Chensheng:

Crystal Ball in signal extraction, better consistency with run 10 (both Gaussian and Crystal ball). Is this method better? Not necessary, but may indicate that this difference is due to signal extraction and not tracking or something in hardware. Needs more study n future

Please add Crystal Ball parameters to see if there are no large variations.

Rongrong
J/psi in Au+Au run 14:
Why nSigmaPion is used for PID? Not nSigmaMu in MuDst, mean is close to mean for muon (0.5 difference). This will require attention for the efficiency calculation because dE/dx is is reliable in the embedding (at least it was not in the past).
- scale factor: for Mix/LS - looks like bin counting is more stale, especially for low statistics bins
- lower stat. error from fit compared to bin counting - why? Are other errors on the fit (other then amplitude) included? Rongrong said they are since the parameter is an area in gausn
- suggestion - check if like-sign background (same event) gives consistent  signal with the even-mixing. Negative slope in the e+e- mass (while there should be cc-bar contribution) arises from different shapes of mixed event mass spectrum and real combinatorial background.

Upsilon: how the signal was optimized? - by scanning cuts and taking the best signal (with the best significance)
- what if you use J/psi cuts? - signal is ~50% lower, but there is still a signal
- mass plot on slide 23 - this is unlike mass distribution after the mixed event was subtracted. We need also plot with raw counts and raw like-sing bg and residual background.
- widths of peaks - need attention and an embedding to have good handle on the signal and purity of states

Takahito:
J/psi v2 with MTD:
VPDMBZDCNoVtx - table with recentering and flattening factors obtained from minimum bias sample (to not be biased by the MTD trigger)
- suggestion: focus on syst. error evaluation and leave invariant mass v2 method for the future (after QM'15)

MTD trigger efficiency:
- consistent results for production mid and low.
- slide 15: some indication of Vz dependence, try use wider bins and fit straight line
- some eta dependence seen already, need to estimate of syst. error
- need estimate an effect of the efficiency variation vs pT when extrapolating to the NoVrtx events
- purity-  extrapolation is difficult because no handle on the extrapolation. Embedding cannot help because we do not have timing there.
- purity estimation - try to use simulations or muons from J/psi to have better handle on the distribution and then purity.
- production mid and high have different purity - different comes from different timing window (?)
- efficiency study - if all the parameters are varies, than purity and efficiency will change in the same time. So there might not be any functional dependence of efficiency on purity
- will try to get more sample by using delta ToF
- maybe use pure muon sample from cosmic ray to have a template for muon signal? It might be difficult because the pT may not be well defined

Liang:
- used tracks from different eta halfs (exclusive bins) while run 4 used inclusive eta range -> use the same method
- check if pT cuts are the same for this study and run 4
- check v2 vs pT for a give centrality bin to check if efficiency vs pT makes any difference
- peak in the like-sign mass spectrum  - is this fluctuation or a signal? Likely fluctuation
- new v2{2} results have larger stat. error compared to previous measurement. Maybe due to different hadron v2?