[Rivet] ATLAS prompt photon analysis

[Mike Hance]

Hi Frank,

> I have one more question: Is there any reason why you number the third 
> histogram as d01-x01-y04 instead of d01-x01-y03? I suppose on HepData 
> it's going to appear as d01-x01-y03 and since Rivet is supposed to be 
> compatible with the files exported from HepData that's also what it 
> would have to be in Rivet. If there is no reason for this, could you 
> make this small adaptation?
> I guess it also affects the following line in the code?
>   if(fabs(_eta_bins[i] - 1.37) < .0001) continue;

It's really only a book-keeping thing - we have a gap in pseudorapidity, from 1.37 to 1.52, in which we do not measure 
the cross section.  It doesn't much matter to me how the accounting for that gap region is done - we can easily modify 
the Rivet routine to skip the gap region, rather than consider it its own region.  It looks like you'd prefer it if we 
just have the three good regions be labeled "1,2,3" rather than "1,2,4", so I'll try to fix that tonight and get you the 
new code.

Thanks!

-Mike



> Cheers,
> Frank
> 
>>
>>
>>
>> Frank Siegert wrote:
>>> Hi Mike,
>>>
>>> Apart from the nb/pb issue the analysis seems to work well. Take a
>>> look at
>>> http://fsiegert.web.cern.ch/fsiegert/atlas/yjets/
>>> where I'm comparing it to the Sherpa 1.2.3 gamma+jets validation
>>> sample that we talked about in the MC validation meeting today:
>>> http://indico.cern.ch/conferenceDisplay.py?confId=127609
>>> Ignore the first one or two bins, since the sample has a pT_gamma>15
>>> GeV cut (on the ME level) which affects those. But apart from that it
>>> looks pretty good.
>>>
>>> If you want, I could add the analysis to Rivet's SVN now, or did you
>>> want to do any more checks?
>>>
>>> Cheers,
>>> Frank
>>>
>>> On 17/02/11 21:22, Mike Hance wrote:
>>>> Hi Frank,
>>>>
>>>> You're right, we should just put everything in pb. I took the numbers
>>>> from the tables in the appendix of the paper, which were in nb, and
>>>> mentally normalized away the extra 10^3 terms I saw in the rivet 
>>>> output.
>>>> Thanks for spotting that!
>>>>
>>>> I've attached a new tarball with that fixed, and with the
>>>> files/functions/plots/etc renamed to have the proper SPIRES ID (thanks
>>>> Jon!). I'm sure there are other mistakes though, so feel free to point
>>>> them out. In the mean time, I hope Giovanni (and others) will be 
>>>> able to
>>>> quickly spot-check the numbers in these files to make sure we've given
>>>> you the right numbers.
>>>>
>>>> Cheers,
>>>>
>>>> -Mike
>>>>
>>>>
>>>>
>>>> Frank Siegert wrote:
>>>>> Hi Mike,
>>>>>
>>>>> Thanks a lot and don't worry about it taking longer, it's a
>>>>> complicated analysis after all!
>>>>>
>>>>> One issue that I noticed without having done detailed testing yet: The
>>>>> .aida reference data seems to be in nb while Rivet's crossSection() is
>>>>> by default in pb and thus that's what the MC histos are being filled
>>>>> the way the code is now. Since the plot in the paper is also in pb I
>>>>> would suggest adapting the reference .aida file, or what do you think?
>>>>>
>>>>> Cheers,
>>>>> Frank
>>>>>
>>>>> On 17/02/11 17:05, Mike Hance wrote:
>>>>>> Hi All,
>>>>>>
>>>>>> I've attached a tarball that (I think) contains everything needed
>>>>>> to put
>>>>>> the ATLAS prompt photon results into HEPDATA, as well as a
>>>>>> corresponding
>>>>>> RIVET routine that seems like it might be doing the right thing.
>>>>>>
>>>>>> One thing I should note up front is that, as far as I can tell, the
>>>>>> paper doesn't have a proper SPIRES ID. I suspect this will change 
>>>>>> when
>>>>>> PRD finally publishes the paper (it's been accepted for
>>>>>> publication, but
>>>>>> I don't think it's actually made it into an issue of the journal 
>>>>>> yet),
>>>>>> so in the mean time, I've called the analysis
>>>>>> "ATLAS_2010_S1234567". My
>>>>>> attempts at calling the analysis something like
>>>>>> "ATLAS_2010_PROMPTPHOTONS" were rebuffed by RIVET, so I'm not sure of
>>>>>> the right thing to do in this case. If you want me to relabel it as
>>>>>> something else, just let me know; otherwise, feel free to call it
>>>>>> whatever you like.
>>>>>>
>>>>>> The contents of the tarball are:
>>>>>>
>>>>>> ATLAS_2010_S1234567.info - The .info file that describes the 
>>>>>> analysis.
>>>>>>
>>>>>> ATLAS_2010_S1234567.raw - A flat file that breaks down the different
>>>>>> (E_T, eta) bins and the corresponding measurements, uncertainties, 
>>>>>> and
>>>>>> so forth.
>>>>>>
>>>>>> ATLAS_2010_S1234567.dat - This is the 'final' flat .dat file
>>>>>> corresponding to the ATLAS measurements, for comparison with the 
>>>>>> RIVET
>>>>>> output.
>>>>>>
>>>>>> ATLAS_2010_S1234567.aida - The .aida file produced from the .dat file
>>>>>> above.
>>>>>>
>>>>>> ATLAS_2010_S1234567.plot - The .plot file for the three histograms
>>>>>> associated with this analysis.
>>>>>>
>>>>>> ATLAS_2010_S1234567.cc - The RIVET routine.
>>>>>>
>>>>>>
>>>>>> I've tested this running PYTHIA standalone, but the cross sections 
>>>>>> for
>>>>>> prompt photon production are low enough that the histogram never
>>>>>> really
>>>>>> gets populated. I also ran the routine using the ATHENA RIVET
>>>>>> interface
>>>>>> against some ATLAS photon+jet (EVGEN) samples, and the results seemed
>>>>>> reasonable, but I wouldn't be surprised if there are still some bugs
>>>>>> lurking in there. I'll try to test this stuff against some filtered
>>>>>> dijet samples next week.
>>>>>>
>>>>>> In any event, please feel free to let me know if there's anything
>>>>>> unclear/wrong with what I've done, and I'll try to fix it ASAP.
>>>>>>
>>>>>> Finally, thanks to everyone who helped (and bugged) us to get this
>>>>>> done
>>>>>> - I apologize it took us so long. Next time should be much faster, 
>>>>>> now
>>>>>> that we know what we're doing.
>>>>>>
>>>>>> Thanks!
>>>>>>
>>>>>> -Mike
>>>>>>
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> Rivet mailing list
>>>>>> Rivet at projects.hepforge.org
>>>>>> http://www.hepforge.org/lists/listinfo/rivet