[Rivet] Problem with Rivet LEP Four Jet Analysis

[Frank Siegert]

In addition to what Hendrik pointed out: Do you happen to be using *all* 
particles in your internal analysis, or only the charged particles?
I also saw discrepancies between my own implementation of this analysis 
and the Rivet one, and it turned out to be important to restrict your 
analysis to charged particles only. Rivet uses charged particles and from 
what I've been told by the author of the analysis this is what it should 
be like.

Frank

Simon Plaetzer, Wednesday 23 September 2009:
> Dear Rivet developers,
>
> when running the LEP four jet correlation analysis the Rivet output is
> not compatible with the builtin Herwig++ analysis. Besides the fact
> that Rivet seems to use much less events than the Herwig++ analysis
> selects for the observables, the shape of especially \alpha_{34} is
> completely different.
>
> Attached you can find the plots from the Herwig++ analysis and the
> Rivet analysis, originating from the very same run of 500k events (in
> this case Herwig++ 2.4.0 and Rivet 1.1.3, both using fastjet 2.4.1).
>
> The Rivet analysis issues the warnings
>
> Rivet.Analysis.DELPHI_2003_WUD_03_11: WARN  numdurjets   = 36515
> Rivet.Analysis.DELPHI_2003_WUD_03_11: WARN  numjadejets  = 63975
> Rivet.Analysis.DELPHI_2003_WUD_03_11: WARN  sumofweights = 499960
>
> Best,
> Simon