[Rivet] Rivet bug in LEP Jet Resolutions Y12 ?

[Peter Skands]

Hi Frank,

Don't know if it helps you, but here is a code snippet from my own 
program that computes the rates and appears to get both the Y and Rates 
correct. The starting point are merging scales extracted from FastJet, 
as I believe you are probably also doing. The inputs to FastJet were 
events that include neutrinos, with ISR switched off.

     // Extract merging scales
     double lny56 = log(clustSeq.exclusive_ymerge(5));
     double lny45 = log(clustSeq.exclusive_ymerge(4));
     double lny34 = log(clustSeq.exclusive_ymerge(3));
     double lny23 = log(clustSeq.exclusive_ymerge(2));
     double lny12 = log(clustSeq.exclusive_ymerge(1));

     // Fill Jet resolution histograms with results from FastJet	
     // (wYij are just constants that take the bin sizes into account)
       HLnY12->Fill(-lny12,wY12);
       HLnY23->Fill(-lny23,wY23);
       HLnY34->Fill(-lny34,wY34);
       HLnY45->Fill(-lny45,wY45);
       HLnY56->Fill(-lny56,wY56);

     // Histogram jet rates
     // Loop over resolutions
       Hr1j->Reset();
       Hr2j->Reset();
       Hr3j->Reset();
       Hr4j->Reset();
       Hr5j->Reset();
       Hr6j->Reset();
       int iy=-1;
       for (double lny=-10.1; lny<0.0; lny += 0.2) {
	iy++;
	if (lny12 < lny) sumN1j[iy] += 1;
	else if (lny23 < lny) sumN2j[iy] += 1;
	else if (lny34 < lny) sumN3j[iy] += 1;
	else if (lny45 < lny) sumN4j[iy] += 1;
	else if (lny56 < lny) sumN5j[iy] += 1;
	else  sumN6j[iy] += 1;
	double err, val;
	// 1j (sumw = number of events for unweighted sample)
	val = sumN1j[iy]/sumw;
	err = sqrt(sumN1j[iy])/sumw;
	int iBin = Hr1j->FindBin(lny);
	Hr1j->SetBinContent(iBin,val);
	Hr1j->SetBinError(iBin,err);
	// etc for 2j, 3j, 4j, ...
       }
     }



On 3/11/11 5:28 PM, Frank Siegert wrote:
> Hi Peter,
>
>> I think the same bug as in the Y1, Y2, ... may also be present in the
>> jet rates, see, e.g.,:
>>
>> http://mcplots-dev.cern.ch/?query=plots,ee,zhad,rate-2jet,Main
>
> Hmmmm... I admit that they do look strange, but it can't be related to
> neutrinos or charged/neutral particles, since they are being filled with
> the same clustering value as the differential distributions, which as
> you say work fine now.
>
> I guess I have to look at how they are being filled again and check
> whether that makes sense. It's probably worth filling them by
> integrating the differential distributions in the end, instead of "on
> the flight" as they are now, but that'll take me a bit of time to check
> and implement.
>
>> On the good side, the Y1, Y2, etc, now really do look like they should
>
> Cool, thanks for the confirmation plots!
>
> Cheers,
> Frank