[asimpletune]

But wasn't the bias in the paper something like 50.5% vs 49.5%?

[mewpmewp2]

Yes, but I used more extreme numbers for ease of calculation and to clearly indicate the direction of a probability.