I was naively paranoid that maybe we ran last night with the target out of the beam somehow, or off to the side slightly. And that this would explain why the polarization did not degrade much. (?)
I called in and talked to Ryan Zielinski, who checked the encoder position recorded automatically during owl shift. He said it was the correct value (whew!) and did not change (I knew that). I remind TO to check this encoder position when on shift. My mistake not to.
Ryan also said the polarization would not exist if the target were not positioned exacly right. It seems it must be centered up on the B field.