[Rivet] Shape comparisons

[Andy Buckley]

Not a waste of time: a well-formed question where we might actually be
able to help... that's the best kind of enquiry! Wish we got
(proportionally) more of them :)

Best wishes,
Andy


On 13/07/12 14:12, James Henderson wrote:
> Hi Andy,
> 
> I'm sorry, I seem to have been doing something incorrectly before as now
> everything works as expected after I re-built the rivet analysis. Most
> of the time I want the analysis to normalise to cross-section but there
> are a few times when I just wish to do a shape comparison so I wasn't
> keen to hard-wire the normalisation and yes, this is a personal analysis.
> 
> Thanks very much for your useful advice and sorry to waste your time,
> Jim
> 
> 
> On 13/07/12 14:43, Andy Buckley wrote:
>> Hi Jim,
>>
>> Rivet's been able to do that for some time, but it requires the version
>> of HepMC being used to support cross-section passing.
>>
>> So are you saying that passing an argument of -x 1.0 to the rivet
>> command line tool does not override the value that comes from the HepMC
>> input? I'd personally consider that to be a bug.
>>
>> However, modifying the analysis to hard-code the cross-section may not
>> be necessary. Is this a personal analysis or a built-in one? If a
>> personal one, and you always want to just do shape comparisons, then the
>> obvious thing to do is to put some normalize(...) calls in the
>> finalize() method. If it's an "official" analysis, or you don't *always*
>> want your plots to be normalised, then the rivet-rescale script allows
>> you to do post hoc normalisation: it's a bit hacky, since making a
>> complete programmatic interface via the command line is ~impossible, but
>> should be enough for simple cases like this. People on this mailing list
>> should be able to help with the latter if you have trouble.
>>
>> Finally, just a note that if you're only interested in shapes, then
>> there's no need for your analysis to make use of cross-section
>> information: just don't make any calls to crossSection(), and remove any
>> use of "NeedsCrossSection" from your .info file. But it's always a good
>> idea to at least normalise histograms to a fixed area rather than the
>> number of events used -- I see the latter all the time, but it just
>> makes it harder / more error prone to compare to other samples which may
>> be a different size -- perhaps to a subtle extent which doesn't
>> immediately set off warning bells!
>>
>> Cheers,
>> Andy
>>
>>
>> On 13/07/12 13:29, James Henderson wrote:
>>> Dear Rivet Experts,
>>>
>>> I have recently updated my Rivet version to 1.8.0 and it seems that now
>>> Rivet is able to automatically read in the cross-section of a sample
>>> straight from the HepMC file. I was trying to normalise two sample with
>>> different actual cross-sections to 1pb in order to do a shape comparison
>>> but I can't seem to overwrite the automatic cross-section finder.
>>>
>>> Is there a way to do such a shape comparison without explicitly altering
>>> the .cc file in order to not normalise to the cross-section?
>>>
>>> Thanks,
>>> Jim
>>>
>>> _______________________________________________
>>> Rivet mailing list
>>> Rivet at projects.hepforge.org
>>> http://www.hepforge.org/lists/listinfo/rivet
>>>
>>
> 
> 
> 


-- 
Dr Andy Buckley, SUPA Advanced Research Fellow
Particle Physics Expt Group, University of Edinburgh



The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.