Simplified ELO: Best players of all time

[Magicwand]

i think that list is most unbiased list i ever saw.
i think if you can stress more points for recent games played then it will be perfect for current rating.

[Joaz Banbeck]

...There certainly seems to be a strong bias towards more recent players. I think perhaps this is because younger players tend to player older ones at their weakest, which gives a statistical perception that the young players are stronger. But I intentionally didn't try to model player strength as changing over time, because it's a very tricky thing to get right and is sensitive to parameter choice...

If you were to use only games in which the players were of approximately the same age, would there be enough data?

[topazg]

...There certainly seems to be a strong bias towards more recent players. I think perhaps this is because younger players tend to player older ones at their weakest, which gives a statistical perception that the young players are stronger. But I intentionally didn't try to model player strength as changing over time, because it's a very tricky thing to get right and is sensitive to parameter choice...

If you were to use only games in which the players were of approximately the same age, would there be enough data?

I don't believe this would work. People peak at different ages, peak for different lengths of time, decline at different rates, and stay active until a different proportional amount of ability decline. Age is not a good enough proxy for this confounder IMO.

[Kirby]

...

Now that I've pulled the data, I can certainly make such lists if you'd like . Maybe that can be a seperate post.

From a statistics perspective, I'm actually trying to answer a very specific (if ultimately unanswerable) question: if any two players play against each other, which is more likely to win? And it is actually objectively measurable, in a way: I could take the data from 2000-2004, train the model, then see how well it predicts the results of the matches from 2004-2009, for example. This may be the next thing I do, actually.

...

Okay, okay. Maybe I was playing the devil’s advocate when I suggested that such ranking lists are “pointless”, because I actually find these types of studies kind of interesting. However, it’s still important to remember that, with this type of problem, model selection is something that is difficult to be objective about. Some people may feel that particular inputs are more relevant to the “who would win in a hypothetical match” than others.

So I guess I will say that I do not find your study pointless. But I do feel that it is not possible to find a model/appropriately-allocated-input-data that will please everybody.

[John Fairbairn]

Inspired by this thread I googled and found an interesting piece at:

http://en.wikipedia.org/wiki/Methods_fo ... ut_history

One of the methods outlined there was to ask grandmasters who they believed best of all time. On that basis, I'd say that the only past master of go who would come high in a modern list would be Honinbo Shuei. This is based on number of mentions I've read by pros as to who they admire and whose games they studied. I'd say the jury would be out on Dosaku - too long ago and no close rivals (though the latter could be argued for Shuei, too, he seems to get the nod for his super-smooth style). If you count Go Seigen as a past master, then he would certainly be very high, and maybe top. Note also that Go was alleged to have said Yi Ch'ang-ho was no genius compared with Cho Hun-hyeon. Go also rated Huang Longshi very highly, and regarded Ogawa Doteki as the greatest ever prodigy. If Chinese players were canvassed, Fan Xiping and Shi Dingan would surely also be in the frame, and maybe Zhou Xiaosong (the Chinese Honinbo Shuwa), too, as a dark horse.

However, lack of mention of some players is suspicious. Honinbo Shusai is rarely mentioned as an exemplar, though he was obviously a sensational player. Perhaps he is still too controversial to mention, or (more likely, I think) his go lacks the innovation in fuseki that modern pros like to study, which aspect would explain high ratings for Go and Shuei, and also a low rating for another true great, Sakata).