From: 	  ghill
  Hi Aya, and the EHE gang - here are some questions about the analysis.

   Thanks,
   Gary

 1. How did you actually fit the E_th and alpha from the data? I don't see errors from a fitting of the model (e.g. MINUIT output errors). Also the first bin in the fg-cos theta -1...+1 contains 300 events for data, and the model predicts 100. So normalisation wasn't allowed to float, or that bin wasn't used. How was data used in the fitting?

 2. What is the criterion for choosing the final cut? Is the event passing rate plot showing integrated events above each cut level? So if you choose level 8.5 you expect 0.02 GZK above that cut? If the expected  background is 10^-3, you are cutting too hard - the F/C 90% limit would be 2.44 for a MRF of 2.44/0.02. If you relaxed all the way back to level  0, you still only expect 0.1 background, and thus basically the same limit but you double the expected signal and halve the MRF for a expected limit that is twice as good.

 3. Sensitivity - at cut 8.5, you are at about 100 times the GZK flux (2.44/0.02), 50 times GZK at cut zero. What would the MRFs be for the other models shown (Z-burst, TDs, etc)? Is there a table of predicted numbers for those fluxes somewhere too? What about the limit for a pure E^-2 spectrum, as a benchmark to compare to AMANDA, IceCube, RICE and ANITA-lite (which is at the 10^-6 level for their E^2 limit in the energy range (10^9-10^15 GeV).