Comments and questions to the NPE analysis by D. Kikola, 2021.04.09:

1. Electron identification with BEMC and trigger efficiency
a) Can you please prepare plots with p/E distributions from experimental data, which you use for electron identification?
b) The analysis relies on simulations to calculate BEMC efficiencies (for electron identification and the trigger). Can you please prepare a plot with a comparison of BEMC data (energy and p/E) and simulations to demonstrate that there is a good agreement?

2. Systematic uncertainties
I support the general approach outlined in Sec. 3.7. However, it produced results that rise a few questions:
a) Fig. 31 and Fig. 33: the overall systematic uncertainty is very small (on a per mille level) and it looks like it is 0 for some bins.
On the other hand, the recent report of the TPC tracking task force showed that the systematic uncertainty on the tracking is ~2% (https://drupal.star.bnl.gov/STAR/meetings/star-collaboration-meeting-march-2021/plenary-session-2/tpc-tracking-task-force-report-0, slide 10). Let’s try to find out which value is more reliable.

b) Fig. 43 and 44: Systematic uncertainties on BEMC at high pT are 0, which suggests we have a perfect understanding and simulations of the BEMC response to electrons. I suggest we review these results.

c) Fig. 46: can you please put values from this plot in a table, similarly as you did for PHE reconstruction efficiency in Tab. 5?

d) Error transfer formula (Eq. 16 on page 59): In general, the error propagation formula for a function f(x,y,z,..) is sigma^2 = (df/dx)^2*sigma_x^2 + (df/dy)^2*sigma_y^2 + (df/dz)^2*sigma_z^2 + ...

For some functions it indeed simplifies to Eq. 16 in the technical note, which is a sum of relative uncertainties in quadratures. If I’m not mistaken, it is not the case for the NPE yield calculation with formula NPE = (N_{inc. e}*p - Npho)/epsilon.

For example, if we only have an uncertainty on purity p, then sigma_NPE = [d(NPE)/dp]*sigma_p = N_{inc. e}/epsilon*sigma_p, which is not NPE*sigma_p/p.

d) Systematic uncertainty due to HDE yield is tiny (below 0.15%). The HDE yield is calculated using experimental data as an input (which have uncertainty on the level of 10% or larger for J/psi) and the NPE/HDE ratio is ~16%. Thus I do understand why overall syst. uncertainty on of the HDE components is so small.

Comments to analysis description in the paper:
l. 286 – 301: motivation for the rapidity distributions of π0 and η (based on Landau hydrodynamics) seems far-fetched, given the data are from p+p reactions. I would add a comment why this assumption is reasonable for p+p data.
l. 439: Connection between luminosity and the equivalent number of minimum bias events of the collected data is not provided.
Fig 1: purity: I would add a comment about a peak at 5 – 10 nsigma_e