The 12 mph page was originally suspended in February following various criticisms. You can access the old page (here)

A cyclist called Guy Chapman thought the previous page was worthy of ridicule and has set up the following web page: (click here)

Obviously Mr Chapman completely failed to understand the purpose or intention of the work. He also made the following errors of fact or understanding:

"vulnerable road users." No Guy. If the work had addressed vulnerable road users the headline number would have been even lower than 12 mph.

"This assumes that the drivers will not brake" Yes Guy. Of course it does. That's the point. 

"you're using the formula for V when it's clearly meant to be applied to delta V" But of course we're entitled to assume that delta v is equal to v because it's driver response in most cases which separate the two.

"you can't apply this formula to the average speed" We agree that this leads to an estimation error, but the estimation error is always to overestimate the speed from the probability of death. So the error, if corrected, would make the headline conclusion smaller. Perhaps 10 or 11 mph instead of 12 mph.

"The 70% don't speed all the time" with a ratio of 1,500:1, it hardly matters if the speeding proportion is 30% or 70%. (and in the free travelling conditions where speed enforcement has its greatest effects it's usually on the high side) It's still a massive ratio. 50% of 1,500:1 is still 750:1.