Author: Peter Berger
Date: 20:34:06 06/05/01
Go up one level in this thread
On June 05, 2001 at 17:24:47, Chris Carson wrote: >1. Unplug the computer while it is thinking, you win on time. > >2. Search for games (comp or human) that the program has lost, > turn off learning and set program to default. Play the moves that > has beat the program before. > >3. Set your time to whatever you like and set the program to game/1 min > (or any unfare variation that ensures you win or draw). > >4. Just get some old PGN's that the program lost and put your name on > them as the opponent. > >5. Make up the PGN's. > >6. Play other programs aginst your program and publish the games as > if you are the opponent. > >7. Get your GM friend to play the program and just claim the PGN's > as your own or reset the program to default and replay the moves. > >8. Just tell everybody that programs are so weak that they are not > worth your time and besides you deserve to get paid to play them. > >9. Remember one game does not mean anything if the computer wins (after > all, they are all just one game), but if you win or draw, obviously > the computer must be rated 800 points below your worst performance. > >10. Take 3 months and play the program everyday in serious games. Reset > the computer and only replay the games that win. This is the method > prefered by the World Champ, so you know it must be the best. > >;) > >Best Regards, >Chris Carson > >Did I miss any? Just take your favourite program and hook it to ICC . Use an open formula that allows anyone from about Rating 1800 to play you and issue some blitz game seeks . Come back about 3 days later and have a look at the games - you might be surprised . And before someone comes up with those people all cheating - looking at the games doesn't confirm that . The reason this fact is not that well-known is probably the ridiculous formulas most computer accounts use . You'll probably see several easy draws your program allowed . You'll probably see a few stonewall approaches that worked and killed the prog . For sure this gets even worse if you look at the won games , too . You'll see several games where the human _nearly_ killed the prog only to blunder later . And note : we are talking about 3 0 or 5 0 blitz here not about games in 2 hours . I have tried this with commercial and amateur programs before and I haven't seen much improvement so far . pete
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.