[Rivet] W finder and PartonJet question

[Andy Buckley]

On 15/05/14 13:16, kristin Lohwasser wrote:
> Hi Andy,
> 
> thanks for the quick answer. Let me iterate a bit.
> 
> 1) hm, but then the questions is: What is with analyses, that should be
> implemented in Rivet, that actually are Born level (there are at least
> some I believe)

I believe (in ATLAS) that the publication policy is *always* to publish
a dressed lepton version without Born corrections, and *optionally* to
also do MC-specific Born corrections and publish those as an extra if
there is theory demand for them. In Rivet we always implement the
dressed version and not the Born, for the reasons I gave before...
primarily that there is no such such thing as a cross-generator
definition of a Born W or Z (nor is there even such a thing physically:
it's a correction back to a particular truncation of the perturbative
expansion!)

> and for those, as well as for the dijet cross section
> measurements, you'd never be able to make the data/comparisons plots
> using rivet, no? (this is a general question)

Correct, but as above we don't need to because more physical, less
"corrected" options should always be available. What is the issue with
dijet cross-section?

> 3) ok, this is not so urgent, and for my application it can be achieve
> in a different way.

Phew :-)

> 2) ok, thanks this explains it. I started from the non-weighted events,
> since I have a truth (unfolding) code, where I cross check against.

Glad to help!

Andy


> On 05/15/2014 12:55 PM, Andy Buckley wrote:
>> Hi Kristin,
>>
>> There is no mechanism in Rivet to find Born-level W/Z or parton jets --
>> in fact the whole design of Rivet is explicitly to discourage such
>> approaches because they are not portable between generators, and their
>> exact physical meaning is also rather vague.
>>
>> 1) If you want to construct these, then turn off QED FSR and any other
>> EW corrections in your MC: the bare leptons will then be "Born".
>>
>> 3) For parton jets you would need to turn off QCD FSR, hadronization,
>> and decays in the generator.
>>
>> At this point you pretty much just have parton level events of the sort
>> that would appear in an LHE file... and to my mind you might as well
>> just use an LHE parser and analyse the parton events directly rather
>> than fight against a system which is explicitly designed to encourage
>> more "final-state type" analyses!
>>
>> 2) When you say that you reapply the MET cut, do you mean the cut on
>> nu.pT()? This isn't quite the same as MET: the MET cut that we apply
>> also includes any other neutrinos in the event (including those from
>> hadron and tau decays) as well as ET imbalance due to acceptance
>> limitations. This is something where we might add more options in future
>> if there is demand, e.g. in response to the ATLAS (& wider) truth
>> definition workshops. Would this difference explain the change in
>> numbers that you are seeing? (By the way, it's a good idea to not just
>> use raw int cut-passed counters but to add the event weights instead --
>> that way you still get a physical answer if you use weighted generators.)
>>
>> Andy
>>
>>
>> On 15/05/14 11:30, kristin Lohwasser wrote:
>>> Dear all,
>>>
>>> I have a (couple of) question(s) concerning the WFinder implementation
>>> in Rivet and how to get Parton Jets in Rivet (is this possible at all?).
>>>
>>> Essentially I want to get BORN level electrons, for this I am using the
>>> WFinder (since as I understand, this is the recommended way to
>>> reconstruct W's in Rivet -- please correct me, if I am wrong)
>>>
>>> 1) CLUSTERING DEFINITION
>>> --> For Born level I am not sure, what setting for the Photon clustering
>>> to use, essentially for the below settings, the WFinder::NOCLUSTER and
>>> WFinder::CLUSTERNODECAY yield the same results (but these are different
>>> from my cross check numbers for Born level)
>>>        WFinder wfinder_born_el(fs, -2.5, 2.5, 20*GeV, PID::ELECTRON,
>>> 0*GeV, 8000*GeV, 25*GeV, 0.2, WFinder::NOCLUSTER    );
>>> so I am not sure, whether any of the ENUMS will actually selec W-Born
>>> level.
>>>
>>> 2) Difference for MET
>>> I understand, you set the MET cut in the WFinder Declaration (25*GeV in
>>> the above case).
>>> However, I apply a MT cut afterwards as well and there I also doublely
>>> applied the MET cut -- however when I *REMOVED* it, the number of
>>> selected events *CHANGED*, which should not happen (code follows now):
>>>
>>>
>>>        if (!wfinder_born_mu.particles().empty()) {
>>>          const FourMomentum& nu =
>>> wfinder_born_el.constituentNeutrinos()[0].momentum();
>>>          if ( calcMT(nu,
>>> wfinder_born_el.constituentLeptons()[0].momentum())  > 40*GeV &&
>>> nu.pT()>25*GeV) {
>>>            keepevent=true;
>>> }}
>>>
>>>        if (!keepevent) {
>>>          MSG_DEBUG("event does not pass mT and MET cuts");
>>>          vetoEvent;
>>>        }
>>> COUNT_PASS_W++;
>>>
>>> 3) How can want one get Parton Jets in Rivet? I tried to find some other
>>> analysis using this, e.g. ATLAS jets, but it seems, they are not
>>> implemented. It would be great, if you could suggest me something.
>>>
>>> Thanks & Best
>>> Kristin
>>>
>>>
>>>
>>>
>>>
>>
> 
> 


-- 
Dr Andy Buckley, Royal Society University Research Fellow
Particle Physics Expt Group, University of Glasgow / PH Dept, CERN