Sign convention in AFBtot observables

[blackstonep]

In the course of trying to reproduce the results of [1], viz. Figure 1 and related values in Table III of [2] for A_FB (B0->D*lnu), I seem to have identified a sign discrepancy between the SM prediction reported in [2] and that of flavio. That is,

>>> flavio.sm_prediction('AFBtot(B0->D*munu)')
-0.19969275389101449
and
>>> flavio.sm_prediction('AFBtot(B0->D*enu)')
-0.2041770212914189

whereas the corresponding values in [2] are the positive of these. It does not seem to me that the difference is a definitional one. One may seek line 360 in flavio.physics.bdecays.bvlnu.py for the offending minus sign but its presence is not apparent to me, an intermediate user of python.

Making this adjustment brings my $\Delta A_{\rm FB}$ contour into qualitative agreement with Fig. 1 of [1], which was produced using flavio, but I do not see mention of this in past issues if the authors there encountered it. Is this a bug in the code worth addressing, or does the fault lie with my understanding?

[1] = 2106.09610
[2] = 2104.02094

[peterstangl]

I think the reason for the different signs is just different conventions used in the literature for the S_i angular observables for i in [4, '6s', '6c', 7, 9]. AFB depends on 6s and 6c, so the different conventions lead to an overall sign difference in AFB. In the bvll case, the "experimental convention" used by LHCb has been adopted in flavio, by changing the signs with respect to the previously used "theory convention". It seems that for bvlnu the signs in flavio follow the "theory convention", while the papers you quote use the other convention.

Since this seems to be just a matter of convention, it is not a bug. But I agree that it can be confusing and it might make sense to think about which convention we should use in flavio for bvlnu.