search

cta-observatory / cta-lstchain

LST prototype testbench chain
https://cta-observatory.github.io/cta-lstchain/
BSD 3-Clause "New" or "Revised" License
26 stars 77 forks source link

Efficiency of theta cut is calculated relative to all events, not just the "central part" of the psf #997

Closed moralejo closed 1 year ago

moralejo commented 2 years ago

It seems to me that when calculating the theta cut which results in a certain efficiency for gammas, the "100%" used as a reference includes all events surviving the gammaness cut:

https://github.com/cta-observatory/cta-lstchain/blob/64abf6265a67d793dd9aee122bd46ca89d402e0a/lstchain/io/event_selection.py#L198-L208

This 100% therefore includes (mostly below 100 GeV) also all gammas which got a wrong head-tail assignment, i.e. they are in a "ring-shaped" population separated from the gaussian-shaped central part. It seems to me that what we really want is to integrate a given % of the "central" part of the PSF, excluding that ring. The cut is calculated with MC, so one could simply identify and exclude those wrong head-tail events, or, perhaps, simpler, to set a large value of theta (say, 0.7 deg) which defines what is "the 100% of the central PSF", relative to which the desired percentage is calculated. This value of theta could be an additional configuration value.

maxnoe commented 2 years ago

@moralejo I concur

maxnoe commented 2 years ago

We have the true sign already in the mc dl2 files I think, so it would be simple to compare disp sign to true sign.

moralejo commented 2 years ago

I think it would be good to have that as an option too. On the other hand the max angle alternative is "reproducible" on real data excess, which may be an advantage.

On Wed, 22 Jun 2022, 17:42 Maximilian Nöthe, @.***> wrote:

We have the true sign already in the mc dl2 files I think, so it would be simple to compare disp sign to true sign.

— Reply to this email directly, view it on GitHub https://github.com/cta-observatory/cta-lstchain/issues/997#issuecomment-1163279351, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADMCCOWDQUG4IONVVYTTA5LVQMX47ANCNFSM5ZQVWQWA . You are receiving this because you were mentioned.Message ID: @.***>

-- Avís - Aviso - Legal Notice - (LOPD) - http://legal.ifae.es http://legal.ifae.es/

vuillaut commented 2 years ago

Being able to apply the same method on both MC and real data is better imo.