1. Calculate MC asymmetry for reconstructed tracks matched to GEANT pions in events passing MB trigger.

2. Do the same for events passing JP2 trigger.

3. Take the difference MB - JP2 for each of the regular GRSV scenarios, DSSV, and the P* and M* distributions.

4. Assign an asymmetric, bin-by-bin systematic taking the larger of the max difference between MB and JP2 asymmetries or the max uncertainty on the JP2 asymmetry.

The size of the systematic depends of course on the scenarios that are included. We have ample evidence from previous measurements for excluding GRSV-MAX from the calculation. Here are three calculations of the systematic uncertainty that exclude different scenarios. The annotations in parentheses after each value indicate whether the systematic was assigned because of the difference between MB and JP2 (value) or because of the limited statistical precision of the JP2 asymmetry (stats):

Dropping only MAX

Calculating Systematic Uncertainty for pi-
[2.00 - 3.18]   -0.012 (value)  +0.008 (stats)
[3.18 - 4.56]   -0.013 (value)  +0.010 (stats)
[4.56 - 6.32]   -0.016 (value)  +0.012 (stats)
[6.32 - 8.80]   -0.021 (value)  +0.013 (stats)
[8.80 - 12.84]  -0.009 (value)  +0.018 (value)

Calculating Systematic Uncertainty for pi+
[2.00 - 3.18]   -0.008 (stats)  +0.008 (stats)
[3.18 - 4.56]   -0.035 (value)  +0.014 (value)
[4.56 - 6.32]   -0.027 (value)  +0.012 (value)
[6.32 - 8.80]   -0.041 (value)  +0.031 (value)
[8.80 - 12.84]  -0.010 (stats)  +0.010 (stats)

Dropping MAX and MIN

Calculating Systematic Uncertainty for pi-
[2.00 - 3.18]   -0.012 (value)  +0.005 (stats)
[3.18 - 4.56]   -0.013 (value)  +0.008 (value)
[4.56 - 6.32]   -0.016 (value)  +0.008 (stats)
[6.32 - 8.80]   -0.021 (value)  +0.010 (stats)
[8.80 - 12.84]  -0.009 (value)  +0.011 (value)

Calculating Systematic Uncertainty for pi+
[2.00 - 3.18]   -0.005 (stats)  +0.005 (stats)
[3.18 - 4.56]   -0.035 (value)  +0.014 (value)
[4.56 - 6.32]   -0.027 (value)  +0.011 (value)
[6.32 - 8.80]   -0.041 (value)  +0.012 (value)
[8.80 - 12.84]  -0.010 (stats)  +0.010 (stats)

Dropping MAX, MIN, and P*

Calculating Systematic Uncertainty for pi-
[2.00 - 3.18]   -0.005 (stats)  +0.005 (stats)
[3.18 - 4.56]   -0.005 (value)  +0.008 (value)
[4.56 - 6.32]   -0.007 (stats)  +0.007 (stats)
[6.32 - 8.80]   -0.005 (value)  +0.003 (stats)
[8.80 - 12.84]  -0.004 (value)  +0.011 (value)

Calculating Systematic Uncertainty for pi+
[2.00 - 3.18]   -0.004 (stats)  +0.004 (stats)
[3.18 - 4.56]   -0.013 (value)  +0.014 (value)
[4.56 - 6.32]   -0.011 (value)  +0.011 (value)
[6.32 - 8.80]   -0.021 (value)  +0.012 (value)
[8.80 - 12.84]  -0.007 (value)  +0.004 (stats)

Finally, here are the plots of the MB and JP2 asymmetries for each scenario including statistical uncertainties. I am concerned that the MB asymmetries do not agree with real NLO pQCD calculations, but I’ve never found a way to fix that.