[Vincia] [Pythia8] Gluon splitting to quarks

[Peter Skands]

Hi Torbjorn,

Yes, the LHC D* spectra in jets are intriguing, and speculation was made 
concerning whether the discrepancy with Pythia could be due to missing 
g->c/b splittings. However, it is somewhat difficult to understand then, 
why that problem only shows up for the lowest jet pT bins (largest 
discrepancy in lowest bin at 25 GeV), while for higher-pT jets the 
problem appears to go away. Further adding to my confusion is the ALEPH 
D* spectrum, which also shows Pythia coming in too low at low D* x 
values. So at the moment I don't understand how this can all hang 
together, just that there are some significant discrepancies with 
several data sets and between individual codes.

Peter

On 17/01/14 12:05, Torbjorn Sjostrand wrote:
> Hello Peter and Steve,
>
> Thanks for input, it is most welcome. I have a master's student who
> should begin a 4 month project on Monday, and c/b(/t) was on my list
>  to study in the shower. From a completely different perspective,
> however, inspired by criticsm I have received that Pythia is producing
> too few c/b. For instance, in this workshop you mention, study slide 4
> in Michelangelo's presentation to see how Pythia undershoots
> c production by about a factor of 2. Of course one will need to
> understand what has been done, also also separate Pythia 6 vs. Pythia 8,
> which is why I look forward to the promised input from Steve.
> Peter, was this slide discussed in the same context of too much/little
> in showers? Or any other ideas you have?
>
> Anyway, we should stay in touch. However, so as not to spam too much,
> maybe we should restrict to those expressing an interest.
>
> Best, Torbjörn
>
> On 16/01/2014 13:20, Peter Skands wrote:
>> Hi Pythia and Vincia'ers
>>
>> In the context of a workshop on charm at LHC this week at CERN, I took a
>> look at how much g->ccbar Pythia actually produces at LEP, and compared
>> it with Vincia.
>>
>> Of course, the absolute rates are low, of order one in a thousand
>> events, but there seems to be a clear factor 2 difference between the
>> two, see the attached plot, which shows the spectrum of xCharm =
>> 2|p|/eCM on a log scale, for Z -> uu/dd events. The case for bottom is
>> included as well, and exhibits the same behaviour.
>>
>> I realize that g->QQ is quite an uncertain process from a shower
>> perspective, but a factor of 2 nonetheless seems disturbing, and makes
>> me wonder eg whether all factors of TR = 1/2 (or TRhat = 1, in Vincia's
>> case, with 1/2 coming from summing over either side of the gluon) have
>> been implemented consistently in the two codes. Have any of you looked
>> at similar issues? There is also a question about whether to use mQQ as
>> the renormalization scale (Vincia) or pT (Pythia), but I doubt that
>> could lead to this big a difference.
>>
>> The background is also that an experimentalist at the workshop raised
>> that Pythia seems to be overpopulating the g->cc component of W+charm
>> events, according to a study performed for CMS SMP 12 002. He said the
>> plots were not in the paper but that they could be seen on public plots
>> on the analysis web page but I was not able to find them on there :( I
>> did not take down his name.
>>
>> Cheers,
>> Peter
>>
>>
>> _______________________________________________
>> Vincia mailing list
>> Vincia at projects.hepforge.org
>> http://www.hepforge.org/lists/listinfo/vincia
>>
>
> _______________________________________________
> Vincia mailing list
> Vincia at projects.hepforge.org
> http://www.hepforge.org/lists/listinfo/vincia