8th Amateur Championship Division 5

Discussion of computer chess matches and engine tournaments.

Moderator: Ras

User avatar
Graham Banks
Posts: 45333
Joined: Sun Feb 26, 2006 10:52 am
Location: Auckland, NZ

Re: Final Standings

Post by Graham Banks »

Andres Valverde wrote: Thanks for the tourney Graham. It is amazing that a no improved EveAnn has hold the 4th with less trouble against improved opponents. I promise it has no learning implemented :)
Thanks Andres. 8-)
gbanksnz at gmail.com
swami
Posts: 6664
Joined: Thu Mar 09, 2006 4:21 am

Re: Final Standings

Post by swami »

hgm wrote:It does remain strange that Joker always performs enormously worse in CCRL events than it does, for instance, in Open War:

Code: Select all

20:  Arion 1.7           47.0 / 69  
28:  Joker 1.1.14e       41.0 / 69  
29:  Sorgenkind 0.4      40.5 / 69  
33:  Deuterium 080126266 37.0 / 69  
40:  Cheese 1.1b         32.5 / 69  
41:  Zeus 1.29           32.0 / 69  
42:  EveAnn 1.63         31.0 / 69  
46:  Buzz 20080620       29.0 / 69  
48:  Brutus 5.4          27.0 / 69  
If I look how Joker performs compared to Zeus / EveAnn / Buzz / Brutus, the difference is really flabbergasting. I have no really satisfactory explanation for this. It could be that Joker only performs well when running on Core 2 Duo (on which I developed it). Could be the effect of the opening book. Or it could be that playing it only against opponents in such a narrow rating range strongly distorts its rating.
I believe that's the case.

Engine Caligula is obviously little less stronger than Danasah yet it scores better than Danasah in 81 rounds so far. Ie the engine plays better against wide range of participants than it does against the group of few engines around its ratings range.

There are several such cases I could offer.