Updates for fakes with u_T^\mu bins and other fixes #643

cippy · 2025-12-24T08:48:42Z

This PR deals with estimating fakes when splitting the events in bins of uT, which uses the estimate in one bin to port to the other with additional corrections. Details and motivation about the procedure will be given in a dedicated presentation.

This PR also adds other features or fixes, often related to the fakes but not necessarily. Some of these might go into an independent PR, but in many cases they are entwined with other changes needed here.

This PR would require some changes in wremnants-data to add new files, this will come later.

… analysis bins of specific axes (e.g. utAngleSign) which have identically zero yield in the A-Ax-B-Bx regions (cherry picked from commit 773c1d9)

… analysis bins of specific axes (e.g. utAngleSign) which have identically zero yield in the A-Ax-B-Bx regions

…current version of Rabbit, it will be probably reverted when the definitions of poi/noi/nuisanceNotConstrained are fixed

…ion on uT>0 with an hardcoded factor (normalization)

…ream_cherrypick Merge from WMass/main

…th a pT dependent shape

davidwalter2 · 2026-01-04T18:51:40Z

There are differences in the CI:
This PR: https://cmsmwbot.web.cern.ch/WMassAnalysis/PRValidation/PR643/2026_01_03/
Reference: https://cmsmwbot.web.cern.ch/WMassAnalysis/PRValidation/ReferenceRuns/2026_01_02_3d5404b/
Is this expected? If so, why?

cippy · 2026-01-04T20:50:32Z

There are differences in the CI: This PR: https://cmsmwbot.web.cern.ch/WMassAnalysis/PRValidation/PR643/2026_01_03/ Reference: https://cmsmwbot.web.cern.ch/WMassAnalysis/PRValidation/ReferenceRuns/2026_01_02_3d5404b/ Is this expected? If so, why?

I'll check, it was not supposed to change the nominal workflow

…ream_cherrypick Merge from WMass/main

cippy · 2026-01-07T21:21:02Z

I think I fixed it, it was a change in the default width variation for the fit, which I wanted to make only for the width as poi but got applied for any case.

This PR: https://cmsmwbot.web.cern.ch/WMassAnalysis/PRValidation/PR643/2026_01_07/
Ref: https://cmsmwbot.web.cern.ch/WMassAnalysis/PRValidation/ReferenceRuns/2026_01_07_7a59b61/

However, before merging I want to clean a few things up

…herrypick Merge from my branch

…ream_cherrypick Merge from WMass/main

cippy · 2026-01-11T16:26:48Z

I think this version is ready for review, no change in CI as expected
This PR: https://cmsmwbot.web.cern.ch/WMassAnalysis/PRValidation/PR643/2026_01_11/
Ref: https://cmsmwbot.web.cern.ch/WMassAnalysis/PRValidation/ReferenceRuns/2026_01_09_2e3b7c0/

davidwalter2 · 2026-01-12T20:12:57Z

scripts/histmakers/mw_with_mu_eta_pt.py

-    filePath=args.muRmuFPolVarFilePath,
-    noUL=True,
-)
+# # Helper for muR and muF as polynomial variations


can this be removed?

I don't know. We are not using it at the moment, but it was supposed to be used and possibly even become the default one year ago. @bendavid what do you think? If we remove it, I would try to keep track of the commit where it is implemented to make it easy to restore it if/when we need it back

davidwalter2 · 2026-01-12T20:13:06Z

scripts/histmakers/mw_with_mu_eta_pt.py

-                    storage=hist.storage.Double(),
-                )
-            )
+    # if isWorZ and not hasattr(dataset, "out_of_acceptance"):


can this be removed?

I'll check, probably yes

davidwalter2 · 2026-01-12T20:18:11Z

scripts/plotting/makeDataMCStackPlot.py

-                select[axis] = hist.overflow
-            else:
-                select[axis] = int(value)
+if args.selection and args.selection != "none":


In the logic before it was possible to disable the default selection with '--selection none' which doesn't seem to be possible anymore, or is it?

You might be right, I need to check it better, this part was originally added by Ruben because in some configurations the previous code was not plotting the fakes as desired, but maybe this workaround icebreaking something else.

davidwalter2 · 2026-01-12T20:19:37Z

scripts/plotting/makeDataMCStackPlot.py

        lowerLegPos=args.lowerLegPos,
        lower_leg_padding=args.lowerLegPadding,
        subplotsizes=args.subplotSizes,
+        x_vertLines_edges=args.vertLineEdges,


Is this change propagated into wums? Does this require an update of wums then?

yes, it is already in wums

davidwalter2 · 2026-01-12T20:22:34Z

utilities/styles/styles.py

    "pdfMSHT20AlphaS": "<i>α</i><sub>S</sub> PDF",
    "pdfCT18ZAlphaS": "<i>α</i><sub>S</sub> PDF",
-    "pdfCT18ZNoAlphaS": "PDF",
+    "pdfCT18ZNoAlphaS": "PDF no <i>α</i><sub>S</sub>",


I think it would be better to relabel the other entries with "PDF + alphaS" instead and keep the NoAlphaS as simply "PDF"

Ok, I agree and I'll do that

davidwalter2 · 2026-01-12T20:32:22Z

wremnants/datasets/datagroups.py

+        # we need to warn the class that these don't belong to fit axes nor fakerate axes,
+        # but are also not "integration axes" because they are removed before entering
+        # the fake estimate, this can happen for example
+        # with option --select in setupRabbit.py or --presel in makeDataMCStackPlot.py,


But the '--select' and '--presel' set the "globalAction" which is always applied before the fake estimation. Instead of adding "histAxesRemovedBeforeFakes" it is also possible to initialize the histselector with these axes removed. This is how it should be done in setupCombine, if not, would this not be the more suitable solution?

I think the problem was that somehow the list of integration axes is evaluated before the integration actually happens, and then the code doesn't allow you to run the full smoothing when that happens. Or, even if the axis is removed by integration, the list of integration axes is not updated afterwards so the code thinks it still has to integrate those axes, which it doesn't like.
Are you saying the previous code was already able to do the same? To me it was not running when my histograms had N axes but I was using --fitvar and --fakerateAxes with less axes

davidwalter2 · 2026-01-12T20:32:57Z

wremnants/datasets/datasetDict_v9.py

    "DYJetsToMuMuMass10to50PostVFP": {
        "filepaths": [
            "{BASE_PATH}/DYJetsToMuMu_M-10to50_H2ErratumFix_TuneCP5_13TeV-powhegMiNNLO-pythia8-photos/NanoV9MCPostVFP_{NANO_PROD_TAG}",
+            # "{BASE_PATH}/DYJetsToMuMu_M-10to50_H2ErratumFix_TuneCP5_13TeV-powhegMiNNLO-pythia8-photos_ext1/NanoV9MCPostVFP_{NANO_PROD_TAG}",


can this be removed?

It will be removed at some point, the current issue is that in Pisa's machine I don't have permissions to copy the files in the other folder, so I had to temporarily create a new path and put them there.

…herrypick Merge from cippy/main_mergeUpstream_testRuben

…ream_cherrypick Merge from Wmass/main

cippy · 2026-01-19T21:31:34Z

I tried to apply fixes according to some commits.
About the comments, for the moment I left them because it is not clear to me if that code is ever going to be needed now that we have the PDF/QCD scale weight applied through the helicities. If not, we can probably remove all of them completely, otherwise the best would be to make those histogram be produced only on demand when needed, since I think they slow the histmaker down.
For the part about the fakes, I agree that the current logic is a bit cumbersome, but since that part of the code is quite intricate I would leave it as it is for now and try to improve it in a different PR (possibly also fixing the pseudodata for the fakes).

cippy · 2026-01-20T21:29:33Z

This PR: https://cmsmwbot.web.cern.ch/WMassAnalysis/PRValidation/PR643/2026_01_19/
Reference: https://cmsmwbot.web.cern.ch/WMassAnalysis/PRValidation/ReferenceRuns/2026_01_20_27e2e9b/

The only difference should be in the renaming of the PDF impact groups, in particular calling "PDF" as "PDF+\alpha_S"

cippy and others added 24 commits October 17, 2025 15:44

generalize eta labels better for plot_decorr_params

fc992c4

plotting and testing

d330024

Protection on the fakes estimation: fake yield is set to zero for the…

09c630f

… analysis bins of specific axes (e.g. utAngleSign) which have identically zero yield in the A-Ax-B-Bx regions (cherry picked from commit 773c1d9)

Protection on the fakes estimation: fake yield is set to zero for the…

8c5d50d

… analysis bins of specific axes (e.g. utAngleSign) which have identically zero yield in the A-Ax-B-Bx regions

small changes for plotting

e5ed649

consider the polVar as noi - workaround to better interface with the …

6085af8

…current version of Rabbit, it will be probably reverted when the definitions of poi/noi/nuisanceNotConstrained are fixed

fake prediction for uT<0 is now evaluated by scaling the ABCD predict…

5f8ce1f

…ion on uT>0 with an hardcoded factor (normalization)

label change for the utAngleSign axis

c9354b7

small fix

e433ca3

Merge branch 'main' of github.com:WMass/WRemnants into main_mergeUpst…

4bc98b5

…ream_cherrypick Merge from WMass/main

moved reference commit of rabbit

c79eca9

remove Zmumu OOA for theory agnostic with pol var

4c95789

fakes in uT<0 are evaluated by scaling the ABCD prediction on uT>0 wi…

488f6cc

…th a pT dependent shape

decorrelate FakeParam syst in uT bins

9293126

new syst for eta nonclosure in ut<0 bin

3f96a4c

updates for veto (no default change)

fec8ace

merge from rubenforti/signUtAnalysis

e7c51cc

updates

65201f6

updates for fakes vs uT

f1a26ce

updates

ea0dfe2

set noi=False for polynomial variations

2fc7884

fixes and updates to fit uT^mu

811a7b0

foo

ba08ba7

fixes

4d1e7b5

cippy changed the title ~~Do not merge: testing some updates~~ Updates for fakes with u_T^\mu bins and other fixes Jan 3, 2026

cippy added 3 commits January 6, 2026 18:46

fixes

ea60c34

Merge branch 'main' of github.com:WMass/WRemnants into main_mergeUpst…

60299aa

…ream_cherrypick Merge from WMass/main

set default W width variation to original 0.6 MeV

1cd1a13

cippy added 4 commits January 11, 2026 17:18

updates

98d150a

Merge branch 'main_mergeUpstream_testRuben' into main_mergeUpstream_c…

8a3041a

…herrypick Merge from my branch

restore default dictionary for datasets

4369f3a

Merge branch 'main' of github.com:WMass/WRemnants into main_mergeUpst…

faa2675

…ream_cherrypick Merge from WMass/main

davidwalter2 reviewed Jan 12, 2026

View reviewed changes

cippy added 5 commits January 14, 2026 23:19

updates

02f2540

updates

78f3057

Merge branch 'main_mergeUpstream_testRuben' into main_mergeUpstream_c…

5683c7e

…herrypick Merge from cippy/main_mergeUpstream_testRuben

implement some comments for the PR

32b866c

Merge branch 'main' of github.com:WMass/WRemnants into main_mergeUpst…

cb76f64

…ream_cherrypick Merge from Wmass/main

Updates for fakes with u_T^\mu bins and other fixes #643

Are you sure you want to change the base?

Updates for fakes with u_T^\mu bins and other fixes #643

Uh oh!

Conversation

cippy commented Dec 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

davidwalter2 commented Jan 4, 2026

Uh oh!

cippy commented Jan 4, 2026

Uh oh!

cippy commented Jan 7, 2026

Uh oh!

cippy commented Jan 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

cippy commented Jan 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

cippy commented Jan 20, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

cippy commented Dec 24, 2025 •

edited

Loading

cippy commented Jan 11, 2026 •

edited

Loading

cippy commented Jan 19, 2026 •

edited

Loading