Match Fruit 2.3.1 PPC - DanaSah 3.03 PC

[pedrox]

I have a pocket PC, is a IPAQ 1940 with a StrongARM processor to 266 MHz. I always have asked myself who ELO will have the strongest programs. In some analyses I have seen that perhaps its level of game could be more than the 2600 points, this valuation I even saw when they not were the 525 MHz PPC.

Recently I have been able to unite the IPAQ with the computer by bluetooth and by means of the CEboard program I can make match with auto232 between CEboard with engine like Fruit (the best with Toga for PPC) and Arena with DanaSah 3.03 in Pentium IV 2400 Mhz (it is not possibly my better version, used by be in that computer then)

The question is:

Will be able to gain Fruit on pocket PC (8Mb hash) to DanaSah in the PC (64 Mb hash)?

For it I am going to play 10 games of 25 minutes by player.

The rating of DanaSah 3.03 in CCRL was of about 2460 points, if we considered that Pentium IV is approximately half of express we could consider the ELO of DanaSah in about 2400 points.

The ELO of Fruit 2.3.1 in CCRL is 2869 (computer CCRL 20 times faster than the PPC).

The computer in MHz is approximately 10 times faster than the PPC, will be sufficient so that DanaSah wins?

a) Fruit win 10-0
b) Fruit wins by 4 points o more (not 10-0)
c) Fruit wins by less than 4 points
d) 5 - 5
e) DanaSah wins by less than 4 points
f) DanaSah wins by 4 points o more (not 10-0)
g) DanaSah win 10-0

Bet, I will give results in 2 days

To the winners I can give a copy of a new private engine if they wish it

Best,

Pedro

[Graham Banks]

I hate to say this Pedro, but I'll pick Fruit by 4+

Regards, Graham.

[pedrox]

There is no problem Graham, if DanaSah loses will not be a dishonor, many GM have lost against programs that play in PPC.

Recently pocket Fritz 3 can be said that it failed in a tournament, it could not obtain GM norm, he is interesting to ask if because the version who used was erroneous or if because pocket PC still does not have the sufficient force (in that tournament PPC was 625 Mhz)

http://www.ajedrezmartelli.org.ar/magis ... rcosur.htm

Pedro

[swami]

Fruit by less than 4 points. (If the Fruit version is latest)

I'm sure Danasah would do better if it was Fruit 2.1

Regards.

[Tony Thomas]

Considering the stuff I have read before, doubling the speed helps gain the engine about 50 points. So quadrupling the speed of the pocket PC, you get almost the speed of the PC. I guess that means fruit will lose about 200-250 points, that means it will be playing at about 2600-2650. The best you can do is probably to tie the match, I dont expect a complete one sided sweep. What books are the programs going to use?

[Uri Blass]

I have a pocket PC, is a IPAQ 1940 with a StrongARM processor to 266 MHz. I always have asked myself who ELO will have the strongest programs. In some analyses I have seen that perhaps its level of game could be more than the 2600 points, this valuation I even saw when they not were the 525 MHz PPC.

Recently I have been able to unite the IPAQ with the computer by bluetooth and by means of the CEboard program I can make match with auto232 between CEboard with engine like Fruit (the best with Toga for PPC) and Arena with DanaSah 3.03 in Pentium IV 2400 Mhz (it is not possibly my better version, used by be in that computer then)

The question is:

Will be able to gain Fruit on pocket PC (8Mb hash) to DanaSah in the PC (64 Mb hash)?

For it I am going to play 10 games of 25 minutes by player.

The rating of DanaSah 3.03 in CCRL was of about 2460 points, if we considered that Pentium IV is approximately half of express we could consider the ELO of DanaSah in about 2400 points.

The ELO of Fruit 2.3.1 in CCRL is 2869 (computer CCRL 20 times faster than the PPC).

The computer in MHz is approximately 10 times faster than the PPC, will be sufficient so that DanaSah wins?

a) Fruit win 10-0
b) Fruit wins by 4 points o more (not 10-0)
c) Fruit wins by less than 4 points
d) 5 - 5
e) DanaSah wins by less than 4 points
f) DanaSah wins by 4 points o more (not 10-0)
g) DanaSah win 10-0

Bet, I will give results in 2 days

To the winners I can give a copy of a new private engine if they wish it

Best,

Pedro

DanaSah3.03 has no ccrl rating

in blitz we have only

http://computerchess.org.uk/ccrl/404.li ... _length=30

1 DanaSah 2.85k 2506 +34 −34 55.4% −38.4 31.8% 286
97.0%
DanaSah 2.73 2461 +32 −32 45.4% +42.8 33.8% 340
98.4%
DanaSah 2.44 2414 +29 −29 45.4% +35.6 30.1% 422

In 40/40 we have

http://computerchess.org.uk/ccrl/4040.l ... _length=30

DanaSah 3.13 2550 +46 −46 44.8% +31.9 37.3% 153
88.9%
1 DanaSah 2.85 2511 +41 −41 48.3% +15.3 31.9% 204
82.7%
DanaSah 2.73 2481 +45 −46 45.4% +31.3 37.3% 153
70.1%
DanaSah 2.44 2464 +44 −45 33.9% +119.0 35.6% 174



Uri

[pedrox]

Considering the stuff I have read before, doubling the speed helps gain the engine about 50 points. So quadrupling the speed of the pocket PC, you get almost the speed of the PC. I guess that means fruit will lose about 200-250 points, that means it will be playing at about 2600-2650. The best you can do is probably to tie the match, I dont expect a complete one sided sweep. What books are the programs going to use?

Fruit is playing with the book varied.bin (default), I have not paid attention until you have made the question, I also can use performance.bin that I imagine that it will be better, I also will play with it. DanaSah will play with the own book and I will use 4 men of bitbases of scorpio.

[pedrox]

DanaSah 3.03 played in CCRL 40/40, the results were erased when I published version 3.13, then had about 50 games with a ELO of about 2460 I create to remember, I think that if that version had played more games it must have been near version 2.85. (2511)

Pedro

[Tony Thomas]

May be you should let fruit play with performance.bin since you are playing a tournament.

[Ovyron]

I hate to say this Pedro, but I'll pick Fruit by 4+

I did this as well...