Skip to content

Tune BnoC HLT2 BdsTohhhh line cuts

Matthew David Monk requested to merge mmonk_bnoc_4h_tune into bnoc_run3_zewen

Reduce the rate of the BdsTohhhh lines, some of which are unacceptibly high currently -- see issue #571 (closed).

Summary of changes so far:

  • Switch to using different kaon/pion builders (also to avoid cuts on the ghostprob)
  • Make some "easy" cuts to reduce the range of the B-candidate mass to something more typical
  • Tighter pT cut on the sum of the children
  • Add an optional cut to the make_b2x builder to cut on the minimum of the final-state minipchi2 which the 4h lines then make use of
  • Relax slightly the very tight DIRA cut
  • Place an upper limit on the 2- and 3- body combinations to be below the charm threshold
    • Actually, the charm part was intentional to be used as a calibration sample. I found that putting it back in does not affect the rate that much so decided to leave it in (cc @jdalseno)
  • Tighten up the MVA cuts
    • I did not do a particularly fine-grain test for this but it seems to reduce the rate a little without much of a drop in efficiency

Rates

The rate of the 4pi line has not gone down (with the approx. 15000 events of minbias this runs on anyway) but the efficiency of the line is a bit better (and I suspect with more events the rate has probably gone down a little)

Line Rate in master (kHz) New rate (kHz) New rate more events (kHz)
B \rightarrow \pi^+\pi^-\pi^+\pi^- 0.267 0.267 0.24473 \pm 0.0292
B \rightarrow K^+\pi^-\pi^+\pi^- 1.6 0.267 0.18529 \pm 0.0254
B \rightarrow K^+K^-\pi^+\pi^- 0.601 0.200 0.26221 \pm 0.0302
B \rightarrow K^+K^-K^+\pi^- 0.534 0.200 0.12236 \pm 0.0206
B \rightarrow K^+K^-K^+K^- 0.667 0.067 0.00699 \pm 0.0049

Overlap

I tried also running the overlap checker on the 4h and VV lines so that we can try to understand how strong the overlap is between them. It seems the overlap is not that strong between exclusive vs inclusive (e.g. B \rightarrow \pi^+\pi^-\pi^+\pi^- vs. B \rightarrow (\rho^0 \rightarrow \pi^+\pi^-) (\rho^0 \rightarrow \pi^+\pi^-)) but more so between lines with one different final state particle, e.g. B \rightarrow K^+K^-K^+\pi^- vs. B \rightarrow K^+K^-\pi^+\pi^-

cond_probs

jaccards

Edited by Zewen Chen

Merge request reports