MC studies of delta(t) resolution function

I looked at the delta(t) resolution function for  signal and background events.
 

Fits done with a double gaussian fit  - 5 free parameters
Here is a sample of the fitting results:

Signal Events (used b0 signal MC -> fit for signal only, no background contamination)
 

# events in fit Category bias1 width1 bias2 width2 fraction1
2615 total -.10 .69 -.315 2.61 .77
1240 kaon -.11 .70 -.228 2.39 .77
406 lepton -.07 .59 -.243 2.43 .70
391 net1 -.05 .60 -.363 2.30 .76
578 net2 -.13 .77 -.539 3.44 .80

Signal Event pull fits
 
 

# events in fit Category bias1 width1 bias2 width2 fraction1
2611 total -.25 1.54 -.70 5.56 .82
1239 kaon -.27 1.52 -.90 5.53 .83
405 lepton -.40 1.92 .69 7.69 .85
391 net1 -.15 1.20 -.58 4.24 .67
576 net2 -.21 1.35 -.82 4.28 .78

Background plots:
 
 
# events in fit Category bias1 width1 bias2 width2 fraction1
440 total -.23 .89 .28 3.0 .81
226 kaon -.18 .83 -.28 2.9 .81
27 lepton -.28 .61 .23 3.0 .47
65 net1 -.32 1.1 2.0 1.9 .85
122 net2 -.23 .90 .44 3.1 .83
 

Background pulls
 
 
# events in fit Category bias1 width1 bias2 width2 fraction1
440 total -.42 1.9 .48 7.9 .92
226 kaon -.51 1.9 2.6 6.6 .93
26 lepton -1.4 1.6 -.31 3.2 .43
65 net1 -.56 2.3 1.0 1.8 .97
123 net2 -.21 1.5 -5.0 12.0 .93

Background residuals
 
573 fake_total -.11 .80 -.41 3.3 .70
69 flip_total -.59 1.1 .74 4.5 .77

Background pulls
 
 
566 fake_total -.18 1.5 -1.1 7.5 .67
68 flip_total -.33 1.1 -1.4 3.2 .64

Pending more information on the lepton tagging category, it seems reasonable to use the same resolution funtion for the different tagging categories.  This is very useful because:

More results to come, I want to see what happens with larger data sets and dm_refitting.