[Rivet] useInvisibles

[Frank Siegert]

Hi Andy,

I agree that a proper refactoring of this is for the next release and 
actually only wanted to check that this is done correctly in the two Z+b 
analyses we have (CDF_2006_S6653332 and CDF_2008_S8095620), where 
according to Gavin this could make a difference.

For these two CDF analyses though it's not quite clear to me whether 
neutrinos are corrected for or not. I tend to guess that they are not 
from the following text in http://arxiv.org/pdf/arXiv:0812.4458 p12

   "The jets are measured in the calorimeter and corrected to the hadron 
level [35], i.e., they are corrected for the CDF calorimeter response 
and multiple pp interactions. Note that the jets are not corrected for 
the underlying event (underlying event correction) or any changes in the 
energy contained within the jet cone due to fragmentation and any energy 
loss due to out-of-cone parton radiation (hadronization correction)."

In [35] there is no reference to neutrinos as far as I can see, and the 
paragraph overall sounds more like no additional corrections are 
applied. But one can't really know for sure.

These two analyses are quite a mess anyway right now, trying to match 
b-quarks to jets in a certain dR. I hope that's not the correct particle 
level definition of that analysis, and have already rewritten them to 
simply use jet.containsBottom() instead. But I don't have the plots 
right now to show that this doesn't make a difference, so I tend not to 
put this in before the release, and leave both of them as they are.

Frank

On 21/07/11 17:55, Andy Buckley wrote:
> Hi Frank,
>
> Yes, that one is still open. I had been thinking in terms of the task
> being to refactor the FastJets interface to make the required common
> definitions of jet particle content easier, which is a low priority
> thing that could wait for the next next release, but you're right that
> we should fix the currently wrong definitions now.
>
> Can you do this? Feel free to modify the FastJets interface to know
> about muon exclusion, neutrino (but not LSP) inclusion, etc. if the
> alternative would be nasty contortions with IdentifiedFinalStates or
> similar. It's easy to update the FastJets interface in Rivet 2.0 if all
> that needs to be changed is one function call, less easy if a chain of
> projections needs to be refactored!
>
> Andy
>
>
> On 21/07/11 08:43, Frank Siegert wrote:
>> Hi all,
>>
>> since we were talking about a new release these days, I thought I'd
>> mention that this one is still open as far as I know? I have it on my
>> TODO list, but haven't managed to fix it yet. Do you think it's
>> worth/feasible doing so for the imminent release, at least for the heavy
>> flavour analyses where it becomes relevant (I'd probably need a day or
>> two to implement and test)?
>>
>> Frank
>>
>> On 11/07/11 09:27, Jonathan Butterworth wrote:
>>> Just to say that all ATLAS analyses except the published W+jets paper
>>> use neutrinos and muons.
>>>
>>> The ATLAS default behaviour is to use all stable particles in the jet
>>> finder EXCEPT
>>> - prompt leptons
>>> - photons with a cone of 0.1 around prompt charged leptons.
>>>
>>> (These photons are generally summed with the lepton four vector
>>> instead).
>>>
>>> The upcoming (and currently preliminary) W+jets results also use this
>>> default behaviour.
>>>
>>> For technical reasons, this is not strictly implemented like this in all
>>> published ATLAS analyses, since prompt leptons and the photons are also
>>> input to the algorithm. But there is always a lepton/jet isolation which
>>> means that the difference is irrelevant given the local behaviour of
>>> anti-kT.
>>>
>>> Cheers,
>>> Jon
>>>
>>>
>>> On 08/07/2011 17:58, Andy Buckley wrote:
>>>> I think we need to revisit the jet config interface for the next
>>>> release
>>>> -- with all the analyses that I know are in preparation and this
>>>> mismatch becoming more obvious by the minute we're clearly going to
>>>> want
>>>> a 1.6.0. We can look into this when I'm down in Durham in a week or
>>>> so...
>>>>
>>>> Andy
>>>>
>>>>
>>>> On 08/07/11 17:12, David Grellscheid wrote:
>>>>> Hi everyone,
>>>>>
>>>>> It seems that at the moment, neither of the two options does what
>>>>> ATLAS
>>>>> seems to want. From the code comments:
>>>>>
>>>>>> The default behaviour is that jets are only constructed from visible
>>>>>> (i.e. charged under an SM gauge group) particles. Some jet studies,
>>>>>> including those from ATLAS, use a definition in which neutrinos from
>>>>>> hadron
>>>>>> decays are included (via MC correction) in the experimental jet
>>>>>> definition.
>>>>>
>>>>> If useInvisibles is true, Rivet not only includes the neutrinos from
>>>>> hadron decays, but all invisibles. W+jets analyses should be careful
>>>>> there!
>>>>>
>>>>> See you,
>>>>>
>>>>> David
>>>>>
>>>>>
>>>>> PS: I still do not understand why one would want to correct for an
>>>>> undetectable particle, but we've had that discussion before.
>>>>> _______________________________________________
>>>>> Rivet mailing list
>>>>> Rivet at projects.hepforge.org
>>>>> http://www.hepforge.org/lists/listinfo/rivet
>>>>>
>>>>
>>>>
>>>
>> _______________________________________________
>> Rivet mailing list
>> Rivet at projects.hepforge.org
>> http://www.hepforge.org/lists/listinfo/rivet
>>
>
>