[Rivet] Z projection on Rivet

[Frank Siegert]

Hi Gavin,

On 31/01/11 13:37, Gavin Hesketh wrote:
> A couple of points for the Z projection. First, in the analysis
> D0_2010_S8671338.cc (inclusive Z pT), the setup needs to be adjusted.
> Currently:
> ZFinder zfinder(-1.7, 1.7, 15.0*GeV, MUON, 65.0*GeV, 115.0*GeV, 0.2);
>
> ie sum all photons in a cone of 0.2 around the muon. This does not
> correspond to the analysis, which did not sum any photons into the
> muons, so:
> ZFinder zfinder(-1.7, 1.7, 15.0*GeV, MUON, 65.0*GeV, 115.0*GeV, 0.0);
>
> This actually makes a significant difference in the first few Z pT bins.
> I realise ATLAS are handling this slightly differently...

Thanks for finding this (I don't remember why I put 0.2 in there, maybe 
because back then we had been discussing that that should be the ideal 
definition?). Just out of curiosity: Do you happen to have a plot 
showing the change?

> Then some fine-tuning which is a much lower priority. At the moment, it
> looks like the Z projection calculated the invariant mass before adding
> on photons, then adds on photons to both electron and muon channels. So:
> 1) calculate the mass after adding the photons, if the photons are going
> to be added (probably a tiny effect)

That's going to be relatively expensive, and will make it more difficult 
to re-use existing projections like InvMassFinalState. But I guess 
that's what it'll have to be.

> 2) for D0 Z->mm, don't add the photons; for ATLAS Z->mm and all Z->ee,
> add the photons. I'm sure there is an easy way to code a switch here.
> Note that for the D0 Z->mm, still need to select the photons in a 0.2
> cone and exclude them from the jet finder.

Ok, I'm starting to wonder whether we should put all of this into the 
ZFinder projection or give up having a uniform ZFinder and do all of 
this stuff "manually" in the analyses. Opinions?

> The current analyses
> (D0_2009_S8349509.cc, D0_2008_S7863608.cc) instead apply a cut on
> jet-muon overlap, which which was not done in the data analysis.

Right, thanks. That's my mistake, I didn't see the clear specification 
of particle-level cuts which don't contain the overlap cut anymore. It's 
good to have the author of these analyses checking them thoroughly :)

I'll wait for opinions on the above before I go about fixing these.

Cheers,
Frank