mclane wrote:in niggemanns company work many strong chess players. maybe they played with the machine and came to this evaluation.
of course that is possible
but the ratings from Selective Search come from games played against other computers with a time limit of at least 1 minute per move and most are 40/2HRS
this i think is more reliable then "unnamed "players i think
but i am still not sure how many games the Citrine has played for Selective search
sometimes this information is mentioned and sometimes not
mclane wrote:if you like i can give schach niggemann a telephone call tomorrow to ask them about the rating.
maybe we find out what is going on.
Yes Excellent..
Thanks Thor!
BTW..Just finished reading the very interesting articles you wrote for SS when you actively competed with CSTAL
very long articles Eric Permitted you
i think very few writers were givin so much space in SS
The thing is they appear to quote from the SSDF list, but it isn't even on the latest available list. Perhaps it is, however, on the rating list that is not yet available on the net.
The thing is that the SSDF ratings nowadays have all dropped 100 points since the "gold" standard (my expression) of 1996. When I look at the ratings some of these units have with which I am familiar, many seem to be significantly under-rated. I just don't agree with the seemingly arbitrary drop that occurred.
Nevertheless, according to the ratings of other machines on the list, I woud venture to suggest that 1900 ELO is fairly accurate - at least by the standards of the other ratings on the list.
Yes, crazy obviously. Although in the latest Chess magazine I noted an ad for Carnelian II claiming a rating of 2000 ELO. Considering that Carnelian II ocassionally wholesale blunders pieces away for no reason, can't even defeat Excalibur King Arthur when it doesn't blunder and can only draw with Star Ruby if given 3 minutes per move to 2 seconds per move for Star Ruby...
The only computer ratings I really trust these days are ones where I can go back to my own source data (i.e huge games database of 40 in 2 games) and calculate ratings based on that.
Incidentally, my brother has Novag Amber rated at 1905. This is based upon 156 games of 40 moves in 2 hours against 14 other computer opponents with an average rating of 1764. Certainly Star Ruby and Citrine are improved from the Amber, but not as much as many people think. The improvements are actually more theoretical than anything else. As an OTB opponent, I suspect most people would not even notice the improvements (apart from Citrine opening book of course). Infact I've even seen Amber get a slight plus score against Star Ruby in a 10 game match.
