E852 experiment
Analysis of Eta Pi0
system with the decay Eta -> Pi+ Pi- Pi0
********************************************
Draft 5
Date: 25 May 2006
From: Suh-Urk
Subject: MC t' dependence
I have one thought.
Our MC events are already generated with a t distribution.
Suppose they have been generated with EXP[-B*t].
Then we will need to compensate for it by multiplying
all the new t-dependent amplitudes I have proposed---
with a factor EXP[+B*t].
Valodya, your MC events have been generated with
a t-dependence ?
Let me know. Thanks.
From: Suh-Urk
Gary,
If the amplitudes contain a t dependence,
then the MC events cannot have any t dependence.
One should not double-count.
Maybe I have misunderstood your remarks??
From: Gary Adams
Not so. We usually generate MC with t dependence to match the
data. Without prior knowledge of the amplitudes I don't see any other way
to do it.
Gary
From: Gary Adams
Suh Urk:
You are right. I didn't explain my concern very well because I didn't
understand it myself. In Vladimir's note 7 he applies your constraint by
specifying the full t dependence of each wave. The exponential part is
given by to', which is not fitted. I agree that a full fit including to'
would require no t dependence in the MC. But what should we do if the data
do not exhibit this behavior? If t decouples from all the fit variables
then I think you are correct that removing the t dependence of the MC is
good enough.
I thought this wouldn't work, but perhaps it is easier than I was
thinking. If the t dependence of the MC is removed then we should be able
to compare the fit results with the data for any given to'. If they agree
in t, then there is no issue. One could expand the to' grid beyond the two
values that were tried to see if an optimum value can be found.
Regards,
Gary
From: Neal Cason
Hi all:
Don't forget, you have to have some t dependence in the Monte Carlo or
you need a huge amount of Monte Carlo events. Plain old phase space is so
non-exponential that almost none of the generated events are very likely in
practice.
Neal
Response
Our raw MC events are calculated with t' dependence and parameters B=6.47 (GeV/c)^{-2}.
I remind that in PWA we used a ratio acc/raw in each mass bin. The next.
fig
demonstrates the very flat t' dependence of ratio acc/raw in EtaPi0 MC events. So the PWA results is not sensitive to t' dependence in MC events, if we don't consider the seperate t' bins.
