@9270
"If a player who blunders once in 1000 games plays a 100 game match"
++ That is besides the question. When 17 qualified ICCF (grand)masters with engines at 5 days/move play 104 games in the ICCF World Championship Finals, all draws,
then that is enough to conclude all these 104 games are perfect games with no error.
It does not say they played perfectly last year (they did not, there were a few decisive games),
or that they will play perfectly next year (maybe one tries too hard to win and loses a game).
@9269
"We do not have proofs that any model applies to the real world"
++ We do have such proof: models are applied to the real world all the time:
we build buildings, bridges, cars, airplanes, spaceships, computers... all by applying models.
Even the Pythagorean theorem was thought of to establish land boundaries after Nile flooding.
Thales's theorem was thought of to measure the height of the Great Pyramid.