A couple of questions:
Are these Elo ratings really comparable to human Elo ratings? So it's pretty clear that these engines would all beat any human player in history?
(Edit: I know computers have been able to beat human champions for what xkcd points out is about half of my lifetime, but Deep Blue had a lot of resources behind it that I would imagine these engines don't.)
How do they deal with differences over time in the hardware that the engines run on? (both in terms of Elo calculation consistency and in terms of historical comparisons between engines' performance ... how do we know how much improvement is due to programming improvements vs. faster hardware? do we know how earlier generation engines would fare today running on the same hardware that these are running on?)
Do the engines in a particular match run on the same or different hardware? (I thought I saw people in the chat talking about this but I didn't exactly understand what they said.)
http://tcec.chessdom.com/live.php
Stockfish 8 squaring off against Houdini 5, 100 games back to back, no rest days.
Go, Stockfish!