[Rivet] [Professor] chi2 distribution in Profile1D

[Andy Buckley]

Hendrik Hoeth wrote:
> Thus spake Hendrik Hoeth (hendrik.hoeth at cern.ch):
> 
>>>>> I just came to my mind that for the actual degree of oversampling
>>>>> we should probably bin in (N_runs-Nmin)/Nmin.
>> Like this?
>> http://users.hepforge.org/~hoeth/plots/oversampling2/

Cool. So the scaling is certainly better than linear, but not enough to 
make cubic preferable to quadratic for a fixed number of runs, in this 
case at least.

> Maybe N_runs/N_min would be even better. It's the same as
> (N_runs-Nmin)/Nmin, but the x-axis is shifted by 1. That brings us back
> to the "oversampling by a factor of XXX" which we were talking about in
> the first paper ...

Yep, sounds reasonable.

Andy

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

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