Matching Studies
  
 No-match events studies and good events situation (11/29/06)

  MC numbers are from top175 minintuple_ttopag_loose_pretag.root
  Cuts are Ntight=4 nLoose=0 or nLoose>=0


           TotalMC |  l+jets     | chis<200   |

 nLoose>=0   13383 | 11700 87.4% | 8345 62.3% |
 nLoose=0     8520 |  7409 87.0% | 6004 70.5% |

 REAL EVENTS

             Total | signal| match.|range(m)| rest|range(all)| backg.|range(with back.)
 nLoose>=0     183 |   151 |  94   | 2.0?   |  57 |  ??      |  32   |
 nloose=0      116 |    99 |  72   | 2.3 GeV|  27 | 2.5GeV   |  17   | 2.8 Gev

Comments:
1.  Background and total events for nLoose>=0 estimated from MEAT,
    adding whatever was necessary to get rid of the QCD cut ( 60% QCD
    background rejected, 7% of other back. rejected, 5% signal rejected).

2. Ranges for signal shown above for nLoose=0 recalculated from John's results
   (they were for 116 events, l plus jets fully matched, or total sim. events).

3. range for the nLoose>=0 are calculated the same way as in point 2. above.
   Not clear that this is correct

4. Range (with back.) obtained by Paul with full background handling.

BOTTOM LINE: We do not seem to do better than MEAT at this time.
             They have stat error of 2.5 GeV after taking care of no
             match and backgrounds. We have 2.8 for nLoose=0. Can we
             improve on this by getting the ISR background included
             and finding a way to have an improvement on the taus?

             We can try to increase statistics by adding the nLoose>0 events.
             Not clear that we can get below 2.8 if we use also the
             nLoose>0, as things may get worse because of the extra
             jets. But we should try.
 
PLOTS:

  Top plots: l+jets, no matching. Middle: l+jets, chisq<200.
  Bottom plots: one jet matches ISR parton better.  
  MW and Mtop for nloose=0  

   MW and Mtop for nloose>=0