That wouldn't matter (at least as I understand the article explanation). There's 2^3 = 8 different stamp configurations, two of which (RRR/GGG) trivially have 1

That wouldn't matter (at least as I understand the article explanation). There's 2^3 = 8 different stamp configurations, two of which (RRR/GGG) trivially have 100% match likelyhood. The other six are all equivalent, and they all predict the same thing. There's (3 choices of doors) * (3 choices of doors) = 9 ways to expose the stamp configuration, and in any of the six equivalent stamps there's five ways to match colors to four ways to differ, 5/9 = 55.55%, the expected distribution. [1] The problem is that quantum mechanics predicts (and experiments show) that the distribution you actually get is 50%, which is lower than what you would expect no matter what stamp configuration you get. On the other hand, it's possible that the way you choose which door to open is not distributed randomly. One of the ways to reconcile Bell's theorem is that there's a passport stamp that mind-controls you into picking doors with a 50% distribution. It's not a very popular theory, but it exists. [2] [1] Assuming RGG without loss of generality 1-R 1-R = Match 1-R 2-G = Miss 1-R 3-G = Miss 2-G 1-R = Miss 2-G 2-G = Match 2-G 3-G = Match 3-G 1-R = Miss 3-G 2-G = Match 3-G 3-G = Match Five matches to four misses. [2]http://en.wikipedia.org/wiki/Superdeterminism http://en.wikipedia.org/wiki/Superdeterminism