[andresgottlieb]

He's talking about cases in which preventative measures could be taken regarding those 200 predictions (eg. 200 asteroids with high probability of crashing into earth, maybe we could track them, identify one coming towards us, and blow it up)

Even though, they were 200, the fact that one of them was identified correctly, saved human civilization as a whole.

The point being, 1 in 200 is not a bad success ratio per se

[catalogia]

As a practical matter, if somebody has a 1/200 chance of being right when sounding the alarm, by the time something actually happens I think nearly everybody will have written them off as the 'boy who cries wolf.'

[baddox]

It's a bad success ratio if the 200 predictions didn't have good explanations for why they were made, because that means they are essentially 200 arbitrary predictions. In that case, the fact that one prediction came true is irrelevant.

To put it another way: if you're making arbitrary predictions without explanations, then you increase the likelihood of a prediction coming true by increasing the number of predictions you make. But clearly we shouldn't value predictors simply based on the number of predictions they make.