Once you disconnect the yellow wire, there isn't a complete circuit for the frequency to ride on. You have to leave it connected. The high voltage wire could be inducing another frequency to the wire, which corrupts the signal from the MAF. The question you have to resolve is why the frequency does not seem rational, and does not vary in response to engine air flow.

 

Here are some more updates on this still unresolved issue.

I checked the resistance of the yellow wire between the MAF and the PCM. Since it was just barely greater than zero I ran a new wire. This had identically zero resistance by my multimeter so I'll call that good. This didn't solve the problem but I was able to detect a very slight change in the average frequency after I ran the new wire. The frequency was about 8 khz at idle and bumped to about 9k at 3000 rpm.

I figured that means the sensor is at least reading, so I tried replacing the PCM, but that didn't change anything. Still reading 471 g/s on Datamaster. So now I'm completely stumped. Bad PCM ground maybe?

I did pick up a copy of TunerCATS so I can at least tune for speed density, since that is how I'm going to have to roll until I get this figured out. Or maybe if I can get the speed density tune good enough then I'll just leave the MAF out of the equation.