Track finding was done using the September 16 build of t00.59.00. The usual 50, 3 and 1 GeV muon events were reconstructed using CFT+SMT. Here are the results:
run CPU eff fake misr fcsq mcsq nclu ---- ---- ---- ---- ---- ---- ---- ---- m50 487 .966 .000 .007 21.1 8.1 20.2 (plots) m03 492 .956 .000 .009 21.7 9.3 20.2 (plots) m01 450 .897 .000 .010 25.7 9.0 20.2 (plots)Here are results for CFT only matched with the MC track at the inner CFT layer:
run CPU eff fake misr fcsq mcsq nclu ---- ---- ---- ---- ---- ---- ---- ---- m50 304 .966 .000 .000 12.2 5.2 16.0 m03 310 .956 .000 .001 12.5 6.4 16.0 m01 308 .897 .000 .001 16.3 7.0 20.2Comparing with t00.58.00, there is a little degradation in the fitting for CFT only. The effect is worse with decreasing pT. This may be due to bug fixes in the handling CFT multiple scattering in the reconstruction.
There is still an efficiency loss in the CFT. Much of this cones from problems in the detector simulation. This should be fixed in the next release.
The CFT+SMT fitting looks better but still the match chi-square is still high and has a 1% misreco tail.
Here are the rcp and obs files used in the above studies: