On Fri, 9 Dec 2005, Heiko Lacker wrote:

> Hi,
> 
> if I'm not mistaken there is a bug (at least in a part) in SP8
> where not all B->D**lnu modes have been simulated. That could
> explain the difference you see in the lepton spectra between
> data and MC. 
> 
> This has been reported at the last Collab meeting. I don't know
> where though.

I found it:

http://www.slac.stanford.edu/BFROOT/www/Organization/CollabMtgs/2005/detSep05/Thur1a/jochen.pdf

Here's what it says (look at it yourself for more details):

Apparently the branching fractions for the B->Dpi l nu (ie,
non-resonant B->Xc l nu) modes were accidentally set to zero in DECAY.DEC.

They made a new release with the correct branching fractions, 
but by then they had already done some simulation production.

The result is that for Runs 2 and 3, the total semileptonic 
branching fraction is 4% too low.  This affects semileptonic studies.

It looks like Run2 and Run3 will be redone.  In the meantime, 
Jochen proposes (to the simulation people) to:

* Provide weighting function to scale up broad D** resonances 
to get correct semileptonic branching fraction

* For the few semileptonic analyses that are sensitive to the 
shape of the nonresonant contribution, produce a dedicated 
nonresonant MC sample.

But I am not sure if they ever did what he proposed.

Happy new year,

sheila