The difference is that the whiteboard formula is actually called the Clauser-Horne-Shimony-Holt inequality, not the Bell inequality, and it is a slightly more sophisticated version, concocted by the four physicists it is named after, five years The first two numbers were 0.56 and 0.82. The third was –0.59, so it seems I would have to take this way from the running total. The fourth number, another 0.56, should then have left me with a total of 1.35 and victory for Einstein.

That’s not what I showed. http://www.bbc.co.uk/iplayer/episode/b04tr9x9/the-secrets-of...

In fact, the subtlety is that the third term, the one that had a negative value, was already negative. The inequality read:

P(a,b) + P(a,b’) – P(a’,b’) + P(a’,b) ≤ 2,

So, plugging all the numbers, this looks like:

0.56 + 0.82 – (–0.59) + 0.56 = 0.56 + 0.82 + 0.59 + 0.56 = 2.53

So, sorry Einstein, victory goes to Bohr instead.