consistent 20-40mV difference between O2's at WOT
 

Yes, I know, they're narroband, bla bla bla, but both averages and just a general impression from looking at the graph, it looks like one bank is leaner than the other (at WOT so it's stable) by about 20mV on average, but with swings usually under 40mV.
Is that normal for 87k mile old O2s to do, or is it a loose header or something?
Any other common causes for such behavior?
Thanks,
Marcin

Could be a lazy O2 sensor ... if you swap them (driver to passenger passenger to driver) and the problem follows the O2 then it's probably a lazy O2 sensor

but if the problem stays on that bank , could be an exhaust leak around the collector, or at the head, could be a failing fuel injector (or clogged)

It takes only a little EGT (exhaust gas temperature)
difference side-side to make that much output voltage
deflection, at the lower AFRs. More influence from
that than from actual mixture difference, out on the
flat top of the curve. Especially on LT headers with the
large side-side differences in O2 position and cooling
possible, I'd call this a thing not to waste time chasing.

Agreed. I stopped chasing that ghost when I swapped my sensors side to side and saw the same difference in mv as before. The difference did not follow the sensor, it stayed on the same side.