Author: Slater Wold
Date: 13:02:36 07/09/01
Go up one level in this thread
On July 09, 2001 at 15:52:24, Ricardo Gibert wrote: >On July 09, 2001 at 15:38:19, Robert Hyatt wrote: > >>On July 09, 2001 at 15:33:28, Ricardo Gibert wrote: >> >>>On July 09, 2001 at 15:09:48, Slater Wold wrote: >>> >>>>On July 09, 2001 at 14:22:06, Ricardo Gibert wrote: >>>> >>>>>On July 09, 2001 at 13:38:00, Slater Wold wrote: >>>>> >>>>>>Perhaps that's not a bad idea. But the simple truth is, most people play for >>>>>>that rating. >>>>>> >>>>>>Take "RebelRex" for example. His first 600 games, he playing anyone, or >>>>>>anything, rated over 2500 unrated or rated. At any timer. His rating was >>>>>>approx. 3000. He has recently changed, to only play humans, and is now 3200+. >>>>>>Why did he do this? To compete with Spitfire? Because he is only interested in >>>>>>results against humans? Who knows?! Who cares!? >>>>>> >>>>>>The reason I do NOT believe in computers that play humans only, is just as Bruce >>>>>>Moreland stated. I _LIKE_ to play humans, more so than I like to play other >>>>>>computers. However, I am rated a lowly 2950, with a high of 3080. Now, who is >>>>>>a 3000 rated GM going to play? A 3400+ or a 2950? Regardless of WHO the best >>>>>>is, the higher rated will almost always be played. >>>>>> >>>>>>I just took a look at my DB, and well over 80% of my rated games on ICC have >>>>>>been against other computers. With the other 20% being 5 or 6 FM's, IM's, and >>>>>>GM's. >>>>>> >>>>>>The truth is, that GM's on ICC are rating pigs just like Spitfire. If they >>>>>>offer lessons, the one with the biggest rating will get more stundents, more >>>>>>simul's, more money. >>>>>> >>>>>>I think Scrappy was a valid experiment. And I think it was interesting. >>>>>>However, I feel it was _VERY_ unfair. Let me explain: I understand 100% why >>>>>>Bob did it. Because if you ever saw Crafty before Scrappy was around, there was >>>>>>usually a line to play it. GM's were being cut off by computer accounts, and >>>>>>Bob has always said, he is more interested in playing humans than computers. >>>>>>This makes a second account completly understandable. However, let's look at it >>>>>>in this aspect; does Bob pay for the account Scrappy? I mean, I understand he >>>>>>pays for Hyatt, but what about Crafty and Scrappy? I think if ALL the (C) >>>>>>operators on ICC got 3 accounts, we would all be a lot better off. I mean, we >>>>>>could have one for (C)'s only, one for humans only, and one for ourselves. >>>>>>Perfect! However, that's $150 a year, just in a "recreational" hobby. I don't >>>>>>know many people who are going to go for that. I also understand that Hyatt is >>>>>>the author, and he has special privleges for this. However, SMK, Christophe, >>>>>>Ed, Frans, NONE of them have ICC or FICS accounts. It's not like we are taking >>>>>>away directly from the programmers. Also, you must again take into account, who >>>>>>is a GM going to play? A 2950 rated commercial program, or a 3400+ rated >>>>>>Crafty? Once again, it just all makes better sense. Once agian, I do not >>>>>>disagree about what Bob did, it did make sense. Just making the point that if >>>>>>we were all afforded the luxury of 3 accounts for the price of 1, we wouldn't be >>>>>>having this conversation. >>>>> >>>>> >>>>>I guess the above is directed at Bob. Good luck. >>>> >>>> >>>>Not directed at anyone. Just a simple observation on the situation. >>>> >>>> >>>>> >>>>>> >>>>>>I use my account for bullshit. I won't use a book here, use a screwy book here, >>>>>>use this here, that there, etc., etc. My 2950 rating is BS. And I am happy >>>>>>with it. I can get a game just about 24/7 on ICC, against the best computers, >>>>>>and SOME GM's. >>>>>> >>>>>>Let's also think back about 2 months ago. JRLOK issued a match to ME, on ICC. >>>>>>NOT to Spitfire. Let's all take a minute, and think why that is. >>>>> >>>>> >>>>>I don't understand this. If your account plays other comps, then it should hurt >>>>>JRLOK to play you in comparison to playing Spitfire, since your comps rating >>>>>will is lower than it would be if it played humans only. He did you a favor, >>>> >>>>We played unrated. And yes, it was a favor. Just not one that was asked. That >>>>was the strange part. Look below: >>>> >>>>>BTW, JRLOK has played spitfire 11 times for a record of -4 +2 =5. Since he was >>>>>generally rated about 75 points lower than spitfire in his losses (except 1), >>>>>his rating was not really harmed. This is what you would expect from a program >>>>>that only plays humans. If he played it more often, he probably would have >>>>>zeroed in on its weaknesses and even gained rating points. >>>> >>>> >>>>Like I said, I did not play JRLOK rated. It wasn't for points. Also, it was a >>>>25/5 game. How many of those 11 games were at a standard time control. AH, >>>>forget it. Point was, I think JRLOK wanted to play the "official" stronger >>>>program on stronger hardware. But like I said, just forget it. >>>> >>>> >>>> >>>>>Comp ratings are deflated, since they spend most of the time playing each other >>>>>and are avoided by humans. Comps upgrade their hardware and software, but their >>>>>ratings remain realtively low since they are just trading rating points with >>>>>each other, so the improvements do not get reflected in their ratings over the >>>>>long term. They are an almost separate rating pool within the larger rating >>>>>pool. >>>> >>>> >>>>Agreed. However, please take this into consideration: JRLOK lost most of his >>>>games against Scrappy. He still ended up with the highest rating on ICC ever. >>>>Bob will tell you, humans notice little difference between his 4x550 and 4x700. >>>>That's a full 600mhz. Upgrade from a 1.0Ghz to a 1.6Ghz and the computer >>>>operators will take notice, however, the strong GM's probably won't. >>> >>> >>>JRLOK has a plus score over the last 50 games against scrappy. Over the last 100 >>>games, he is only -2. I don't think you are making a strong point with this. For >>>all intents and purposes, they are equal. >>> >>>[snip] >> >> >>Check the games. I had Scrappy playing some very specific (and narrow) openings >>to isolate a couple of weaknesses and fix them. He would tell you himself that >>if he could break even at blitz, he would be terribly happy. > > >If you give me the pertinant eco code(s), I'm willing to wade through the game >history and exclude them for a more representative count. Go to Hyatt's FTP site. Download from the "games" directory. Load it into a viewer, and tell me the score of ALL games with JRLOK vs Scrappy & JRLOK vs Crafty. (I am not @ home right now, or I would be more than happy to.) Not 40% I am guessing. Perhaps I am wrong, or slightly off. But when I was preparing for my game against JRLOK I waded through them, and I know he truly doesn't have a very good record. Only recently, has he "learned" Crafty, and now can more/less keep up. Sometimes going a little over, sometimes a little under. Hopefully that'll give me arguement a TAD more substance. Slate
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.