Author: Moritz Berger
Date: 13:46:15 09/09/98
Go up one level in this thread
On September 09, 1998 at 15:49:40, Ed Schröder wrote: >The general interest in auto232 in low, say 2-3%. Not here in CCC of course, >here it is hot. Most of Rebel customers never heard of auto322. But I admit >I am punishing my very best and loyal customers. I guess I have to live with >that. ... I bet all the upcoming reviews will complain about this: "We would have liked to tell our readers about playing strength improvements, but we can't. Why? Because Ed Schröder wants us to do spend 300 hours (time required for about 40 tournament games) of manual testing and we simply cannot do this. In our opinion, this shows a disrespect of our work as independent journalists because Ed Schröder wants us to rely solely on his own assesment about the expected gains.". How about this? >Not true Moritz. The criticism was on SSDF not Chessbase. *My* criticism was mainly on ChessBase because they created the problem by not making the autoplayer publically available. This will change soon and resolve this problem for me. I read your complaints also as a criticism on ChessBase (you always implied that they do test Fritz against Rebel and benefit from this while you cannot do likewise), and your statements have been instrumentalised by Ossi Weiner (www.computerchess.de) without any disagreement on your part. >Well I do hope that Rebel is more than auto232 alone :) But Rebel is clearly *less* without auto232. Even MUCH less, to me. >No revenge, that is silly. Good to hear, that's a starting point to convince me to carry on this argument here :-) >Ok, I will explain. > >5-6 weeks ago I sent email to SSDF. Two points: >- please remove Rebel results from next SSDF list; >- please do not include Rebel10 on SSDF list; >The email explained in detail all my motivations. I do not think that your demand to remove *all* Rebel games from the SSDF list is justified. In fact, I believe it's very much unfair. Reasons: (1) You did HUGE advertising with SSDF results when Rebel was the top honcho. Will you also revoke these ads now??? You will look like a clear moral loser when you cash in the benefits as long as you please and withdraw when you don't like the results. Oh yes, Bobby Fischer remains world champion until he dies... (2) You do not honor the massive work the SSDF has put into testing Rebel over the years. I feel that you simply have not the right to deny them to publish the results of their work when they are - as far as Rebel's games are concerned - 95% independent of potential effects of the issue you raised as your basic complaint. Rebel didn't play against Fritz *that* often, and maybe other programs had worse results against Fritz, even helping your relative position in the group of "follow-ups". That's the nice thing about doing as many different matches as the SSDF does: The impact of one participant on the whole rating pool is small - only the final position affects all programs with worse scores by "moving them one slot down the list". (3) All other ratings are based (directly or implicitly) on games played against Rebel. It would be unfair for all other programs who scored better than average against Rebel to remove these games from the list. (4) If you cannot prove that the SSDF testing methods are biased - and you don't even claim this as far as I understand - you even could not legally have them remove Rebel from their list. >I still have no answer. Instead of that SSDF released a new list with Rebel. >I considered such an attitude towards me as the worst possible scenario one >can think of. At least I expected a 'yes' , 'no' or whatever. I agree that they owe you at least a clear rejection of your demands. >So the only guarantee I have Rebel10 not being on SSDF is to remove >auto232 from Rebel10. That's no 100% guarantee, either. Chessmaster 5000 is also on that list, but not tested against opponents where it might have scored better because the SSDF was frustrated with having to play hundreds of manual games ... I remember the bad manual SSDF results for Fritz 4 when I had at the same time excellent scores against stronger opponents it never met at the SSDF because it didn't have an autoplayer. >It's a decision for one release. I simply wait and see how others will deal >with the raised "hidden auto-player" problem and then make up my mind >again. > >- Ed - Not having Rebel listed at the SSDF is detrimental *to Ed Schröder* because: (1) people looking for strong programs on the list will never consider Rebel, because it doesn't even show up. Implication is that it is not relevant in some way with respect to the other programs listed there, and this clearly hurts your sales. (2) if somebody tests Rebel e.g. without having the main opening book enabled and gets bad results, in the past these results were suspicious because you always had the SSDF list for guidance, showing Rebel among the very best programs (this is also the case with the latest rating list where Rebel is rated only slightly behind the #1 program). Now this person will suddenly have a much higher credibility, because we will not have any rating for reference. (3) part of the bad press ChessBase got will now cast echoes on Rebel and Ed Schröder for repeating the same mistake of making no autoplayer publically available. Even worse, now that ChessBase releases auto232 compatible products, this will make them look like really good guys and you will have to take all the heat. Don't expect my condolence when you start appearing on the www.computerchess.de pages in the news section ... Hey, come on Ed, you really have nothing to hide ... Rebel's playing strength is beyond any doubt for me, don't spoil the fun by disabling the transmission of Rebel's mating moves via auto232 ... Moritz
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.