Support #2761

Significant residuals in background simulations for limited radius at low energies

Added by Knödlseder Jürgen about 6 years ago. Updated about 6 years ago.

Status:ClosedStart date:11/20/2018
Priority:NormalDue date:
Assigned To:Knödlseder Jürgen% Done:

100%

Category:-
Target version:1.6.0
Duration:

Description

Anastasia Sokolenko reported some strange features when simulating background events for the 1DC South 50h IRF for a simulation radius of 0.5 deg. There seem to be energy dependent excess events at low energies compared to the predicted number of background events. Increasing the simulation radius, the effects are not significant anymore (see below).

residual_5.0deg_maxrate1_10subbins.png (33.9 KB) Knödlseder Jürgen, 11/20/2018 04:40 PM

residual_0.5deg_maxrate1_10subbins.png (31.8 KB) Knödlseder Jürgen, 11/20/2018 04:41 PM

residual_0.5deg_maxrate1_10subbins_seed1.png (31.8 KB) Knödlseder Jürgen, 11/20/2018 05:58 PM

residual_0.5deg_maxrate1_10subbins_seed2.png (32.7 KB) Knödlseder Jürgen, 11/20/2018 05:58 PM

residual_0.5deg_maxrate1_10subbins_seed-change.png (30.9 KB) Knödlseder Jürgen, 11/20/2018 05:58 PM

Residual_5.0deg_maxrate1_10subbins Residual_0.5deg_maxrate1_10subbins Residual_0.5deg_maxrate1_10subbins_seed1 Residual_0.5deg_maxrate1_10subbins_seed2 Residual_0.5deg_maxrate1_10subbins_seed-change

Recurrence

No recurrence.

History

#1 Updated by Knödlseder Jürgen about 6 years ago

I looked the entire day into this problem, making many cross and double checks, without finding any problems.

I noticed, however, that changing the seed value for the simulations had quite some impact on the results. The left panel below shows the result for seed=1 and the mid-panel the result for seed=2. I then incremented the seed value after the computation for an energy bin, and got the result in the right panel, which no longer shows correlations with energy.

So my conclusion is that the feature appeared due to running simulations with the same seed value. One should not forget that the random number generator returns “pseudo random numbers”, and the sequence of returned values is fully deterministic, and depends on the seed value. So running the generator for different energy bins and the same seed value give the same “trends”.

#2 Updated by Knödlseder Jürgen about 6 years ago

  • Tracker changed from Bug to Support
  • Status changed from In Progress to Closed
  • % Done changed from 90 to 100

Also available in: Atom PDF