[Rivet] Rivet validation plots

[Jamie Tattersall]

Hi Chris,

Yeah, I thought this might be the case. I was kind of hoping this may be 
a way to go under the radar! Come the revolution....

Cheers,
Jamie


On 06/05/16 15:31, Christian Gütschow wrote:
> Hi Jamie,
>
> I understand your point completely, but unfortunately the foot 
> soldiers will get in trouble if we make these run cards public without 
> seeking approval first. Collaboration policies and all that -- I’m 
> afraid you're gonna have to redirect your complaint towards the ATLAS 
> management.
>
> For now, we have to rely on this being sufficiently described in the 
> paper. I understand this hasn’t always been done properly in past, 
> which is one of the reasons why this time (for Run 2, that is) we have 
> aimed to document the MC setup (and the employed tunes) as much as 
> possible in dedicated pub notes. I appreciate that this isn’t optimal 
> either, but I hope it will still improve the situation.
>
> Cheers,
> Chris
>
>
>> On 6 May 2016, at 15:08, Jamie Tattersall 
>> <tattersall at physik.rwth-aachen.de 
>> <mailto:tattersall at physik.rwth-aachen.de>> wrote:
>>
>> Hi Christian,
>>
>> Thanks for the quick reply.
>>
>> One point just to emphasise is that the results can change 
>> significantly depending on Monte Carlo tunes and param card choices. 
>> I want to emphasise this because just saying 'Pythia 6' is really not 
>> enough. I want to make this clear because I know from SUSY examples, 
>> different choices of for example the matching scale produce 
>> completely different results.
>>
>> This is why it would be so good to include the actual monte carlo 
>> cards as well.
>>
>> Cheers,
>> Jamie
>>
>>
>> On 06/05/16 14:58, Christian Gütschow wrote:
>>> Hi Jamie,
>>>
>>> just to comment:
>>>
>>>> My particular interest was actually in ATLAS_2013_I1190187 as we 
>>>> could never reproduce the p_T(lep lep met) distribution and I just 
>>>> saw that there was now a rivet implementation. We were not the only 
>>>> theory group that had this issue but we could never really get to 
>>>> the bottom of it (even after correspondance with ATLAS). 
>>>> Unfortunately (unless I've missed something) it seems like the file 
>>>> here only contains the data and not the validation.
>>>>
>>>
>>> Unfortunately this one was submitted already ~2 years ago, i.e. a 
>>> long time before we got accustomed to submitting validation plots 
>>> along with the routines.
>>>
>>> That said, if a routine has already been used to produce the curves 
>>> in the paper, then we usually don’t produce separate validation 
>>> plots and only submit the routine/reference data even today.
>>>
>>>
>>>> In general, we would love to see as much as possible! Histograms, 
>>>> with the data vs monte carlo is the obvious one. I've just had a 
>>>> random look through some of the analyses present and stumbled 
>>>> across ATLAS_2015_I1351916. I guess the plots here are just what we 
>>>> are looking for but particularly important for us is to also know 
>>>> exactly which Monte-Carlo was used (I have to admit that I haven't 
>>>> dug up the paper so perhaps this is made clear here). In any case, 
>>>> it would be very useful to have the information contained in the 
>>>> validation directory.
>>>
>>> That’s a good point, I shall make a mental note to check that this 
>>> information is provided somewhere in the legend. For this particular 
>>> case, we used a Pythia6 sample for the validation, which I think is 
>>> described in the paper, but I see your point: an informative legend 
>>> entry won’t hurt. :-)
>>>
>>> Cheers,
>>> Chris
>>>
>>>
>>>
>>>
>>>> On 05/05/16 21:06, Andy Buckley wrote:
>>>>> On 05/05/16 19:28, Jamie Tattersall wrote:
>>>>>> Hi Andy,
>>>>>>
>>>>>> This is potentially a very stupid question..... but where do I find
>>>>>> validation material (i.e plots) for the Rivet analyses? (I've been
>>>>>> clicking for a while now but not got anywhere).
>>>>>
>>>>> Hi Jamie,
>>>>>
>>>>> Good question! This is probably something that we should make a 
>>>>> systematic point of putting online. For the last year or so, we 
>>>>> have managed to get the experiments to supply validation plots 
>>>>> along with the code, and those plot sets are in the tarballs at
>>>>> https://www.hepforge.org/archive/rivet/contrib/
>>>>>
>>>>> For a lot of the SM ones, there are also fairly up-to-date 
>>>>> comparisons against modern generators at http://mcplots.cern.ch/
>>>>>
>>>>> But we don't yet have a repository of e.g. generator comparisons 
>>>>> to all plots on the Rivet website. It is something that we could 
>>>>> add, though... just putting our new nightly validation outputs on 
>>>>> the website for example.
>>>>>
>>>>> I guess that you are more interested in the BSM analyses, though. 
>>>>> What sort of validation would you like to see for those? As you 
>>>>> know, they haven't been a major emphasis so far, but that is 
>>>>> changing so any input is welcome!
>>>>>
>>>>> Andy
>>>>>
>>>>
>>>> _______________________________________________
>>>> Rivet mailing list
>>>> Rivet at projects.hepforge.org <mailto:Rivet at projects.hepforge.org>
>>>> https://www.hepforge.org/lists/listinfo/rivet
>>>
>>>
>>>
>>>
>>>  —
>>>
>>>  Dr. Christian Gütschow
>>>
>>>  TU Dresden
>>>  Institut für Kern- und Teilchenphysik
>>>  Zellescher Weg 19
>>>  01069 Dresden
>>>
>>>  > at CERN: 104-02-C02
>>>  > at IKTP: E17, ASB
>>>  > chris.g at cern.ch
>>>
>>>
>>>
>>>
>>>
>>
>
>
>
>
>  —
>
>  Dr. Christian Gütschow
>
>  TU Dresden
>  Institut für Kern- und Teilchenphysik
>  Zellescher Weg 19
>  01069 Dresden
>
>  > at CERN: 104-02-C02
>  > at IKTP: E17, ASB
>  > chris.g at cern.ch <mailto:chris.g at cern.ch>
>
>
>
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.hepforge.org/lists-archive/rivet/attachments/20160506/ba760fd4/attachment.html>