moments update Click to edit Master subtitle style
moments update Click to edit Master subtitle style update on moments & sampled singles w. j. llope bulkcorr PWG, June 26, 2014 Independent code that reads 4 TH 2 Ds (net-X, tot-X, X, Xbar) vs rm. Ncorr at all 7 √s. NN… TH 2 Ds for X=p from xiaofeng luo (net-p paper), X=p, q, K from daniel mcdonald (q and K at QM 2012) produces: moments from data, & (N)BD and ``sampled singles” baselines http: //wjllope. rice. edu/fluct/protected/moments_20130417. pps now calculating sampled singles uncertainties from delta theorem now applying efficiency corrections to XFL’s net-p results working towards efficiency corrections (including tof match eff!) for dmac’s results _____________________________________ Outline overview of (N)BD and sampled singles baseline approaches XFL net-p, with and without efficiency correction dmac net-p, net-q, and net-K, then efficiency corrections to dmac net-q using nihar efficiencies comparison of dmac net-p and XFL net-p Bulkcorr PWG meeting, June 26, 2013 1
STAR Results (QM 2012) Click to edit Master subtitle style net-protons √s. NN <m. B>* 7. 7 421 11. 5 316 19. 6 206 27 156 No non-monotonicity seen 39 112 Significance of net-p dip? 62. 4 73 200 24 X. Luo for STAR, QM 2012 D. Mc. Donald for STAR, QM 2012 net-charge Bulkcorr PWG meeting, June 26, 2013 net-K 2 * Cleymans et al. PRC 73, 034905 (2006) moments update
moments update (N)BD Click to edit Master subtitle style T. . J. Tarnosky & G. Westfall, Oct. 2012 http: //arxiv. org/pdf/1210. 8102 v 1. pdf Functional form describes the (particle identified) multiplicity distributions ranging from NA 22 & UA 5 to PHENIX Inputs: mean (μ) & variance (σ2) Then, the values of Ck , Sσ , & Kσ2 are predicted. Bulkcorr PWG meeting, June 26, 2013 3
moments update “Sampled Singles” Click to edit Master subtitle style Sampled Singles is not “mixed events” The only input is the 2 D distributions of Npos vs. rm. Ncorr and Nneg vs. rm. Ncorr. where pos = p, K+, q+ and neg = pbar, K-, q 11. 5 19. 6 27 39 62. 4 200 Npbar 7. 7 same plots for K±, q± Np Filled at the same spot in analysis codes where the deviates are saved… …i. e. includes all the same track cuts, PID, and run&evt QA as the independent analysis… refmult. Ncorr Bulkcorr PWG meeting, June 26, 2013 4
moments update Sampled Singles method Click to edit Master subtitle style 7. 7 11. 5 19. 6 27 39 62. 4 200 Np Npbar In every slice of rm. Ncorr, sample one value of Npos and one value of Nneg, Nevt times. refmult. Ncorr In each sampled event at a given rm. Ncorr, one then has a value for Npos and Nneg Then form Nnet = Npos-Nneg and Ntot = Npos+Nneg Fill 2 D plots of Nnet and Ntot vs. refmult. Ncorr Then extract the moments (products) with the CBW corrections as usual… Destroys all intra-event correlations between Npos and Nneg, reproduces singles distributions, & has the same statistical certainty as the data, by construction… Bulkcorr PWG meeting, June 26, 2013 5
moments update Sampled Singles method: Oversampling Click to edit Master subtitle style Approach used up to a ~1 week ago sampled NEVT times for a given slice of refmult. Ncorr. Was observed however that the resulting CBW-corrected moments products had a noticeable variance (lowest root-s generally) with the choice of TRandom 3 seed… This is not physical but can be fixed easily with only expense being CPU time… Now sample each slice (pos or neg) NREPEATS*NEVT times and fill resulting Nnet and Ntot TH 1 Ds with a weight of 1/NREPEATS Same statistical uncertainties Same moments uncertainties (Δ thm. ) Stable sampled singles results will now show some plots for NREPEATS = 1, 2, 4, 8, 36, 49, 81 Bulkcorr PWG meeting, June 26, 2013 6
moments update Sampled Singles method: Oversampling Click to edit Master subtitle style RMS of Samp. Sing <Kσ2> values vs NREPEATS by root-s 0. 1 0. 001 RMS(<Kσ2>) decreases with increasing root-s, and with decreasing centrality at any NREPEATS Increasing NREPEATS decreases RMS as a power law. . Values stable for NREPEATS ≥ ~36 Bulkcorr PWG meeting, June 26, 2013 7
moments update Sampled Singles method: Oversampling Click to edit Master subtitle style Samp. Sing err(<Kσ2>) vs NREPEATS by root-s 0. 1 0. 001 0. 0001 Bulkcorr PWG meeting, June 26, 2013 8
moments update Sampled Singles method: Oversampling Click to edit Master subtitle style Delta theorem’s Kσ2 uncertainty for many runs with NREPEATS=1 Average value of Δ-thm uncertainty exactly reproduces that from data even for NREPEATS=1 But there is some variance. largest at lowest root-s. Bulkcorr PWG meeting, June 26, 2013 9
moments update Sampled Singles method: Oversampling Click to edit Master subtitle style Delta theorem’s Kσ2 uncertainty for many runs with NREPEATS=36 Average value of Δ-thm uncertainty still reproduces that from data Variance in Samp. Sing uncertainty now tiny. Bulkcorr PWG meeting, June 26, 2013 10
moments update Sampled Singles / Data comparisons Click to edit Master subtitle style Oversampling makes the sampled singles results much more stable. Code now takes an hour to run on my laptop, not ~3 minutes as before… Now turn to comparisons to Xiaofeng’s net-p results (net-p paper). Efficiency corrections performed to both data and sampled singles. Efficiencies are charge-independent and are taken from the analysis note. Note eff(39) = eff(62) = eff(200) was assumed Corrected cumulants depend on Ck(net), Ck(tot), Ck(pos), Ck(neg), Npos, & Nneg Bulkcorr PWG meeting, June 26, 2013 11
moments update Uncorrected net-p Sσ vs. centrality by root-s Click to edit Master subtitle style cyan bands are sampled singles uncertainties from delta theorem Bulkcorr PWG meeting, June 26, 2013 12
moments update Uncorrected net-p Kσ2 vs. centrality by root-s Click to edit Master subtitle style Bulkcorr PWG meeting, June 26, 2013 13
moments update Uncorrected net-p Sσ & Kσ2 vs. root-s for 0 -5% Click to edit Master subtitle style very good agreement to Xiaofeng’s uncorrected results in general, except 200 (some v. slight differences at lowest root-s) Sampled Singles reproduces the data values… (N)BD also does a good job for net-p… Note: TH 2 Ds from Xiaofeng at 200 Ge. V include 40 M events, plots in net-p paper have 240 M events. Bulkcorr PWG meeting, June 26, 2013 14
Now, perform the efficiency corrections… moments update Click to edit Master subtitle style Sσ vs centrality by root-s open: uncorrected solid: corrected 7. 7 11. 5 19. 6 27 39 Kσ2 vs centrality by root-s Bulkcorr PWG meeting, June 26, 2013 62. 4 200 15
Efficiency-corrected Sσ & Kσ2 with the Sampled Singles values moments update Click to edit Master subtitle style Sσ vs centrality by root-s measured(WJL) measured(XFL) sampled singles 7. 7 11. 5 19. 6 27 39 Kσ2 vs centrality by root-s Bulkcorr PWG meeting, June 26, 2013 62. 4 200 16
moments update Efficiency-corrected Sσ & Kσ2 with the Sampled Singles values Click to edit Master subtitle style Bulkcorr PWG meeting, June 26, 2013 17
moments update Click to edit Master subtitle style Now… switch over from XFL net-p data to dmac net-p, net-q, net-K data Bulkcorr PWG meeting, June 26, 2013 18
moments update Uncorrected dmac net-p Sσ vs. centrality by root-s Click to edit Master subtitle style Bulkcorr PWG meeting, June 26, 2013 19
moments update Uncorrected dmac net-p Kσ2 vs. centrality by root-s Click to edit Master subtitle style Bulkcorr PWG meeting, June 26, 2013 20
moments update Uncorrected dmac net-p Kσ2 vs. root-s for 0 -5% Click to edit Master subtitle style Bulkcorr PWG meeting, June 26, 2013 21
moments update Uncorrected dmac net-q Kσ2 vs. centrality by root-s Click to edit Master subtitle style (N)BD and Samp. Sing similar for 39 and below Samp. Sing blows up for 62. 4 and 200 and except for peripheral… Bulkcorr PWG meeting, June 26, 2013 22
moments update Sampled singles net-q central at highest root-s! Click to edit Master subtitle style Same events! Intra-event correlations are important… Bulkcorr PWG meeting, June 26, 2013 23
moments update Uncorrected dmac net-q Kσ2 vs. root-s for 0 -5% Click to edit Master subtitle style Intriguing that sampled singles ~reproduces the data for 39 Ge. V and below. What turns on between 39 and 62. 4? !? Bulkcorr PWG meeting, June 26, 2013 24
moments update Uncorrected dmac net-K Sσ vs. centrality by root-s Click to edit Master subtitle style Bulkcorr PWG meeting, June 26, 2013 25
moments update Uncorrected dmac net-K Kσ2 vs. centrality by root-s Click to edit Master subtitle style Bulkcorr PWG meeting, June 26, 2013 26
moments update Uncorrected dmac net-K Kσ2 vs. root-s for 0 -5% Click to edit Master subtitle style Bulkcorr PWG meeting, June 26, 2013 27
Efficiency corrections to dmac net-q (using nihar’s chg-independent eff. ’s) moments update Click to edit Master subtitle style Sσ vs centrality by root-s unlike net-p (slide 15), corrected net-q values are larger than the uncorrected values open: uncorrected solid: corrected 7. 7 11. 5 19. 6 27 39 Kσ2 vs centrality by root-s Bulkcorr PWG meeting, June 26, 2013 62. 4 200 28
moments update Summary Click to edit Master subtitle style Completely independent analysis code for data, sampled singles, and (N)BD. Now including sampled singles uncertainties (Δ-thm) and some efficiency corrections. Oversampling makes the sampled singles results stable for different TRandom 3 seeds. Sampled singles approaches and (N)BD compared to XFL’s net-p paper data. Sampled singles accurately reproduces the moments values from the data. ~no (impact from) intra-event correlations (on these observables) for net-p. Sampled singles curves based on net-p paper data provided to Xiaofeng. net-p paper GPC is presently discussing whether to show these in the net-p paper. Same comparisons done for dmac’s net-p, net-K, and net-q results. ~no (impact from) intra-event correlations (on these observables) for net-K. Strong (impact from) intra-event correlations (on these observables) for net-q and √s. NN≥ 62. 4. Plans I now have all embedding data for π±, K±, p and pbar at all BES root-s values… Presently calculating the π± and K± tofmatch efficiencies (p, pbar already in hand) Will allow me to efficiency-correct dmac net-p, net-q, & net-K data… Bulkcorr PWG meeting, June 26, 2013 29
moments update Uncorrected net-p Sσ vs. centrality by root-s, Comparison Click to edit Master subtitle style Samp. Sing not oversampled Bulkcorr PWG meeting, June 26, 2013 30
moments update Uncorrected net-p Kσ2 vs. centrality by root-s, Comparison Click to edit Master subtitle style Samp. Sing not oversampled Bulkcorr PWG meeting, June 26, 2013 31
moments update Uncorrected net-p Sσ & Kσ2 vs. root-s, 0 -5%, Comparison Click to edit Master subtitle style Despite widely different acceptances (~50% in <net-p> at 7. 7) the present d. E/dx-only analysis and the RICE “requiring TOF” give the same answer for 0 -5% collisions. Ks 2 Ss this is kind-of amazing to me. root-s Bulkcorr PWG meeting, June 26, 2013 32
moments update Click to edit Master subtitle style BACKUP SLIDES Bulkcorr PWG meeting, June 26, 2013 33
moments update Net-K Click to edit Master subtitle style sampled singles uncertainties are not drawn, but are similar to those on the data points by construction… Bulkcorr PWG meeting, June 26, 2013 34
moments update Net-Q Click to edit Master subtitle style Sampled singles values go beserk for ~central events at the highest √s. NN…. Bulkcorr PWG meeting, June 26, 2013 35
moments update Net-p Click to edit Master subtitle style sampled singles approach reproduces the net-p moments for all √s. NN & centrality… Bulkcorr PWG meeting, June 26, 2013 36
moments update Net-p vs. √s. NN, 0 -10% central Click to edit Master subtitle style Bulkcorr PWG meeting, June 26, 2013 37
moments update Present net-p paper values w/ and w/out efficiency correction. Click to edit Master subtitle style Bulkcorr PWG meeting, June 26, 2013 38
- Slides: 38