Is that for global?

Just to expand for Amit, all the parameters for a quake prediction affect the probability.

Small quakes happen more often than large quakes. That means that hitting a small quake is easier, thus it's probability is higher and it's "score" is lower.

Location is important too. A circle in a place with lots of quakes has a higher probability (lower "score") than the same size circle in a place with fewer quakes.

A larger time window will result in a higher probability (lower "score") than a short time window.

Now, when you have these three different things all affecting the probability and mixing together, it gets more complicated. Again, it may not "make sense", but the math doesn't lie.

Brian