[mox1]

I think it's easier to relate to a coin flip if we use an "unfair coin".

90% of the time the coin flip returns tails (aka fail).

10% of the time it returns heads (aka win).

For a given coin flip, their algorithm can predict the results 50% of the time. At this point I don't remember the calculations off the top of my head, but it involves a Binomial distribution.

[RogerL]

Hmm, I can predict that 90% of the time, so I don't follow.

[kylebrown]

That prediction doesn't improve over the prior. So its no better than the (biased) coin flip.