Mirror of the best candidate showers page.
The Kolmogorov-Smirnov Test in HitLogLikelihood_Primitive: (Code.)
I updated the write-up of what is going on in the code to include a description of the KS test (page 2).
As Chris suggested, I integrated the TDF by hand and wrote it into the code, rather than doing numerical integration. This made it run much faster. The form of the integral is shown here.
Just to check that the code is doing the right thing, I made a few plots of the TDF and the integral over time. These are all for logE = 20, cosTheta = 0.95. The plots show the TDF (black) and the integral (blue) versus time for a few values of rperp:
Reconstructed showers (compared to most recent version of 3.01):
01-29-03 Similar.