Yep, this is a known problem. I believe we have a fix. This was a feature
of gismo wherein it trimmed the MC tree to a given number of generations to
save space in memory.

 The fix is to only trim (all) children of particles which interact in the
calorimeters and leave all other particles alone. I'm testing this fix now.

Richard

Eugene Guillian wrote:

> Hi everyone,
>
> I'm writing regarding some troubling observations I've made about
> gismo-simulated events -- specifically, regarding monte carlo
> information in an event.
>
> There are several sources of monte carlo information in a given event.
> To be specific, let me talk about ttbar events in the lepton + jets
> decay channel, so t1 -> lep + nu + b1 and t2 -> u + d + b2.  I'll call
> these 6 decay products "primary decay products".
>
>   (1) LCDEvent has a list of monte carlo particles.
>   (2) Each of the primary decay products has a list of daughters.
>   (3) Each calorimeter cluster has a list of associated monte carlo
>       particle.
>
> Using (2), I made a list of FINALSTATE particles with no daughters.  I
> found that the b-quark has many such particles.  However, bbar almost
> always has zero such particles.  The up-type and down-type particles
> from hadronic W decay comes somewhere in between -- often they would
> have zero such particles, but not as often as bbar.  The leptons looked
> normal -- i.e. neutrinos had only one such particle, electrons had quite
> a few (electron + brem. photons), muons usually had one or several (muon
> + brem. photons), and taus ranged from few to dozens.
>
> The above observation leads me to think that monte carlo information is
> being truncated somehow.  In a separate study, I examined the
> association between lists (2) and (3).  I found that there is a 1-to-1
> correspondence between particles in lists (2) and (3).  This suggests
> that, not only is the list of monte carlo particles being truncated, but
> the missing monte carlo particles aren't contributing to cluster
> energy.  In other words, it seems as if the truncation is happening
> before detector simulation.
>
> Let me give some other indication of problems:
>
> * In a list of particles from (2), I found a charmed meson that was
>   INTERMEDIATE, but it had no daughters.
>
> * The number of particles in (1) and (2) are often similar (within a
>   couple).  But not infrequently, (2) has many more particles than (1).
>
> Has anybody noticed such a problem?  Can anybody shed some light to this
> problem?
>
> Gene

--
Richard Dubois
SLD, Stanford Linear Accelerator Center
[log in to unmask]
http://www.slac.stanford.edu/~richard/
650-926-3824
650-926-2923 (FAX)