You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm always frustrated when I have to guess the injection scale where I assume the sensitivity to be and run the injection steps to find out I was wrong.
This could be avoided by first estimating the scale using quick injections with a low number of trials.
That does however require the background TS distribution which would mean that the background TS calculation and the signal injections would have to be submitted in two subsequent steps and not, as it is now, at the same time.
The method implemented now does not require the background TS distribution!
It basically assumes a background TS distribution median of zero. For fit_weights that does not hold and gives an estimate that is too low.
I'm always frustrated when I have to guess the injection scale where I assume the sensitivity to be and run the injection steps to find out I was wrong.
This could be avoided by first estimating the scale using quick injections with a low number of trials.
That does however require the background TS distribution which would mean that the background TS calculation and the signal injections would have to be submitted in two subsequent steps and not, as it is now, at the same time.
sensitivity.pdf
The text was updated successfully, but these errors were encountered: