CTShower Testing - Apr. 14, 2006

Mirror of the best candidate showers page.

CTShower Documentation

I separated the energy and angle fits, using the same method as is used in the chisq fits: first adjust only angle/time, then only energy, then only energy/core location. A few experimental reconstructions showed that leaving out the middle step made the reconstructor more prone to ending up way outside the array.

Examples posted on Apr. 12 were not optimized for Chiquita. Also, the choice of SOTC trial used for the final result was based on the location chisq from Brant's fit with the angle chisq from Shawn's fit. A possible reason for this is that Shawn's fit tends to return very large location chisq values. I have added the line which evaluates the location chisq from Shawn's fit, but the choice of SOTC starting point is still based only on the results from Brant's fit, so this does not affect which reconstruction is chosen as the final result.

I also fixed the bug that was causing some sites to not have a grey circle whey they should have. (It was an i < sites.GetLast() when it should have been i <= sites.GetLast(), causing the last site in the array to be skipped.)

Overall energy distributions:

Checking the sign of X and Y on the reconstructed showers... The code looks like it carries the sign through correctly, but there is just a strong tendency to reconstruct a shower at -X,-Y.