Apologies for having skipped part of this thread, and if it turns out I'm repeating what others have said:
1. Possibly the boiler has a a maximum temperature drop and, if you throttle the flow to get a higher differential, it responds by cycling to try to reduce the differential down to below its maximum
2. Thinking about condensing. If the flow temperature is high and the return is below condensing, it makes sense that condensation will only form on that part of the heat exchanger that is coolest. So, if condensing starts below 2°M (M = a made-up temperature scale), say, and stops at 4°M, a boiler running at 4/2 will always condense well, whereas a boiler running at 6/2 will only give 2/3rds of the condensing capacity because 1/3 of the heat exchanger is likely to be above the dewpoint which is 4°M. So there is likely to be more to it than just the boiler return temperature. Not my original idea, but heard it somewhere and it makes sense (although probably a slight simplification of the actual physics at work).