search

TopEFT / topeft

15 stars 24 forks source link

Adding fwd regions and lt #387

Open bryates opened 11 months ago

bryates commented 11 months ago

Replacing PR #382 with updates after repo refactoring.

kmohrman commented 11 months ago

Thanks for porting over these changes @bryates. I might suggest holding off on merging this till the HistEFT stuff is in place (just to avoid adding unnecessary complications into the validation).

Additionally, I would suggest that someone (perhaps Wynona, since it seems to fit in with her general sensitivity improvements studies) would run this and double check the size of the sensitivity improvements before merging.

Finally, as a general suggestion, I think it would likely be useful to have an easy method of flipping on and off all of these sensitivity improvements from category splitting (e.g. Wynona's off-Z category stuff, and this forward region stuff) since I would imagine that at some point in the future, you will want to know how big of an improvement each change has compared to the "baseline" (if "baseline" is defined as TOP-22-006 categorization). So I think an option to run with the "baseline" categorization would be probably quite useful moving forward.

bryates commented 11 months ago

Thanks for porting over these changes @bryates. I might suggest holding off on merging this till the HistEFT stuff is in place (just to avoid adding unnecessary complications into the validation).

Additionally, I would suggest that someone (perhaps Wynona, since it seems to fit in with her general sensitivity improvements studies) would run this and double check the size of the sensitivity improvements before merging.

Finally, as a general suggestion, I think it would likely be useful to have an easy method of flipping on and off all of these sensitivity improvements from category splitting (e.g. Wynona's off-Z category stuff, and this forward region stuff) since I would imagine that at some point in the future, you will want to know how big of an improvement each change has compared to the "baseline" (if "baseline" is defined as TOP-22-006 categorization). So I think an option to run with the "baseline" categorization would be probably quite useful moving forward.

Thanks @kmohrman, I agree with all of this. @ywan2 were you interested in exploring this?