aMC@NLO Higgs production not correct

Hi Ralph,

I think that the numbers you find are quite reasonable.

First of all NNLO QCD corrections to the inclusive Higgs cross sections are still quite large; if you check the literature, you'll find something around +25% with respect to the NLO result at the LHC (see for example arXiv:1101.0593).

Then you need to consider the fact that the numbers you find online are for the *total inclusive* cross section, while the cross section computed by aMCatNLO contains a pT cut on real radiation (pT jet >10 GeV if you used the default run_card), so is somewhat smaller.

I think that these two facts account for the largest part of discrepancy that you see.
Other minor sources of discrepancy are the PDFs, the value of alphas(mZ) and the central scale choice.

In the end, the aMCatNLO cross section should be as accurate as an NLO result (so not super precise in worst-case scenarios like higgs production, where NNLO corrections are still important); and remeber that the number you get is not the total inclusive cross section, but the total cross section after the cuts you imposed.

I don't know if Marco or Kentarou have somethinge else to add to my answer, anyway I hope it helped to clarify the situation...
Cheers,

Federico

Hi again,

I played a little with aMCatNLO in order to see the impact of the various factors I mentioned before.

I must correct some of my previous statements.
1) Lowering the pT cut on real radiation (from the default value of 10 GeV down to 2, for example) doesn't affect the cross section by much; so this is a minor source of discrepancy and I apologize for my wrong statement.
2) The choice for the central value of the scale has in fact a large impact on the final number, not a minor one.
Anyway my conclusions don't change: NNLO corrections are sizable, so you can't expect that the numbers in the twiki agree with the ones from aMCatNLO.

As a further simple check of the aMCatNLO result, you can use the Higgs cross section online tool by Grazzini
http://theory.fi.infn.it/cgi-bin/higgs.pl
(there are also other ones, like ihixs by Anastasiou).
If I consider the mtop->inf. limit, use mstw2008 pdfs, and fix the scales to muF=muR=mH (=125), I get
35.554 pb (Higgs cross section)
3.580e+01 +- 3.5e-01 pb (aMCatNLO)
that are in very good agreement.
You can also see that if you choose a different scale (like mH/2) the cross section is significantly affected, and that the NNLO cross section output by the Higgs cross section tool is much larger than NLO.

Cheers,

Federico