Bring back the computer champions

One great source of contention that has always existed in chess is how to measure the relative talents and ratings of players of the past compared to the present. It is unlikely that the ELO ratings of the past are applicable to the present, for instance, and we can never be sure how older players would have done without modern opening theory and such. It doesn't have to be this way, though.

We know (very roughly) the ratings assigned to the old-time computer chess champions--Kaissa, Machack VI, Cray Blitz, etc. We know this because they played rated humans as well as computers, although in most cases the number of games that they played were not conclusive. Because they got better at chess as time went on, we have ratings running the gamut from the 1200's and even below this up to elite grandmaster strength.

I think that we should use these old computer champions as benchmarks upon which we base our current ELO ratings. It would help to keep some reasonable idea of the relative weight of ratings in different eras, and these programs in their reconstructed versions would stay around forever.

Finally, now that it seems clear that human chess supremacy will not last much longer (Hydra's near 3000 ELO rating confirms this), it will no doubt become less interesting to have man vs. machine matches in the future, since there isn't any challenge. However, we still have old players from the past that we can play with forever (albeit mechanical ones), and they come from a time when the machine-human chess contest was still very much alive. We may not be able to bring back Philidor, Morphy, Tal, or Lasker, but we can definitely bring back everyone from MacHack to Deep Blue.

Deep Blue is especially important, because it came along at a time when the best humans and the best computers were almost evenly balanced. It is a historical champion/player in the same sense that Philidor or Steinitz is, and we have the opportunity to have it around forever.

Originally posted by Gioachino Greco
One great source of contention that has always existed in chess is how to measure the relative talents and ratings of players of the past compared to the present. It is unlikely that the ELO ratings of the past are applicable to the present, for instance, and we can never be sure how older players would have done without modern opening theory and suc ...[text shortened]... same sense that Philidor or Steinitz is, and we have the opportunity to have it around forever.

I think for the past computer champions it would also be interesting to reconstruct the original hardware they played on. They'll then be exactly as good playing strength-wise as they always were, and we can measure current ELO ratings against that. That way, a 1600 ELO today would be exactly the same as a 1600 ELO 100 years in the future.

Originally posted by Gioachino Greco
I think for the past computer champions it would also be interesting to reconstruct the original hardware they played on. They'll then be exactly as good playing strength-wise as they always were, and we can measure current ELO ratings against that. That way, a 1600 ELO today would be exactly the same as a 1600 ELO 100 years in the future.

Actually, it's just the opposite. I'm looking for a reliable way to rate the human players, so that a 1600 rating for a human is always a 1600 rating whether it's from 2006 or 2025.