Explanation of systematics of CMS p+Pb high multiplicity di-hadron data at sqrt{s}_(rm NN) = 5.02 TeV

In a recent article (arXiv:1210.3890), we showed that high multiplicity di-hadron proton- proton (p+p) data from the CMS experiment are in excellent agreement with computations in the Color Glass Condensate (CGC) Effective Field Theory (EFT). This agreement of the theory with several hundred data points provides a non-trivial description of both nearside ("ridge") and away-side azimuthal collimations of long range rapidity correlations in p+p collisions. Our prediction in arXiv:1210.3890 for proton-lead (p+Pb) collisions is consistent with results from the recent CMS p+Pb run at $\sqrt{s}_{\rm NN} = 5.02$ TeV for the largest track multiplicity $N_{\rm track}\sim 40$ we considered. The CMS p+Pb data shows the following striking features: i) a strong dependence of the ridge yield on $N_{\rm track}$, with a significantly larger signal than in p+p for the same $N_{\rm track}$, ii) a stronger $p_T$ dependence than in p+p for large $N_{\rm track}$, and iii) a nearside collimation for large $N_{\rm track}$ comparable to the awayside for the lower $p_T = p_{T}^{\rm trig.}=p_{T}^{\rm assoc.}$ di-hadron windows. We show here that these systematic features of the CMS p+Pb di-hadron data are all described by the CGC (with parameters fixed by the p+p data) when we extend our prediction in arXiv:1210.3890 to rarer high multiplicity events. We also predict the azimuthally collimated yield for yet unpublished windows in the $p_{T}^{\rm trig.}$ and $p_{T}^{\rm assoc.}$ matrix.