[Rivet] Minor analysis changes

[Gavin Hesketh]

Hi Andy,
I think there are really three bits of info needed for each parton 
multiplicity: the cross section and # events with no cuts, which I 
currently get from the pythia/agile output; and the number of Z passing 
cuts, ie the denominator in the ratio, which is currently lost.
If you'll be at the atlas UK thing this week, it might be easier to sort 
it out then!
cheers,
	Gavin

On 04/01/11 18:54, Andy Buckley wrote:
> On 04/01/11 15:14, Gavin Hesketh wrote:
>> Hi Frank,
>> yes, exactly. You process individual samples with different # partons in
>> the final state all the way through pythia, rivet. Then (manually)
>> weight them to the same lumi and add together. So, (Z+j)/(Z) comes from
>>
>> a.(Z+j)1 + b.(Z+j)2 +...
>> ------------------------
>> a.(Z)1   + b.(Z)2 + ....
>>
>> where a,b are the lumi weights, (XX)1, 2 are from the samples with
>> different partons. If all you have is [(Z+j)/(Z)] for each of these
>> parton samples, I haven't figured out a way to do this...
>
> Rather than explicitly create new histograms that aren't in the paper:
> we could have Rivet write out the total cross section (without cuts) at
> the end of the run. You could then extract that from the log file if we
> make the printout format fully predictable. Would that help?
>
> Andy
>