Recent precision comparisons between data and energy loss/jet quenching calculations are making serious use of systematic and statistical uncertainties. Unfortunately, there is a significant difference between the STAR pi+- RAA data and the newest PHENIX pi0 RAA. Previous comparisons have been made by Jamie and Thomas:

http://www.star.bnl.gov/protected/highpt/dunlop/comparisons_star_worlddata/stardata_comparedwithworlddata.html

The same issue is illustrated here (STAR results have stat+sys error drawn):

where the h+- RAA results is also shown. Interestingly, at high pt, the old STAR h+- is closer to the PHENIX result.

There are two things pertainign to this comparisons that I would like to discuss. Firstly, the STAR result uses a fit to p+p as the baseline. This to some extent underestimates the uncertainties. A bin-by-bin division (with interpolation), gives

For completeness, this is the ratio between the p+p data and the fitted Levy distribution (note again the large uncertainties at high pt)

It would probably be interesting to add the new (preliminary) EMC-triggered pion spectra to this figure.

While the larger uncertainties from the data-to-data ratio reduce the disagreement wirh PHENIX, there is still a sizeable disagreement at lower pt (2-6 GeV). Moreover, it also seems that the p+p spectra from STAR and PHENIX are in rather good agreement, while the real difference is in the Au+Au data. This is visible in one of the plots from Jamie's page:

*Added 4 June 2010 *

This plot shows the ratio of STAR and PHENIX data for p+p and Au+Au separately. It looks like the largest disagreement is in the Au+Au data: