What sample of data should this be aiming for? - B- sys limit B0 sys limit ratio : stat=sys 4% and diff is 5%. ouch => seems not to bother. Or need to work on all sys. [ discussed -- we see that the charm is the key ] What is the current (lp2005 or more) set of numbers. [ done -- see web pages] Optimization studies on cuts? Correlation matrix on fit? Fit prob determination? Check run 1 analysis and paper. [ being done ] Use of L00? Goal of 7458 was to repeat run 1. similarly, qq used for match to run 1, isgw in evtgen why not fit alpha and use its error for sys. how good is the fit? (yes, ok, answered in issue2 ] How do you know you have the right MC from the point of view of getting the correct blend of missing/dead svx? How much sys is there in that? What if you used all one run? Nb. incomplete MC issue 2 on post blessing raises the point to me: if the mixture matters, what is this telling us? issue3: not clear how lifetime shifts with s. ====== Bs tbs sys=stat ratio sys dom. 360pb-1 ========================= Can you assign the pion mass to the same charge as thelepton, instead of the kaon mass for any bkg estimate? Can you use the TTT to get charm how many events are there in prl prd now? Evt Frac mu- D0 21971 .425 mu- D*+ 2947 .907 e- D0 8539 .471 e- D*+ 1219 .908 mu- Ds+ 1591 .503 e- Ds+ 706 .503 last prd: Evt Frac l- D0 5198 .526 l- D*+(kpi) 935 .086 l- D*+(kpipipi-) 1166 .183 l- D*+(kpi0pi) 2858 .366 l- Ds+(phipi) 220 l- Ds+(k*0k-) 125 l- Ds+(k0sk-) 33 l- Ds+(phimunu) 205 Can you account for the combinatoric background in MC and assume the rest is charm -- does that give a better charm fraction estimate? The 0+ prd says you use wrong sign to get the estimate of b,c bkg. presumably this was wrong since you use it since there is no signal but there was no phase space, so the test was invalid. Mention this? What is the treatment of the PV? What is contamination of charm in sidebands -- as it is not combinatorical? What is combintaorical contamination in charm (prompt-enhanced) sample? why 260 on b+/0 Why are chi**2 cuts on muons different for Bs and B0+? Why not COT hit requirement for e,mu in the Bs? what about L00? does s get messed up by material so that L00 does not help? Why not svx/cot hit req in bs, but there is in b+0? NO optimizaiton mentioned for B+0, while given for Bs. Also no cone in either for the search region around the lepton? Track categories and dependencies thereon caused problems in the Bc. What is the set of tracks you select here? (you mention for the pi_s but not others) PRD+0 has P_b-P_v and you don't mention primary vertex. why? You could get better sys with more stat -- your error is stat lomited. why publish on this set, and not 1fb-1? plan to publish again? when? does publishing now exclude future? Do we follow notation for previous prd? use Monte Carlo, although it is mostly avoided in the PRD+0 Parametric vs full is complex to describe. What is going on? NO pt tune in PRD+0== why not? was around ? now an error? Why the more complex comb bkg param over what was done before: motivation? is what we had before ok? fit prob on e- is .001 Why do we now need special control samples for ct*? In prd+0 and Prd-s we did not. Loosening lepton id by dropping offline cuts seems uncontrolled -- in fact does it depend on run? Was level 3 etc always the same? Why not use a direct cut on track relative to PV? (as opposed to plenty) -- How does such a selection change your estimate of s? What is charm contamination in the prompt? does this affect s? ie: fake lepton plus: D from B+ D from charm D from Bother Combinatorics Do they all have the right s? We see for example that the charm has a very different s. Can you actually understand what the origin might be? number of tracks, momentum, material? It seems we are poking blindly in the dark... What other methods have been considered/rejected by B group? ( insert a fake track on a real D, etc) Why is resolution (s) so different for bs vs b+0? Why don't you use stat error on s when you do your systematics? WS now shows up with stats we have. did not show up in previous case. (are you unable to see the bump if you use same stat as run1?) There is NO bump in Bs so why worry? Can you list they ways bkg gets into b+0 and bs? are they the same? You fit templates for b+- and use exp for bs. why the differnce? We need to understand what the electron cuts are doing to the charm suppression. can we make the "same" cuts on the muons: that is, make it suppress charm in the same way? cutting out charm seems an advantage! Put another way, you can also select events within the isolation cone in muons that would have been rejected by electrons, enhancing the sample. ================== Bs paper A number of cuts on page 3 change relative to PRD -- why this? chi**2 x cmu 7 was 9 chi**2 z cmu 9 was 12 chi**2 x cmp 7 was 9 Iso/pt is 2.5 was 1.0 ces dx 1.4 was .5 ces dzsintheta 2 was 3 ces chi*2 is 15 was 10 pt pi is .9 was .8 m(l k+ k- pi+) is 3-5.4 was 3-5 De/Dx was not used here, was used before. Why is it not used? sigma ct(ds) .02 was .1 sidebands was 1884-1934 is 1900 to 1930 2004-2054 is 2015 to 2035 signal 1947.5-1982.5 now was 1944-1994 What notation do you wish to use for Lb? (page 7) On page 8 you talk about parameterization of Mc vs non parameterized. This could be an issue of time but why not be consistent? Seems hard to put sensibly into a publication. The parameterization of the pt spectrum was not in the prd but seems old hat. Why is this? In your definition of f**, you refer to a ration with decays of B->l nu Ds X. Doesn't the choice of BR's for X affect this? In fact, doesn't the differenesmentioned in the last part of section 4 imply a sensitivty to X? Why are you using two exponentials now for combinatorics whereas you used one in the previous PRD? Notation on Eq 1 is not very regiorus -- what is x? presumably ct*? pg 11 : You use loose leptons for resolution -- what goes into this s and how much does it depend on the precise lepton cuts? How do you know you sample the same tracks and hence resolution when you change composition in this way? pg 13 : seems you should use a constraint on fc that it cannot be negative. also, can you show us the templates? Table 4: electron fit probability bad -- cannot use this page 16: can you provide a table like table 7 in the b0+ for the b background? page 17: Can you cut on the isolation in the wedge for muon events to see if you get the same efficiency for charm events in electrons? And (as mentioned above) can you take the non -isolated muons and use them to test the charm background? page 19: after you fit, can you give a correlation matrix of the parameters? page 21: your alpha range does not reflect your answer in the blessing update page 24: Can you detail the correlated errors in B0+ and Bs? B0+- Note --------- Again, changes in cuts page 3 pt > 8 (was 9, then 8) chi**2 x cmu 6 was 9, is 7 in Bs chi**2 z cmu 5 was 12, is 9 in Bs chi**2 x cmp 6 was 9 is 7 in Bs In a combined paper you need to justify these differences. I can imagine run 1 vs run 2 is different, but bs vs b0+? Iso 2.5, none in PRD same as Bs now. There were no details in the PRD on the CES cuts (I guess it referred to a generic id) page 4: you dont mention optimization. did you use a cone to search for the tracks: you don't mention it. page 5 : sigma ct* B sigma ct D0 0.02 was 0.05 D*+ window is 1.835-1.895 was 1.83-1.90 We note the stand alone tracks are mentioned. This caused headaches in Bc to Jpsi pi when not clarified. What tracks are used and is there a standard B physics group blurb?! page 6 dm sideband is .15 -.2 was .15 to .17 notation page 7: was vh-vp now just vb. page 11 The resolution scale factor works out differently than Bs. How is this? Is it a result of fake composition and how do you know the fake composition? - or - Is there a systematic dependency on the loose cut defintion (again, just using no lepton cuts in reconstruction does not seem well-defined) The background changes with lepton pt -- how does this affect the analysis? It seems that if you give up some stat for systematic you can cut harder and get rid of the fc problem. 8.1: you can look at the lepton pt dependence here. Can you take a shot at seeing if you can compute the absolute background. YOu know the charm production cross section and BR. It seems we know the production modes, we know where tracks are and we can guess how well we reconstruct tracks near a jet (this coudl be gotten from MC as a first approx) Our chance of a track misid as a lepton is a number we can get from single mc. We also need to conside rconversions from pi0->gamma gamma and the gamma fakes a lepton. I imagine some of this is known. You parameterize templates but some parameters seem poorly determined. How good is the fit and cna you use a simpler one? I dont understand Fig 11 vs table 5. Seems the D* signals have very little prompt but you show zero. The D0 has a lot. Also it seems that the combinatoric form is getting correlated with the prompt charm and you end up with negative results. What if you remove the prompt charm in those fits. How good are they? The d0 fit probablities seem bad. Table 6 parameters for lambda+, f- and lambda- are poorly determined. Outstanding questions for Bs ----------------------------- Q1: k*K mode: Q2: need to understand this one Q5: what generator will you end up using for both? Q10: What is the status of the probablity of the electron fit? Q15: is it correct to ignore bbbar background or not?! Outstanding questions for B0+ ============================= Q7:We also refer to this, but what about using L00. The motivation to repeat run I is not important now. How would this affect charm background fitting? or the Systematic on s? Q12: again, for a publication this is not a good answer: we need to be consistent. Q14: We allude to this above. The fact that the standard cuts dont affect it is good -- so why bother with the poorly defined cuts? Q15: Try looking at the number of charm events that change between cuts so that you get a clear statistical meaning: that is measure delta Fc. Q20: This narrower alpha range was not implemented in the cdf note -- is it in the tables and results with systematics?