive, >pathologically untrusting ways of the chess computer world, but one >thing is for sure: the SSDF may very well end up being the victim of all >this. Meanwhile, everybody seems to wait for everybody else's move. >Wait, do nothing and let the SSDF go into deep trouble. > >Enrique
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.