[Rivet] Deciding on a dR cut for Photos

[Peter Skands]

Hi,

Just a thought. I'm worried about pi0's (and etas) looking like photons. That 
could make a large cone dangerous, unless pi0's can be kept stable in the 
generators.

Also, the DR of the photons could well be a function of the pT of the leptons, 
with more boosted leptons sending their photons into a narrower cone, and slower 
ones dumping them all over? But I guess looking at pp->W gives a lepton spectrum 
sufficiently similar to that of Z that we don't have to care so much about that. 
Ideally, perhaps you could set mW=mZ for these studies?

Cheers,
Peter

Thus spake Emily Nurse :
> Hi all,
> 
> The CDF and Dzero analyses that measure the Z pT distribution attempt to 
> correct for QED effects by using Photos to correct the l+l- pT 
> distribution back to the true Z pT. We therefore have to cluster back 
> any photons that we think Photos would have simulated and add them to 
> the lepton 4-momenta in the relevant Rivet routines. In order to do this 
> we have to define a reasonable cone size to use, where any photons 
> within that cone around the lepton are clustered back in. This is true 
> for both electrons and muons.
> 
> In order to address the issue of where to put the dR cut I simulated  
> W->enu and W->munu events using Photos and made a plot of dR between the 
> lepton and any photons. The first attached plot shows the dR 
> distribution for photons in W->enu and the second is for photons in 
> W->munu.
> 
> The table below shows, for different dR cuts, the fraction of photons 
> with dR < cut and the fraction of events with dR < cut (this third 
> column includes events with NO photons in the numerator and 
> denominator). I think the third column is the number we care about.
> 
> ELECTRONS
> ***************
> dR cut     % of photons     % of events
> 1.0            94.3                  98.3
> 0.5            89.0                  96.8
> 0.2            80.7                  94.4
> 
> MUONS
> *********
> dR cut     % of photons     % of events
> 1.0            87.1                   98.0
> 0.5            76.5                   96.3
> 0.2            60.1                   93.7
> 
> 
> Andy rightly pointed out that to do this properly we need to also get an 
> idea of the number/energy/distribution of photons from other sources 
> (initial state radiation, within jets etc) that should not get clustered 
> back in. This I am not set up to do that easily and it will be a more 
> detailed study. Perhaps we can see how much the result changes as we 
> vary dR, if it changes a lot we need to consider it further?
> 
> In the mean time I would propose 0.5 as a default, but this is clearly 
> not optimised.
> 
> Cheers,
> Emily.
> 
> 
> 
> ------------------------------------------------------------------------
> 
> 
> 
> 
> ------------------------------------------------------------------------
> 
> 
> 
> 
> 
> 
> 
> 
> 
>