I have seen about 180 difference between cylinders when coming up from ambient but then they stabilize equal. 800 degree variance in my opinion would equate to a very noticeable if not audible performance imbalance. So if you do not notice that, I would have to bet you are seeing probe failure. Switch the leads from channel A to B and vice versus and see if the problem follows the probe. That will tell you for sure. You could also infra red guage the y pipe at each probe location and compare those temps also. They will not be the same as probe tip temps but should be equal to each other across each location unless you truely have a temp discrepancy as the guage is indicating.