massive reconfiguration time again.

How are you tracking your O2 sensor? By eyeball? Or does your gauge datalog?

Unless the data is being logged electronically, I question whether it may help with diagnosing such a transient event as a bog. Also, the sensor can 'lie' because if your mixture doesn't light off (too rich, or too lean) the sensor can read it as 'lean' either way (you probably already know this, though).

It may be that your solution will have to include a combination of timing and mix/pump changes. A big pumpshot will help cover the loss of signal when going WOT from low revs, but since the velocity is also going to drop also fuel isn't going to atomize as completely. YR has already described how your plugs show evidence of fuel droplets - maybe that's as much of an incomplete mixing situation as it is a lots-of-fuel situation? If so, it seems to me that more ignition advance in the 1200-1800 rpm range might help? I know you've mentioned run-on when you turn it off, so bumping the initial may not be in the cards, but getting it to come in faster might?

Just a couple random thoughts...