RMS results - from ELT to EP

[nimretred]

Hello,

I just got my results from RMS, which include:

• Event Loss Table, witch include Mean Loss Net, Correlated and Independent standard deviations and annual rates

• The other file does not really have a name, but it is the one with the return periods and net losses. Is this what's called Exceedence Probability, if I take 1/return period and plot it against the losses?

I'd like to know how the two tables are related, that is, how to get from one to the other.

Thanks

[Cho Da]

You need the document titled, "RMS&tm; Exceedance Probability Methodology".
RMS can provide this if you are a licensee, otherwise you should be able to get this from whoever ran your data.

[notreallyme]

Originally Posted by nimretred

Hello,

I just got my results from RMS, which include:

• Event Loss Table, witch include Mean Loss Net, Correlated and Independent standard deviations and annual rates

• The other file does not really have a name, but it is the one with the return periods and net losses. Is this what's called Exceedence Probability, if I take 1/return period and plot it against the losses?

I'd like to know how the two tables are related, that is, how to get from one to the other.

Thanks

The second one is an EP (exceedence probability) curve.

To get from one to the other first you have to decide what you are creating.
Aggregate Exceedence Probability Curve (AEP)
Occurence Exceedence Cuve (with secondary uncertainty)
Occurence Exceedence Curve (without secondary uncertainty)

-- I suppose one could create an AEP without secondary uncertainty, but I have never heard of such a thing and think it's use would be exactly 0.

The Occurence without secondary uncertainty is pretty easy, and it can be seen right on the Event Loss Table (ELT). Though we used to use it, it's sort of a dinosaur now.

To create a OEP (with secondary uncertainty).
- Find points on the secondary uncertainty curve and calculate that way.
For example if the Expected Loss for an event is $1,000,000 with a SD of $500,000 you can find the Beta Parameters for this event and get 100 new 'events' from that one event.

To create the AEP curve you need to consider correlation as the above poster mentioned this is outlined in the RMS documentation*. -- I have never, nor know of anyone who has ever, actually built something to create this. But I would guess someone has done it somewhere.

* There is also a document titled 'Calculating Secondary Uncertainty' or something. Use 'Secondary uncertainty' as a search on the RMS documentation and you will find it.