The "three air core toroids" is confusing, as we have 24 coils. (And I know this is one version of the ATLAS boilerplate, as I have made this comment several times.) I asked people around ATLAS to answer the question "ATLAS has N air core toroids. What is N?" and get 8 and 24 as answers, but never 3.

This statement is essentially 100% also in the detector paper: one barrel and two endcap toroids.

Luminsoities with 3 significant digits in the plots is a bit much. (With 3.4% luminosity undertainty, we're just at +/- 1 pb) Also, we should be consistent with the number that we use: some muon plots use 33, others 32.6.

all plots are redone with 2 significant digits.

In Figure 1, we're seeing that the shape isn't quite as predicted. (We've seemn this before, but the effect is getting more strinking) What, if any, affect does this have on the A's and C's? (The electron spectrum, for example, is harder than predicted) Which uncertainty in Table 4 accounts for this? Electron energy scale or electron reconstruction?

Electron energy scale would affect e+ and e- the same direction. This effect is covered by MC model and PDFs, as we discuss e.g. in l.270-273, and therefore goes into A and C uncertainty.

There are too many digits of precision in Table 1. Also, column one should have no uncertainty. We observe exactly 72207 W+ candidate events. There is a statistical uncertainty of 1/sqrt(72207) that must be applied to any inference we make from this number, but there is no question that we observed this many. Same comment for Table 5.

Ok, we remove the sqrt(N) uncertainty.

Are the y-axes "entries" or "events"?

We use "Entries / [bin width]" in accordance with ATLAS plot Style guide.

Are there any WW or ZZ events in this sample?

Yes, we subtract these from MC samples. Dibosons are on the 0.1% level.

Figure 5 (lower bottom) has a half-invisible table on it that should probably be removed.

The wording in the conclusions is very confusing: "The Z -> mumu cross section is measured with an experimental precision of 1.1%" is in conflict with a 3.4% luminosity uncertainty. Now, I think I see where you are going with this: "apart from the luminosity, which we have little control over, we can measure the cross-section to 1.1%". But luminosity is experimental. I wouldn't go down this road in this part of the text: the information is all in Table 7 and in Section 5 anyway. I'd also drop the "experimental accuracy" part of line 398. (And it's precision, not accuracy)

The important property of the luminosity uncertainty is, that it is 100% correlated across all possible cross section measurements. Thus, it is really a special uncertainty. We mention in several places, that the luminosity uncertainty is 3.4% and enters all cross sections.

Figure 8 is really, really impressive. However, there is so much text on the plots that the data occupies only a time part of it. I wonder if this can't be improved. A similar comment with Figure 9: the data band occupies only 10% of the plot.

axes have been zoomed

Did we do the predictions in Table 11 ourselves? If so, is that because no published calculation exists?

There was a big effort to understand properties and limitations of various theory programs, ATL-COM-PHYS-2010-695. One often cannot use published values, because e.g. different programs, different SM parameters etc. are used and thus values are not really comparable.

-- MassimilianoBellomo - 07-Mar-2011