[G12] G12 Kinematic Fitter for pions

Sun Dec 20 09:53:14 EST 2015

Lei,

Michael’s pull distributions for the two-pion channel are clearly not properly tuned. This is most likely because of the covariance matrix, any cuts (PID, timing, etc.) have an impact on the pull mean values but a smaller effect on the pull widths and the CL distribution. His widths are too small, which implies that the errors are too big and explains why the CL distribution is rising. I don’t really understand why this issue comes up now.

Since Zulkaida is currently not around, any kinematic fitter discussion in a meeting tomorrow wouldn’t make sense from my side. I suggest we postpone it until early next year.

Best wishes,

	Volker

> On Dec 19, 2015, at 9:31 PM, Lei Guo <lguo at jlab.org> wrote:
> 
> I have seen something similar before... People do different cuts to preselected events, such as pid, timing, etc. when u do a more stringent precuts, the cl looks better. Not sure if this is the cause for the current issue
> 
> Regards.
> 
> Sent from my iPhone
> 
> On Dec 19, 2015, at 1:40 AM, Michael C. Kunkel <mkunkel at jlab.org <mailto:mkunkel at jlab.org>> wrote:
> 
>> Greetings,
>> 
>> My issue is with the difference of the pulls under the same conditions. This should not be. This can only be true if somehow the data sets are different. I do not agree with you that we an live with somewhat different pull and CL distributions and I believe having a meeting on this would be beneficial.
>> 
>> I understand we have been through this several time, in fact it is always FSU that brought this up and it was discussed. Now I am bringing it up and wish to discuss it. I would hope for the same professional courtesy as FSU received on the subject. 
>> BR
>> MK
>> ----------------------------------------
>> Michael C. Kunkel, PhD
>> Forschungszentrum Jülich
>> Nuclear Physics Institute and Juelich Center for Hadron Physics
>> Experimental Hadron Structure (IKP-1)
>> www.fz-juelich.de/ikp <http://www.fz-juelich.de/ikp>
>> On 12/18/15 6:16 PM, Volker Crede wrote:
>>> Michael,
>>> 
>>> We have been trough this several times and I am afraid we will never agree on these fine-tuning parameters. On the other hand, as long as the actual code is the same, I believe we can live with somewhat different pull and CL distributions … provided they look okay.
>>> 
>>> - Volker
>>> 
>>> 
>>> On Dec 18, 2015, at 12:11 PM, "Michael C. Kunkel" < <mailto:mkunkel at jlab.org>mkunkel at jlab.org <mailto:mkunkel at jlab.org>> wrote:
>>> 
>>>> This is obvious, and hence why Lei asked for everyone to show their pulls.
>>>> BR
>>>> MK
>>>> ----------------------------------------
>>>> Michael C. Kunkel, PhD
>>>> Forschungszentrum Jülich
>>>> Nuclear Physics Institute and Juelich Center for Hadron Physics
>>>> Experimental Hadron Structure (IKP-1)
>>>> www.fz-juelich.de/ikp <http://www.fz-juelich.de/ikp>
>>>> On 12/18/15 6:09 PM, Volker Crede wrote:
>>>>> Apparently, there is more to it … Our distributions look better.
>>>>> 
>>>>> - Volker
>>>>> 
>>>>> 
>>>>> On Dec 18, 2015, at 12:06 PM, Michael C. Kunkel <mkunkel at jlab.org <mailto:mkunkel at jlab.org>> wrote:
>>>>> 
>>>>>> Incorrect. If you run a diff on the codes from FSU and the one shown there is no difference in any file.
>>>>>> BR
>>>>>> MK
>>>>>> ----------------------------------------
>>>>>> Michael C. Kunkel, PhD
>>>>>> Forschungszentrum Jülich
>>>>>> Nuclear Physics Institute and Juelich Center for Hadron Physics
>>>>>> Experimental Hadron Structure (IKP-1)
>>>>>> www.fz-juelich.de/ikp <http://www.fz-juelich.de/ikp>
>>>>>> On 12/18/15 5:35 PM, Volker Crede wrote:
>>>>>>> Michael,
>>>>>>> 
>>>>>>> I don't think this is what FSU uses for the pions … Our pulls and CL distribution look slightly better.
>>>>>>> 
>>>>>>> - Volker
>>>>>>> 
>>>>>>> 
>>>>>>> On Dec 16, 2015, at 5:49 AM, Michael C. Kunkel <mkunkel at jlab.org <mailto:mkunkel at jlab.org>> wrote:
>>>>>>> 
>>>>>>>> Greetings, 
>>>>>>>> 
>>>>>>>> Below is a link that shows the pull probability and pull distributions look like for the tune channel of p pip pim. These parameters are what FSU, ODU, Bochum, and others use for pions. 
>>>>>>>> We had agreed that the pull distributions were adequate if the pull probability is adequate.  However as you can see from the plot, there is a small rise in the pull probability when it approaches 1. 
>>>>>>>> 
>>>>>>>> https://www.jlab.org/Hall-B/secure/g12/mkunkel/G12_MEETINGS/excl_ppipi_10.pdf <https://www.jlab.org/Hall-B/secure/g12/mkunkel/G12_MEETINGS/excl_ppipi_10.pdf>
>>>>>>>> What do you think about this?
>>>>>>>> -- 
>>>>>>>> BR
>>>>>>>> MK
>>>>>>>> ----------------------------------------
>>>>>>>> Michael C. Kunkel, PhD
>>>>>>>> Forschungszentrum Jülich
>>>>>>>> Nuclear Physics Institute and Juelich Center for Hadron Physics
>>>>>>>> Experimental Hadron Structure (IKP-1)
>>>>>>>> www.fz-juelich.de/ikp <http://www.fz-juelich.de/ikp>_______________________________________________
>>>>>>>> G12 mailing list
>>>>>>>> G12 at jlab.org <mailto:G12 at jlab.org>
>>>>>>>> https://mailman.jlab.org/mailman/listinfo/g12 <https://mailman.jlab.org/mailman/listinfo/g12>
>>>>>>> 
>>>>>>> 
>>>>>>> 
>>>>>>> 
>>>>>>> 
>>>>>> 
>>>>> 
>>>>> 
>>>> 
>>> 
>>> 
>>> 
>>> 
>>> 
>>> 
>> 
>> _______________________________________________
>> G12 mailing list
>> G12 at jlab.org <mailto:G12 at jlab.org>
>> https://mailman.jlab.org/mailman/listinfo/g12 <https://mailman.jlab.org/mailman/listinfo/g12>
> _______________________________________________
> G12 mailing list
> G12 at jlab.org
> https://mailman.jlab.org/mailman/listinfo/g12

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/g12/attachments/20151220/35a3a053/attachment-0001.html>