How to understand propagators from event record

Hello Zach,

I don't fully understand your question. There are no intermediate propagator in the event you give as an example (such propagators have status code 2, and here all particles have status code -1 (incoming) or 1 (outgoing). The mother info in the event ("1 2") just tells that all outgoing particles come from the incoming particles. Finally, the color information 501/502/503 describes how the colors are connected in the event - this info is used by Pythia when it stretches color strings after parton shower (you see this in the part of the Pythia event record that with status codes between 10 and 19).

There is no such thing as a given diagram giving a given event, since a single diagram is not (in general) a gauge invariant quantity. You can get some information about reasonable clustering histories for the event from the color connection information, but remember that this is only valid at leading Nc (while the event distribution and cross section are of course valid at all orders in Nc).

Hope this helps, all the best,
Johan

Hi Zach,

> Is there a nice way to understand what diagram was being used to produce this matrix element? Or is the information gone once we are out of MadGraph? (Perhaps with the LHE file we can get the process number and try to line them up, but that isn't terribly stable).

In fact this information is not physical either due to gauge invariance or due to interference between different diagram.
So the rule that we follow in MG is include in the lhe file only s-channel propagator if they are "close" to the on shell mass.
(The definition of close follows the BW_cut parameter in the run_card). Indeed in such case the interference are negligible.

But behind this particular case, MG is not able to link an event to a particular diagram.

If you want to know the relative importance of the diagrams in the process, this is possible.
You need to look at the relative contribution of the different channel of integration. (the details of each channel are given in the html output)
For more information about the meaning of the channel of integration (and the link with Feynman diagram) please read the following topic:
https://answers.launchpad.net/madgraph5/+question/187387

Cheers,

Olivier

On Jul 13, 2012, at 3:50 PM, Zachary Marshall wrote:

> New question #203058 on MadGraph5:
> https://answers.launchpad.net/madgraph5/+question/203058
>
> Hi,
>
> I'm looking at, for example, ttj production. Some events are u g > t t~ u :
>
>
> I IST ID Mothers Colours p_x p_y p_z E m
> 1 -1 21 0 0 502 503 0.000 0.000 184.553 184.553 0.000
> 2 -1 2 0 0 501 0 0.000 0.000 -942.435 942.435 0.000
> 3 1 6 1 2 501 0 -14.277 39.092 -128.779 219.647 173.000
> 4 1 -6 1 2 0 503 1.914 -70.052 75.939 201.511 173.000
> 5 1 2 1 2 502 0 12.363 30.960 -705.042 705.830 0.000
>
> which in Pythia becomes:
>
> 3 !g! 21 21 1 0 0 2.58899 0.25353 184.54439 184.56273 0.00000
> 4 !u! 21 2 2 0 0 -0.69225 1.35703 -942.42958 942.43081 0.00000
> 5 !g! 21 21 3 0 0 2.58899 0.25353 184.54439 184.56273 0.00000
> 6 !u! 21 2 4 0 0 -0.69225 1.35703 -942.42958 942.43081 0.00000
> 7 !t! 21 6 0 0 0 -13.76723 39.40545 -128.67320 219.60901 173.00000
> 8 !tbar! 21 -6 0 0 0 3.81355 -69.77149 75.91630 201.43209 173.00000
> 9 !u! 21 2 0 0 0 11.85042 31.97660 -705.12828 705.95243 0.00000
>
> That is, the MadGraph record shows these things connected all together, and the Pythia record shows them as not connected to anything. From ExRootAnalysis, we get the ttbar connected (apparently) to the incoming u, and the outgoing u apparently connected to the incoming g - as though the propagators were just completely removed.
>
> Is there a nice way to understand what diagram was being used to produce this matrix element? Or is the information gone once we are out of MadGraph? (Perhaps with the LHE file we can get the process number and try to line them up, but that isn't terribly stable).
>
> Thanks,
> Zach
>
> --
> You received this question notification because you are a member of
> MadTeam, which is an answer contact for MadGraph5.

Hi Olivier and Johan,

That's just the answer I was expecting, I suppose... Thanks! I should really stop trying to cheat with the parton record...

Best,
Zach