Our dedicated chess computers in fact 300-350 elo weaker ??

This forum is for general discussions and questions, including Collectors Corner and anything to do with Computer chess.

Moderators: Harvey Williamson, Steve B, Watchman

Forum rules
This textbox is used to restore diagrams posted with the fen tag before the upgrade.
Larry
Senior Member
Posts: 2269
Joined: Wed Aug 01, 2007 8:42 am
Location: Gosford, NSW Australia

Post by Larry »

Larry wrote:I just happened to be browsing u-tubes today and happened upon
this one. While it won't be the final word in this discussion, it is
relevant to it. It's old, circa 1993, and features the well known
Brit, Eric Hallsworth:
https://www.youtube.com/watch?v=74xHd4A ... e=youtu.be

L
Apologies, this video is in another current thread.
Please disregard... :oops:
L
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Hi Larry,

Don't worry. This video is has some historical context that is very appropriate to this thread.

Best regards
Nick
donkeylane
Full Member
Posts: 679
Joined: Mon Aug 29, 2016 8:31 pm
Location: Cheshunt, Hertfordshire, UK

Post by donkeylane »

With regard to chess computers being overrated by 300-350 points,when the Mephisto Amsterdam came out in 1985 ,S/S number 2 gave it a rating,albeit a tentative one of 2199 ,the latest SSDF rating is 1829, a difference of 370 points,dedicated chess computers were still a novel conception then,and it's no surprise for overenthusiastic ratings,as an anecdote,I played at a local chess club toward the end of the 1980's ,one of the players had a Mephisto Academy,currently rated 1842 on the SSDF list,he came in one night rather chuffed with himself he'd beaten it that day,the point being that the Academy was a difficult opponent for a club player then,and human playing strength has not evolved to keep up with the evolution of chess computers.
User avatar
Steve B
Site Admin
Posts: 10140
Joined: Sun Jul 29, 2007 10:02 am
Location: New York City USofA
Contact:

Post by Steve B »

donkeylane wrote:With regard to chess computers being overrated by 300-350 points,when the Mephisto Amsterdam came out in 1985 ,S/S number 2 gave it a rating,albeit a tentative one of 2199 ,the latest SSDF rating is 1829,
SSDF decreased all ratings in 1992 by 100 Pts to adjust for ratings inflation
so they will show a 100 Pt difference then all other lists
Today Selective Search shows 1949 for the Amsterdam
Wiki shows 1927

you can see this quite clearly here:
http://www.schachcomputer.at/eloliste.htm

The Confusion Continues Regards
Steve
donkeylane
Full Member
Posts: 679
Joined: Mon Aug 29, 2016 8:31 pm
Location: Cheshunt, Hertfordshire, UK

Post by donkeylane »

Cheers Steve,that cleared that up.A little bit clearer than mud now. Regards Steve C.
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Steve B wrote:
donkeylane wrote:With regard to chess computers being overrated by 300-350 points,when the Mephisto Amsterdam came out in 1985 ,S/S number 2 gave it a rating,albeit a tentative one of 2199 ,the latest SSDF rating is 1829,
SSDF decreased all ratings in 1992 by 100 Pts to adjust for ratings inflation
so they will show a 100 Pt difference then all other lists
Today Selective Search shows 1949 for the Amsterdam
Wiki shows 1927

you can see this quite clearly here:
http://www.schachcomputer.at/eloliste.htm

The Confusion Continues Regards
Steve
CCRL 1995 Ply date begs to differ with your 1992 date. 2106 according to that for SSDF (PLY) same Publication CCNS = 2094 = Average across all ratings = 2127 for Amsterdam.

Or are you saying that there were multiple dumb downs and dumb ups?

Kind of contradicts to what Peter Grayson wrote earlier too. Seems to me that by 1995 the ratings would have become pretty reliable since that was about the year where dedicated computers where at their absolute peak with thousands of human vs computer games being sent to CCNS.

It seems incredible to me that so many reputable publications and reputable writers and reputable Testers and reputable games and reputable human-computer tournaments played throughout the world at that time. And they got it all so wrong?

Amazing incompetence right?

I don't think so....regards
Nick
User avatar
Steve B
Site Admin
Posts: 10140
Joined: Sun Jul 29, 2007 10:02 am
Location: New York City USofA
Contact:

Post by Steve B »

spacious_mind wrote:
CCRL 1995 Ply date begs to differ with your 1992 date.
Or are you saying that there were multiple dumb downs and dumb ups?
the 100 elo reduction i mentioned was in 2000...not 1992
(the chart i linked to showed the 2000 downgrade year and the year 1992 for comparison..i typed in 1992 by mistake)

actually there were several SSDF downgrades over the years
in 1989 SSDF adjusted downward by 70 Elo
other downgrades were smaller
AFAIK there have been no downgrades since then but i guess Lars Sandin would know that for certain

BTW...Hallsworth downgraded his lists over the years as did the CCR

Statistical Calibrations Regards
Steve
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Steve B wrote:
spacious_mind wrote:
CCRL 1995 Ply date begs to differ with your 1992 date.
Or are you saying that there were multiple dumb downs and dumb ups?
the 100 elo reduction i mentioned was in 2000...not 1992
(the chart i linked to showed the 2000 downgrade year and the year 1992 for comparison..i typed in 1992 by mistake)

actually there were several SSDF downgrades over the years
in 1989 SSDF adjusted downward by 70 Elo
other downgrades were smaller
AFAIK there have been no downgrades since then but i guess Lars Sandin would know that for certain

BTW...Hallsworth downgraded his lists over the years as did the CCR

Statistical Calibrations Regards
Steve
Statistical Calibrations for what Steve? It can't be as a result of human versus computer play. So what would justify a change. I personally think that there are no justifications for it after 1995 as human-computer play got less and whoever was left playing against them had his collection and would mostly talk about his wins and rarely his losses if at all. You can forget about these people and their opinions. Sometimes they may throw in a report of a loss just to add credibility. But you can forget those.

Just collate the official human-computer matches and that is enough to show the inconsistencies on lists. You don't think that the Bronsteins of this world when they played against computers in tournaments did not come prepared? All Grandmasters go into a Tournament fully prepared. He probably owned some of them.

People show off their best moments and if they error early on well they just turn it off and don't tell anyone. You can't make judgements on stuff like that.

Best regards
Nick
User avatar
Steve B
Site Admin
Posts: 10140
Joined: Sun Jul 29, 2007 10:02 am
Location: New York City USofA
Contact:

Post by Steve B »

spacious_mind wrote: Statistical Calibrations for what Steve? It can't be as a result of human versus computer play. So what would justify a change. I personally think that there are no justifications for it after 1995
SSDF cailbrated their ratings to human ratings ..that is their explanation for the ratings reductions..

more here...scroll down until you you see the section on "Ratings Calibrations"

https://chessprogramming.wikispaces.com/SSDF


personally i dont really care all that much about the details but you seem to so why not take this up with Lars Sandin or contact the SSDF directly
for the details?

Calibrating Regards
Steve
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Steve B wrote:
spacious_mind wrote: Statistical Calibrations for what Steve? It can't be as a result of human versus computer play. So what would justify a change. I personally think that there are no justifications for it after 1995
SSDF cailbrated their ratings to human ratings ..that is their explanation for the ratings reductions..

more here...scroll down until you you see the section on "Ratings Calibrations"

https://chessprogramming.wikispaces.com/SSDF


personally i dont really care all that much about the details but you seem to so why not take this up with Lars Sandin or contact the SSDF directly
for the details?

Calibrating Regards
Steve
That's funny, you don't think these posts are not being read? You don't think that the founders don't even read this a little bit? Let me remind you what this post is about. What's your favorite computer? R40 maybe? Well it's now rated at hmmm let me see.. 2055 ELO! You think I am joking? Then read the thread headline!

But I know you don't care about lists and stuff so you can ignore the rest here as I am not writing it for you. Because I need people that actually care a little about lists to read this!

How about a new list, that's being looked at and discussed today and receiving thumbs up by some:

Image

See R40 is already being deflated (sounds like a recession). Well, I know the poster also has a sense of humor so perhaps it's just a joke. But the thumbs up are a little disconcerting don't you think? Look at the bottom of the list there is R30 again and what do you see? Oh it is now 2138. Because some club player can post and win games against it! Brilliant! Oh if this were to be a USCF List then Fide or European rating would now be 2070! Brilliant just like this thread has shown us what it should become!

Image

OOops... wrong pic... that's an Eric Hallsworth quote talking about some guy playing against computers. Seems the guy's rating is Fide 1714, but I could be mistaken. But then again perhaps if he waits another 10 years and plays them again then he gets to beat them all for sure.

Image

Not just Fritz 3 on a P90 but everyone on that list needs to have their rating recalibrated right?

and some of these Grandmasters too:

Image

And, why would that be so? well... let me see...

Image

Ok, I get it now... Fritz 3 beats the world, but loses to R30 Normal setting at 40 in 2 hrs 7-5 who in turn loses to a club player and therefore everything gets recalibrated downwards for dedicated computers only!!, a trend that started with the founders and just gets better and better. Got it!! Thanks... Brilliant and Fiendishly Genius!!! Well the P90 result maybe worth about 60 points but no more. Add 60 to 2359 and you get 2419. R30 Normal = 2395.

It took me a while to get it....outstanding!

Image
Image

I mean I can't make this stuff up. It's all out there volumes and volumes of stuff....

What has this dedicated chess computer world come to. Perhaps we should all just take a break, go to the beach and chill and drink some Kool Aid...
:P

Best regards
Nick
User avatar
Steve B
Site Admin
Posts: 10140
Joined: Sun Jul 29, 2007 10:02 am
Location: New York City USofA
Contact:

Post by Steve B »

spacious_mind wrote:
But I know you don't care about lists
actually i said i dont care about the details of how they calibrate their lists
again ill leave it for the SSDF to defend their 30+ years of rating the dedicated computers
spacious_mind wrote:
OOops... wrong pic... that's an Eric Hallsworth quote talking about some guy playing against computers.
here too..Ill leave it for Eric to defend his 30+ years of rating lists
spacious_mind wrote:

Perhaps we should all just take a break, go to the beach and chill and drink some Kool Aid...
:P
Great idea
i recommend you take this computer with you
it was never rated so you take comfort in knowing the rating is not overstated by the all the deceitful "founders" all working secretly in concert from around the world to lie to the buying public

https://www.flickr.com/photos/10261668@ ... 922171154/

The Beach Boys Send Their Regards
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Steve B wrote:
Great idea
i recommend you take this computer with you
it was never rated so you take comfort in knowing the rating is not overstated by the all the deceitful "founders" all working secretly in concert from around the world to lie to the buying public

https://www.flickr.com/photos/10261668@ ... 922171154/

The Beach Boys Send Their Regards
You nailed it Steve!! That is exactly what has happened! Everyone in the world is being lied to!

Pick any list play the games and you will find out OMG I am being lied to!

Thanks for finally noticing the problem :P

Best regards
Nick
donkeylane
Full Member
Posts: 679
Joined: Mon Aug 29, 2016 8:31 pm
Location: Cheshunt, Hertfordshire, UK

Post by donkeylane »

I'm looking for that 76 positional ELO calculator,I believe it was set by a German university,and it was Steve B who submitted the link,I intend to give it to my computers and will submit the ELO calculation on this thread . Many thanks Steve C.
User avatar
Steve B
Site Admin
Posts: 10140
Joined: Sun Jul 29, 2007 10:02 am
Location: New York City USofA
Contact:

Post by Steve B »

donkeylane wrote:I'm looking for that 76 positional ELO calculator,I believe it was set by a German university,and it was Steve B who submitted the link,I intend to give it to my computers and will submit the ELO calculation on this thread . Many thanks Steve C.
Actually it was Fern who posted that link:
http://hiarcs.net/forums/viewtopic.php? ... sc&start=0

ELO Regards
Steve
donkeylane
Full Member
Posts: 679
Joined: Mon Aug 29, 2016 8:31 pm
Location: Cheshunt, Hertfordshire, UK

Post by donkeylane »

O.K thanks for that,I'll get onto that shortly ,and then we'll be able to clear up all that misunderstanding regarding the ratings of chess computers over the years.Ha ha. .I did have a look at it, and it did seem a very interesting positional test,but I was unable to find the link. Still learning ,regards Steve C.
Post Reply