Believe it or not, I actually have some drafts. I was just way too lazy to post them…
Continuing from the Contur cliffhanger… Not yet.
Before getting to Contur, there is one part of the project I need to catch up on: what happened to the original plan of changing Higgs interactions and looking for the consequences?
In the early posts, I introduced the Higgs Effective Lagrangian, or HEL, and started varying three Wilson coefficients, written here as cH, cHW, and cHB. These control additional interactions that can change the rate and kinematics of Higgs production.
For this project, HEL provided an effective-field-theory description based on dimension-six operators, focused on Higgs interactions. It belongs within the SMEFT approach used here; moving to a Warsaw-basis implementation means translating how the interactions are expressed. The HEL model paper describes the implementation behind those initial studies.
The first checks were encouraging. I compared predictions from Herwig and MadGraph, starting with zero Wilson coefficients and then switching on individual coefficients and combinations.
One exploratory benchmark used:
- cH = 0.2
- cHW = 0.1
- cHB = 0.1
At 85 TeV, MadGraph returned a cross section of 197 ± 1 pb, while Herwig returned 205.4 ± 0.3 pb. Both predicted a substantial enhancement over the Standard Model rate.
An important correction to my earlier posts: those two numbers are approximately 4% apart, which is larger than their quoted Monte Carlo integration uncertainties. They show similar overall behaviour, but I should not have described them as agreeing within statistical errors.
The large enhancement was also a benchmark result, rather than evidence that this parameter point was physically viable. Establishing that would require further checks, including the validity of the EFT approximation in the energy region being studied.
The next challenge was expressing these HEL settings in the Warsaw basis used by the SMEFT implementation I was investigating.
A basis is a choice of independent operators used to write the effective Lagrangian. Different choices can describe the same physics, but the coefficients need to be translated consistently. A single nonzero coefficient in one description can become several nonzero coefficients in another.
There are tools designed for this purpose, including Rosetta, which translates between operator bases such as SILH and Warsaw. The existence of a translation tool, however, does not remove the need to check the conventions of the particular model files being used.
This was where the project became more difficult.
The cH comparison could be brought into closer agreement after accounting for an additional normalisation factor involving the new-physics scale and the Higgs vacuum expectation value. That helped identify a convention issue in that comparison; it did not establish a general solution.
The momentum-dependent cHW case remained less satisfactory. The translation produced several Warsaw-basis coefficients, but the resulting setup did not reproduce the much larger HEL cross section.
Several coefficients appearing after translation is perfectly possible. The unresolved issue was whether the complete translated setup reproduced the intended predictions.
There was also the practical requirement of generating usable events. Successfully loading a model and obtaining a cross section are useful checks, but the Rivet analysis needs a reliable event sample. Within the project’s timescale, I did not establish a stable translated SMEFT workflow in Herwig.
That changed the direction of the main analysis.
I continued with a controlled Standard Model setup: Herwig’s built-in implementation for the signal, alongside Standard Model background samples. This allowed the event selection, jet tagging, and signal–background comparison to develop on a stable foundation.
The exploratory HEL studies remained part of the project, while direct Wilson-coefficient constraints remained unfinished.
The next step was to investigate what the completed analysis could tell us statistically. That brought me back to Contur, with a new question: how could I test the interpretation pipeline using simulated measurements?
