It seems like ELO skill ratings fall apart when comparing players of two different eras, since ELO Ratings involve the skill level of your opponents too.

Maybe a better metric would be running both players games through a computer to see the "%best move" metric?

> Maybe a better metric would be running both players games through a computer to see the "%best move" metric?

Due to computers being common place now, Magnus would absolutely destroy that metric compared to older eras. Chess has changed a ton in the past 10-20 years due to computers, being able to analyse lines. And now Magnus is known for playing "Ai Lines" which is the kind of stuff the new ML models do which tends to be pretty bonkers and un-human but gets long term results.