Improve forward efficiency for very high PT tracks

cc @flpisani @dcampora @dovombru

Testing the forward on Z->mumu events, it is only around 75% efficient on tracks with PT above 20 GeV, which is a priori surprising. These are super-straight tracks and so I wonder if there is an issue with the momentum guided search window from the UT, that somehow it is very narrow and then if there is some bias in the momentum estimate this puts the window in the wrong place? Anyway I suspect this is worth a dedicated investigation because it might expose some other issues.