It's been assumed that Bashful was halfway out of the beam during the decay channel running. This doesn't appear to be the case. For my MC studies, I have assumed that Bashful has half thickness but is spread across the beam uniformly (which is certainly wrong). Anyway, this would give me the right number of events in Bashful and the right Y-distribution, but the wrong X-distribution.
What I see is too many events in Bashful altogether. See the vertex Z-distribution for events which pass NU_CANDIDATE as well as tight veto wall activity cuts and track angle cuts. Note that the MC is normalized to the number of events in the calorimeter (the "absolute" normalization we used for the Karmino cross-checks). Zooming in on Bashful, we can see that most of the excess peak really is coming from the chamber, not the veto wall upstream. The black histogram on that plot is the data with veto cuts removed -- the events which appear have vertices in the veto wall, which is upstream of Bashful.
The next plot is the X-distribution of events in Bashful. These are events with a Z-position within 20" of Bashful and a vertex Z sigma < 10". I see 48 such events (compared to an expected 20 from the MC). Note that the MC I show is actually the Y distribution, since I got bitten by the Turtle X/Y swapping annoyance. The thing to note is that the data are NOT really asymmetric in X. In fact, half the events are at positive X, where Bashful isn't supposed to be.
Checking the spec file (dcpos.spec0017) shows the first wire of Bashful at X = -110.3". This puts the other edge of the active area of the chamber at X = +7.2". Assuming a 3-inch frame (this is what Sam guesses is the size), the material should end at X = +10.2". But there are plenty of events out to +20". This may be due to a unistrut frame next to the chamber??Eric D. Zimmerman (firstname.lastname@example.org)