Polarized STAR and PHENIX A_L W

[giacomomagni]

Implementation of the polarized asymmetries DY data from RHIC.

STAR datasets:

Dataset	Obs.	Status	Comments	Correlations	Inspire	HepData
STAR_WMWP_510GEV	STAR_WMWP_510GEV_WM-AL STAR_WMWP_510GEV_WP-AL	✅	Figure 5 provides the combination with the 2011-2012 data	Table1 of the Appendix for 2013 run only, not implemented	paper	dataset
		❌	Included in the combination with 2013 data		paper	dataset

PHENIX datasets:

Dataset	Obs.	Status	Comments	Correlations	Inspire	HepData
	$A_L$ in $W^\pm +Z$	❌	see below		paper	dataset
	$A_L$ in $W^\pm +Z$	❌	see below		paper	dataset

[giacomomagni]

Regarding the Phenix measurement we might have a problem.
My understanding is that Phenix has measured $A_L$ in $W^\pm+Z$ process,
despite what they say in the paper title.
In particular I read from here:

Figure 3 shows the combined asymmetry for all of the
PHENIX data sets and published data from STAR [11].
The two data sets cannot be compared directly, because
PHENIX measures the asymmetry from W± +Z decays,
while the STAR result is solely from W± decays. The
comparison can be made through the curves, which account for the specifics of each measurement.

So the numbers that are provided in HepData are not exactly for the same quantity measured by Star.
You can then see a difference of the predictions for the two processes in Figure 3.

@Radonirinaunimi , @enocera do we have a code to produce sensible predictions also for Phenix ?
Do we neglect this difference and use the usual code?
If not, do we want to have them implemented for the future anyway?

[Radonirinaunimi]

Unfortunately, we cannot yet use the mcfm-pol to produce predictions for $Z$ boson production. In principle, that should be doable given that the only difference between the W and Z are the beam functions to which expressions are available here (https://arxiv.org/pdf/1704.05457.pdf) for the latter. But I don't think this will be for NNPDFpol2.0 (?)

That being said, the differences between $W+Z$ and $Z$ seem to be significant for large-pseudorapidities. So maybe we can still use the same codes for $|\eta| < 0.35$.

[Radonirinaunimi]

To allow for these asymmetries to have negative pseudodata.

[giacomomagni]

Looking here:

https://agenda.infn.it/event/38132/contributions/234387/attachments/120751/175847/Trieste-wv.pdf

It seems that our friends do include the 3 datapoints from PHENIX as well in their $A_L$ ($W^+$) datasets...

[Radonirinaunimi]

Looking here:

https://agenda.infn.it/event/38132/contributions/234387/attachments/120751/175847/Trieste-wv.pdf

It seems that our friends do include the 3 datapoints from PHENIX as well in their AL (W+) datasets...

Interesting! I am curious to know how do they do then? Maybe they just don't care about the distinction $W^\pm$ vs $W^\pm \otimes Z$...