# Extreme Value Theory Data Scaling

I have a data set available of almost thirty years of data, with for each month the number of occurrences of a certain event and the total number in the set available.

What I would like to compute is the yearly rate which will not be exceeded in 99.5% of the cases. i.e. which will occur 1 in 200 years.

If I work with the yearly rate, I only have 30 data points, while I could also scale my monthly data (times 12) to obtain the yearly rate each month. What is the effect of the scaling when I'm looking at my result? Say after extreme value analysis on the scaled data I obtain a rate of 10% which will not be exceeded in 99.5% of the cases. This is the yearly rate occurring in a month. So, if I understand this correctly, this means the yearly rate I won't be exceeding in 200 months (~ 17 years)? How do I transform this rate to a 1 in a 200-year event?

• ...but then the probability of exceeding the yearly rate of 10% in one month does not imply in fact I will have a rate of 10% over the full year. :s Sep 11 '15 at 7:49