CHICOS Analysiscut2

Running analysiscut2 on old data:

I am updating autoFit4 to hopefully make it able to go through the list of new showers and find a new set of best candidates automatically. Changes include:

• Cut on radius at 8 km. <- Modified from 10 km.
• Cut on time at 20 us.
• After first reconstruction, deletes hits outside 2-sigma time errors. (There is no requirement that the reconstruction match the intensity well.) Repeats reconstruction and checks again. (Generally one iteration is all that is required.)
• Ignores showers that have fewer than 3 match sites that reconstruct within 2-sigma error bars. (Currently does not check whether the trigger site reconstructs well.)

To get an idea of how this code performs, I ran it on the current set of 120 Best Candidate showers. Breakdown of results:

• One Chiquita shower:
• 06-14-03 Sample rext is a Chiquita shower.
• 5 Showers which do not pass analysiscut2:
• 06-18-03 Included hit at -22 us; otherwise OK.
• 11-20-03 Previously missing; reason unknown.
• 10-19-03 Previously missing; reason unknown.
• 08-24-06 Included hit at 22 us; fit not great.
• 09-12-06 Included hit at 23 us; otherwise OK.
• 16 Showers which pass analysiscut2 but do not reconstruct with 3 matches within 2-sigma errors:
• 01-29-03 Was only ok because it ignores a healthy unhit site.
• 03-03-03 Sample rext had only 3 sites.
• 05-17-03 Had poor time fit.
• 03-01-04 One site off by 2 us, one site off by 4 us. Otherwise good fit.
• 02-01-05 One site off by 3 us when using v3.03. Better fit with v2.35.
• 02-16-05 Sample rext had only 3 sites.
• 07-16-05 Had poor time fit.
• 08-09-05 Had poor time fit.
• 09-23-05 Westridge gets hit twice - need to remove one set of hits to get good fit.
• 10-21-05 Need to use SOTC 50 to get a good fit.
• 11-16-05 Had poor time fit.
• 01-02-06 One site 3 us off. Otherwise good fit.
• 03-21-06 Loses 1 hit at 23 us. Fit to remaining sites not as good.
• 05-15-06 Had marginal time fit.
• 06-19-06 Had poor time fit.
• 09-10-06 Fit with v2.35 was better.
• 98 Showers which pass with good data.

Most showers looked fine. Exceptions:

• 06-27-03 Fit is off, energy overestimated. Shower should be cut.
• 12-26-03 Stuck on a site, energy possibly underestimated.
• 05-24-04 Removes too many CIT 1000s. Not sure if it is possible to check each site individually.
• 06-01-04 Keeps Granada Hills (9 us off). Need START to fit more points. Version 2.35 got better fit.
• 11-29-04 Needs START to get good fit.
• 12-12-04 SOTC/START helps. Version 2.35 got better fit.
• 02-06-05 Keeps hit at Huntington (just under cutoff of 8 km). Fit is better without it.
• 03-03-05 Removes hits from Arcadia and HugoReid that are probably ok. (Very close together.)
• 06-01-05 Hard to fit well. Sample rext was better.
• 06-28-05 Not great fit, but similar to sample.
• 07-09-05 Not great fit, but similar to sample.
• 07-26-05 Not great fit, but similar to sample.
• 09-20-05 Very bad fit. Eliminates too many hits.
• 11-08-05 Keeps hit at Marshall (4 km away) but fit is not too bad.
• 11-30-05 Difficult to get good rext because trigger hit is ignored (both hits >1000). Sample rext was done with dummy.
• 12-12-05 This is ok but for some reason, the color of South Pasadena reflects the average of 2 hits instead of the 1 that is kept.
• 12-26-05 Sample rext does not keep Southwestern, making it a Chiquita shower. Fit in sample was not much better.
• 03-10-06 Removes too many CIT 1000 hits. Version 2.35 fit is somewhat better.
• 04-16-06 (Wrong image.) Core placement in sample rext is odd also.
• 04-17-06 Again missing some CIT 1000s. Sample rext misses trigger also.
• 06-04-06 Not great fit, but similar to sample.
• 06-23-06 Not getting the same fit via web interface??
• 08-30-06 Not great fit, but similar to sample.
• 10-03-06 Removes too many hits. Should be cut on the basis of intensity fit (all 0).
• 01-18-07 Correctly removes Pasadena, incorrectly removes Marshall.
• 02-03-07 Blair wrongly deleted. Could be fixed by making a minimum allowed time spread of ~2 us.
• 03-05-07 (Wrong image.) Keeps PCC, 9 us away. Make cut on largest acceptable discrepancy?
• 03-17-07 Not great fit, but similar to sample.

Notes:

• Some error bars are too small and things are getting thrown out when they should be kept. Suggest making cut on 2-sigma errors or 2 us, whichever is larger.
• Some showers get the intensity fit all wrong. Throw away any showers that have all 0 intensities.