Author: Christophe Theron
Date: 23:36:01 11/19/00
Go up one level in this thread
On November 19, 2000 at 12:22:23, CLiebert wrote:
>On November 19, 2000 at 12:13:07, Christophe Theron wrote:
>
>>On November 19, 2000 at 10:14:46, stuart taylor wrote:
>>
>>>On November 18, 2000 at 20:00:14, Aloisio Ponti Lopes wrote:
>>>
>>>>Thüringen Mohlsdorf 2000
>>>>
>>>>Junior 6.0 Athlon 1000 +1 -1 +1 +0 -1 1 5.0/6
>>>>Deep Fritz 2x P3-866 -½ +1 +1 -1 +½ r 4.5/6
>>>>Nimzo 7.32 Athlon 1000 -½ +1 -0 +1 -1 3.5/6
>>>>Gambit Tiger 1.0 Athlon 500 -1 +0 -1 +½ -½ 3.5/6
>>>>Shredder 4.0 Athlon 800 -1 +1 -0 +0 -1 3.5/6
>>>>Chess Tiger 13.0 P3-840 +½ -0 +1 -1 +0 3.0/6
>>>>Hiarcs 7.01 P3-500 -½ +1 -½ -½ +0 3.0/6
>>>>Goliath Light 2.0ß Athlon 650 +½ -0 +0 -1 +1 3.0/6
>>>>Rebel Century - Athlon 1000 -0 +0 -1 +1 -½ 3.0/6
>>>>Zarkov 5.01 Athlon 1000 +½ -0 -½ +1 +0 2.5/6
>>>>Hiarcs 7.32 Athlon 1100 +0 -½ +½ -0 +1 2.5/6
>>>>Gandalf 4.32f Athlon 1000 +0 -1 +½ -0 -0 2.0/6
>>>>Chigorin's Way Cel. 500 -0 +½ -0 +0 +½ 1.0/6
>>>>Chessmaster 6000 P2-400 +1 -0 +0 -0 -0 2.0/6
>>>>
>>>>I can't understand why Gambit Tiger was running on an Athlon 500. Can someone
>>>>explain please?
>>>>
>>>>A. Ponti
>>>
>>>It shouldn't need more. It's not much difference to the others' timings, much
>>>less than one extra ply.
>>
>>
>>
>>Don't make a fool of yourself by posting such nonsense.
>>
>>The speed difference between the winner on Athlon 1000MHz and Gambit Tiger on
>>Athlon 500 accounts for a 70 elo points handicap for Tiger.
>>
>>Uri also points out the fact that Gambit Tiger did not play with his own book,
>>which is even worse.
>>
>>You should have a good look at Gambit Tiger's performance with such a handicap,
>>and you will see that it is actually a very good performance.
>>
>>Instead of focusing on the fact that Gambit Tiger did less points than the 3
>>programs at the top, which had ALL superior hardware (by AT LEAST a factor of
>>2), you should have a look at all the programs with superior hardware which did
>>WORSE than Gambit Tiger. There are 7 of them, count them.
>>
>>
>>
>>
>>> But the results of GT seem very poor indeed. If it had come equal first, I may
>>>have said that the few missing mhz. justifies it not getting more. But not this!
>>
>>
>>
>>You need some more knowledge about computer chess I think.
>>
>>Do the experiment yourself: take the same program, and let it run on a 1000MHz
>>computer and let the other copy of itself run on a 500MHz computer. Let them
>>play against each other.
>>
>>What result do you expect?
>>
>>
>>
>>
>>> But if this is blitz (I don't see it written what it is), that looks fine to
>>>me. Knowledge takes more time, and it is not many games anyway.
>>
>>
>>
>>The Tiger family is not optimized for a given time control. They perform equally
>>well at all time controls, blitz or tournament time controls.
>>
>>
>>
>>
>> Christophe
>
>
>Below you see the final tableau and I agree: in relationship to the hardware
>Gambit Tiger played a very well tournament, indeed!
>
>In general, Christophe, which Tiger is the best according to your tests:
>GT or CT?
>
>Regards
>Christian
They are very close. The difference between the two IN GENERAL is under the
margin of error in all our experiments.
But if you take the results opponent by opponent, then it is clear that against
some opponents CT does better, and against others GT does better. At the extreme
ends of the spectrum, it happens that GT crushes some opponents, but it does not
happen at the same degree with CT.
I tend to believe that Gambit Tiger allows less draws, and so is probably the
best of the two for short tournaments (as demonstrated in France, Holland and in
Thueringen). But I don't have numbers to back up this idea.
Christophe
>1. Junior6 6
>2. Deep Fritz 5
>3. Gambit Tiger 4,5
>
>4. Shredder4 4
>5. Chess Tiger 13 4
>6. Nimzo732 3,5
>7. Hiarcs701 3,5
>8. Goliath Light 3,5
>9. Zarkov5 3,5
>10. Rebel Century 3 3
>11. Hiarcs732 3
>12. Gandalf 2,5
>13. Chessmaster 2
>14. Chigorin 1
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.