add some comments about the priors in the chi2 analysis

1 files changed, 17 insertions, 2 deletions

diff --git a/utils/chi2 b/utils/chi2
index 7a8342d..bb62af3 100755
--- a/utils/chi2
+++ b/utils/chi2
@@ -45,8 +45,23 @@ def printoptions(*args, **kwargs):
         np.set_printoptions(**original)
 
 # Uncertainty on the energy scale
-# FIXME: These are just placeholders! Should get real number from stopping
-# muons.
+#
+# - the muon energy scale and resolution terms come directly from measurements
+#   on stopping muons, so those are known well.
+# - for electrons, we only have Michel electrons at the low end of our energy
+#   range, and therefore we don't really have any good way of constraining the
+#   energy scale or resolution. However, if we assume that the ~7% energy bias
+#   in the muons is from the single PE distribution (it seems likely to me that
+#   that is a major part of the bias), then the energy scale should be roughly
+#   the same. Since the Michel electron distributions are consistent, we leave
+#   the mean value at 0, but to be conservative, we set the error to 10%.
+# - The energy resolution for muons was pretty much spot on, and so we expect
+#   the same from electrons. In addition, the Michel spectrum is consistent so
+#   at that energy level we don't see anything which leads us to expect a major
+#   difference. To be conservative, and because I don't think it really affects
+#   the analysis at all, I'll leave the uncertainty here at 10% anyways.
+# - For the electron + muon energy resolution, I don't have any real good way
+#   to estimate this, so we are conservative and set the error to 10%.
 ENERGY_SCALE_MEAN = {'e': 0.0, 'u': -0.066}
 ENERGY_SCALE_UNCERTAINTY = {'e':0.1, 'u': 0.011}
 ENERGY_RESOLUTION_MEAN = {'e': 0.0, 'u': 0.0, 'eu': 0.0}