Possible incorrect reconstruction of photon energy.

[Dmitriy]

Dear experts,

I'm generating pp->Zj,Z->ee events with MG5 & pythia8, and then processing them with Delphes3.4.2 using the ATLAS card. Particle level electrons can be reconstructed there in the following ways:
1) As an electron, when event contains EFlowTrack and no EFlowPhoton;
2) As a photon, when event contains EFlowPhoton and no EFlowTrack;
3) As a photon, when event contains both - EFlowTrack and EFlowPhoton.

And in the case 3) the energy of reconstructed photon equals to the energy of EFlowPhoton. These photons satisfy isolation criteria and it is checked that there are no electrons and jets within ∆R < 0.1 of this photon, that is, the energy of EFlowTrack is lost.

Example of third case reconstruction (4-vectors (E, PT, Eta, Phi)):

Reco:(166.294, 145.276, 0.531624, 1.78427)
Track:(105.57, 92.1709, 0.53289, 1.79414)
EFlowTrack:(105.57, 92.1709, 0.53289, 1.79414)
Tower:(271.864, 237.504, 0.531624, 1.78427)
EFlowPhoton:(166.294, 145.276, 0.531624, 1.78427)
Particle:(269.196, 235.028, 0.53289, 1.79414), -11
Parton:(264.189, 229.483, 0.543269, 1.82591)

Reco - reconstructed photon, particle - electron on particle level which corresponds to Reco, parton - electron on generator level which corresponds to Reco. 4-vectors of all these particles lie within ∆R < 0.1 of Reco.

Could you please answer whether this kind of problem is due to some bug with reconstruction or maybe have I misunderstood something?

P.S. The fractions of events of 2-nd and 3-rd kind are approximately 3:2.

Best regards,
Dmitriy