Handshakes for the limits

Values

Toy Models

under svn: https://svn.cern.ch/reps/lhchcg/trunk/cms/tests/ (rev 124)
names for workspace, modelconfig, dataset: w, ModelConfig, data_obs
Workspace Method Code Result Notes
counting-B5p5-Obs6-Syst30B PL Giovanni 7.09868  
Mingshui 7.09867  
Aaron 7.09886  
LEP Giovanni 7.30 ± 0.06 P-vals for r=7: CLb = 0.646 ± 0.006; CLsplusb = 0.0422 ± 0.0013
7.33 ± 0.02 Cross-check that running from a grid works
Mingshui 7.30 ± 0.01 P-vals for r=7: CLb = 0.655 ± 0.005; CLsplusb = 0.041 ± 0.002
LHC Giovanni 7.10 ± 0.05 P-vals for r=5: CLb = 0.608 ± 0.003; CLsplusb = 0.0939 ± 0.0009
Mingshui 7.18 ± 0.11 P-vals for r=5: CLb = 0.608 ± 0.004; CLsplusb = 0.095 ± 0.003
Aaron 7.13 ± 0.02 P-vals for r=5: CLb = 0.608 ± 0.002; CLsplusb = 0.0949 ± 0.0005
counting-B5p5-Obs6-Syst30S PL Giovanni 7.41476  
Mingshui 7.41452  
Aaron 7.41423  
LEP Giovanni 7.62 ± 0.11 P-vals for r=7: CLb = 0.418 ± 0.006; CLsplusb = 0.0876 ± 0.0018
Mingshui 7.61 ± 0.04 P-vals for r=7: CLb = 0.687 ± 0.005; CLsplusb = 0.069 ± 0.003
LHC Giovanni 7.38 ± 0.02 P-vals for r=5: CLb = 0.614 ± 0.003; CLsplusb = 0.0886 ± 0.0009
  P-vals for r=5: CLb = 0.6105 ± 0.0014; CLsplusb = 0.0907 ± 0.0004
Again with RooStats, just larger statistics
Giovanni(2)   P-vals for r=5: CLb = 0.6121 ± 0.0014; CLsplusb = 0.0906 ± 0.0004
RooFit-independent code
Mingshui 7.40 ± 0.06 P-vals for r=5: CLb = 0.612 ± 0.004; CLsplusb = 0.0904 ± 0.003
Aaron 7.38 ± 0.03 P-vals for r=5: CLb = 0.616 ± 0.002; CLsplusb = 0.0909 ± 0.0005

A It was typo "r=5"

Best fit values of the nuisance parameters for LHC frequentist:

Workspace Code Result (μ = 0) Result (μ = 5.0) Notes
counting-B5p5-Obs6-Syst30B Giovanni 0.095151 -0.49363  
Mingshui 0.095156 -0.49383  
counting-B5p5-Obs6-Syst30S Giovanni 0 (exact) -0.46005  
Mingshui 0 (exact) -0.46011  

CMS Toy Models

under svn: https://svn.cern.ch/reps/lhchcg/trunk/cms/toy-gg.WW.ZZ4l-comb/ (rev 124)
names for workspace, modelconfig, dataset: w, ModelConfig, data_obs

Workspace Method Code Result Notes
cms-zz-mh140 PL Giovanni 1.16787  
LEP Giovanni 1.805 ± 0.018 P-vals for r=1.7: CLb = 0.329 ± 0.004; CLsplusb = 0.0189 ± 0.0006
LHC Giovanni 1.691 ± 0.004 P-vals for r=1.7: CLb = 0.285 ± 0.001; CLsplusb = 0.0139 ± 0.0001
Mingshui 1.689 ± 0.008 P-vals for r=1.7: CLb = 0.295 ± 0.002; CLsplusb = 0.0144 ± 0.0004
Aaron 1.66 ± 0.04 P-vals for r=1.7: CLb = 0.31 ± 0.01; CLsplusb = 0.013 ± 0.001
cms-gg-mh140 PL Giovanni 4.29972  
LEP Giovanni 3.805 ± 0.029 P-vals for r=4: CLb = 0.648 ± 0.006; CLsplusb = 0.0290 ± 0.0010
LHC Giovanni 3.96 ± 0.01 P-vals for r=4: CLb = 0.688 ± 0.003; CLsplusb = 0.0343 ± 0.0007
LHC Mingshui 3.98 ± 0.02 P-vals for r=4: CLb = 0.688 ± 0.003; CLsplusb = 0.0344 ± 0.0009
LHC Aaron 4.0 ± 0.04 P-vals for r=4: CLb = 0.662 ± 0.01; CLsplusb = 0.033 ± 0.002
Asymptotic CLs Aaron 3.96289  
cms-ww-mh140 PL      
LHC Mingshui 0.501 ± 0.007  
Aaron 0.526 ± 0.006 P-vals for r=0.5: CLb = 0.186 ± 0.003; CLsplusb = 0.0110 ± 0.0003
Giovanni 0.517 ± 0.003 P-vals for r=0.50: CLb = 0.183 ± 0.004; CLsplusb = 0.0110 ± 0.0006
P-vals for r=0.55: CLb = 0.182 ± 0.004; CLsplusb = 0.0075 ± 0.0005
(error could be underestimated, but it's definitely between 0.50 and 0.55)
cms-combined-mh140 PL      
LHC Mingshui 0.401 ± 0.005  
Aaron 0.439 ± 0.005  
Giovanni 0.418 ± 0.008
0.431 ± 0.006
New NLL code
Old NLL code (and different random seed)
Asymptotic CLs Aaron 0.43459  

ATLAS Toy Workspaces

under svn: trunk/atlas (rev 124)

Workspace Method Code Result Notes Workspace/ModelConfig/Dataset
ZZ/ATLAS_H_4l_140_[...].root LEP Giovanni   P-vals for r=4: CLb = 0.755 ± 0.008; CLsplusb = 0.0421 ± 0.0020 combined, ModelConfig, asimovData
LHC Aaron 4.25 ± 0.02  
Giovanni 4.32 ± 0.05 with fixed code
gamgam/gg_140.root LHC Aaron 5.80 ± 0.03   WS, ModelConfig, dataset
Giovanni 5.76 ± 0.02 P-vals for r=6: CLb = 0.498 ± 0.007; CLsplusb = 0.020 ± 0.001
HWW_Njet_ext/ww2l_140.root PL Giovanni 0.707268   ws, modelConfig, data
LEP Giovanni   P-vals for r=0.7: CLb = 0.458 ± 0.010; CLsplusb = 0.0954 ± 0.0029  
LHC Aaron 0.76 ± 0.02 taken from email  
Giovanni 0.760 ± 0.010 with fixed code  
Asymptotic CLs Aaron 0.746597    
comb_HWW_Njet_gamgam_zz/
140_uncorr_zz.root
PL Giovanni 0.594858   combWS, ModelConfig, combData
LHC Aaron 0.674 ± 0.022 taken from email
Giovanni 0.667 ± 0.009 with fixed code  

Combination

ATLAS version:

  • under svn: https://svn.cern.ch/reps/lhchcg/trunk/combination/atlas_version/140_update.root (rev 203)
  • names for workspace, modelconfig, dataset: combWS, ModelConfig, combData
CMS version:
  • under svn: https://svn.cern.ch/reps/lhchcg/trunk/combination/cms_version/toy-gg+ww+zz/comb_withzz-mh140.root (rev 193)
  • names for workspace, modelconfig, dataset: w, ModelConfig, data_obs

Workspace Method Code Result Notes
Atlas version PL Aaron 0.373875  
Giovanni ??  
LHC Aaron 0.408 ± 0.014  
CMS version PL Giovanni 0.37385  
Aaron 0.373918  
LHC Giovanni 0.410 ± 0.005  

Timing

Workspace Method Code Time/toy Notes
Atlas ZZ LEP Giovanni 0.07 s  
LHC Giovanni 0.53 s  
Atlas WW LEP Giovanni 0.62 s  
LHC Giovanni 6.6 s  
Atlas Comb LHC Giovanni 13 s  
CMS γγ LEP Giovanni 0.09 s  
LHC Aaron 1.8 s  
LHC Giovanni 1.0 s  
CMS Comb LHC Aaron 25 s  
CMS Comb LHC Giovanni 6.4 s  
All Comb LHC Aaron 77s  
All Comb LHC Giovanni 47 s !!  

Notes:

  • for Giovanni's measurement, the time is averaged between S+B toys and B-only toys with fractions 80%/20% (that's the mixture the code throws when computing CLs)
Edit | Attach | Watch | Print version | History: r33 < r32 < r31 < r30 < r29 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r33 - 2011-06-27 - AaronArmbruster
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    LHCPhysics All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback