In theory, it is a 1/15 chance of successful attack. Which is a terribly low success rate of attack prevention.

In practice, it is 15/16 chance of detection of the exploit attempt. Which is an extraordinarily high rate of detection, which will lead to a fix by Apple.

Net net, huge win. But I agree they come across as overstating the prevention aspect.