Hi Maarten, Gerco,

Here are my comments:

Physics: IN the Bs->emu paper we talked a lot of leptoquarks, here it's just R-parity susy. Also the Bs paper should be cited. Are the models you constrain still unaffected by other measurements?

Intro: the physics motivations are a bit short. L.76: You may want to say a bit more on how electron ID is done. What's the efficiency? What's the mis-ID? This is critical here. L.83: why do you do the multiple candidate selection before the BDT? It must be more efficient to do so afterwards. L.127: why train separately? Wouldn't adding the year to the variables do the same? L.219: Computing the difference between weighted MC and unweighted MC is never a good way of determining a systematic. What's relevant is the uncertainty on the weights, not the difference with 1. This uncertainty could be L.229: 30% is a lot. Is that consistent with what we wrote in the Bs->emu paper? L.253: products of which couplings? You say too much or too little.

Abstract: it says "further constrains" here and in conclusion "will help to further..." which much weaker. We should say the same.

Text: L.8: varying by -> spanning over L.29: move ref [8] to where you explain BDTs. L.31: is -> was Fig.1 : It's not arbitrary units, it's dP/dm(emu), see paper 2014-020 for examples. Add a legend explaining which histo is which. Finally change green to another colour. The SB dislikes green. L.91: For simulated samples calibrated with data L.119: why "chi-square" and not $\chi^2$? Use \chisqip macro. L.145: The track multiplicity -> particle multiplicity L.152: events -> candidatea L.153: to the electron. L.171, 180, 185, 207, 243: events -> decays Fig.2 and 3: The labels are much too small. L.210 and 212: D->pipi decays L.241: what is the "Kpi channel"? Be precise L.243: do you need to repeat the fit result? L.250: you should give the expected limit too. L.252: repeat the Belle limit here. Else there's no place where they are long each other. L.253: "supersymmetric models of non-SM physics" does not mean much.

Ref [4,5]: add collaboration names

Cheers,

Patrick