drach09's blog

As a follow-up to the previous comparison of standalone PYTHIA to NLO pQCD, I repeat the study for the existing embedding sample. The statistics for low partonic p_T will not be as plentiful, but the statistics for high partonic p_T will allow a better examination of the high-p_T jets. This may also act as a systematic check of the low-p_T behavior.

For this study, I normalize each partonic p_T bin by mulitplying by the corresponding partonic cross section and dividing by the corresponding number of generated PYTHIA events. In addition, I scale by the "fudge factor," derived to ensure a smooth partonic p_T spectrum. The fudge factors were calculated relative to the 2-3 GeV/c partonic p_T bin, and thus the fudge factor for the 2-3 GeV/c bin is 1. One may, then, expect there to be a constant offset in the comparison of NLO to PYTHIA in the event that the PYSTAT cross section for the 2-3 GeV/c bin is incorrect.

Figure 1

In Fig. 1 I show the comparisons of PYTHIA parton-jet and particle-jet cross sections to that of NLO pQCD jets. Compared to the previous study the high-p_T statistics are vastly improved, while the low-p_T statistics appear to be adequate. As in the previous study, at low p_T, the PYTHIA cross sections are enhanced relative to the NLO cross section. With the improved high-p_T statistics, it is clear there is indeed a constant offset for the whole distribution. Fitting this offset with a constant over the range 15 < Jet p_T < 40 GeV/c, I obtain a ratio of 1.642±0.0075 with χ²/ν = 35.06/25. This may indicate that the PYSTAT cross section for the 2-3 GeV/c bin is not quite right and that the fudge factors calculated relative to this inherit an extra scaling by a factor of ~1.6.

Figure 2

In Fig. 2 I show the result of scaling the PYTHIA cross sections by 1/1.642 to correct for the constant offset. After the correction, the low-p_T enhancement is still around 50% from 5-6 GeV/c, 40% for 7 GeV/c, and 20% for 8 GeV/c in parton jets. Thus, it still appears likely that the low-pT enhancement plays a substantive role in the low-p_T discrepancy between data and embedding.

Removal of Low Partonic p_T Bins

As a check, I also evaluated the jet cross sections after removing the contributions of the 2-3 GeV/c bin and also if I further remove the 3-4 GeV/c bin (Fig. 3). Removal of the 2-3 GeV/c partoinc p_T bin appears to improve the agreement of the lowest jet p_T bins, but the low p_T enhancement still remains to a large degree. Removal of the 3-4 GeV/c partonic p_T bin is clearly too much for the 5 GeV/c jet bin and quite possibly too much for the 6 GeV/c bin. A 15% enhancement still remains for a number of the low p_T bins. Clearly, one must be careful of throwing out partonic p_T bins, but it does not appear that removing a couple of low partonic p_T bins has a significant negative effect on the bins relevant to the jet-patch triggers. This does not help my particular analysis, but it may prove informative for other jet analyses which are concerned only with jet-patch triggers.