What's YOUR Match-up Rate?

Originally posted by Paul Leggett
I was thinking the same thing- comparing a top CC player to Fischer OTB seems incongruous. A strong player with days to move vs a super strong player with minutes per move is an interesting question.

It would have been really cool if Fischer had played a large number of CC games back in the 1960's- simply comparing his OTB rate to his CC rate would be very enlightening.

As incongruous as it might seem, the results with pre-computer CC GMs and super GMs OTB then and now are basically the same.

Originally posted by Paul Leggett
I was thinking the same thing- comparing a top CC player to Fischer OTB seems incongruous. A strong player with days to move vs a super strong player with minutes per move is an interesting question.

It would have been really cool if Fischer had played a large number of CC games back in the 1960's- simply comparing his OTB rate to his CC rate would be very enlightening.

Well I'm not the only person who's done this analysis.
All the pre computer era Correspondence Chess World Championships have been tested & the benchmark thresholds remain intact.
Many OTB FIDE World Championships have been tested (the focus being on those which yield the larger sample sizes) & again these thresholds seem to apply.

All these great players, if they have 3 years to complete a CC WC or if they're an OTB Super GM with several minutes to ponder each move, seem to play an engine-like game to only a certain extent.

Then go online, use the biggest database you can to cut modern theory out of the equation & Mr or Mrs nobody at the head of the ratings list with few (if any) losses & moving daily in 25 games in progress plays a hugely more engine-like game than any of the greats. 😉

It may not be 100% proof, but if the methodology is applied correctly then I think depending on the batch results it can form proof beyond reasonable doubt.

Don't forget a site like this can also utilise other methods to form a case, and these can provide sound smoking-gun circumstantial evidence, but ultimately if you are trying to find someone who is consistently using an engine in their games then it stands to reason that you must see how the moves they elect to play match with those chosen by a strong engine.

Unfortunately modern Super GM's don't play long time control CC whilst being monitored for assistance, but I agree it would be interesting to see how much more engine-like their non-database play is under such circumstances.

Originally posted by Zygalski
it would be interesting to see how much more engine-like their non-database play is under such circumstances.

Agreed. Another thing that interests me, though it’s not feasible to do the analysis in practice, is if we took the latest FIDE top 1000 players and obtained sufficient match up statistics, how much different would the FIDE ranking list be from the corresponding match up ranking list? And to take this further, what is the largest FIDE Elo difference that can be found between two players where the weaker player obtains a higher match up? Anyone got a spare lab of PCs available? 😉

Originally posted by Zygalski
Well I'm not the only person who's done this analysis.
All the pre computer era Correspondence Chess World Championships have been tested & the benchmark thresholds remain intact.
Many OTB FIDE World Championships have been tested (the focus being on those which yield the larger sample sizes) & again these thresholds seem to apply.

All these great play ting to see how much more engine-like their non-database play is under such circumstances.

I think you are answering questions I'm not asking, but your last line gets back on the page!

Ok 🙂
I'm trying to cover some bases of FAQ's as well.

I think the general feeling I get from the questions asked is that people mistake strong human-like play for engine-like play.
In my opinion, regardless of the use of engine analysis of openings/early middlegame by modern players, and the time used to ponder candidate moves, humans will still play like humans & engines like engines!

I can't imagine unassisted CC players regularly going off on tangents just because such & such a move yields a +0.15 advantage at depth of 20 ply!
Human player moves in balanced positions follow plans - be it short or long term - whereas an engine obviously just does brute force number crunching & this is largely what leads to these consistent benchmark thresholds for unassisted play.