MAF Tuning, STFT vs WB variance Q

I set up my scanner to capture MAF Err using STFT's (default from HPTuner) and another using my WB (used hpt afr err math).

The diff between the 2 is significant in the lower hz range. The WB table keeps wanting me to add, the STFT's say we've had enough... and by the way it drives I believe the STFT's.

Why is that? I understand many folks like the WB just for WOT, but why is it that far off in the cruising ranges, might I have a problem or is there a specific area I should address only with the WB and only with the STFT's?

If it's a problem maybe some air is getting to the WB after the O2's? That feels like it would tell the WB it's leaner than it should be. It's a new exhaust with a Ball/Socket joint and a welded WB bung that are after the O2's, no noticeable leaks.

Last log attached.
Attachment 160701

Because the car is intentionally switching lean with closed loop to hit the cats with o2. The average works out to be lean.

whoops, I should have mentioned I had set the tune up to scale the MAF. Standalone setup in C10.
CLOSE LOOP - off via ETC
MAF High RPM disable to 100
DEFCO off
COTS off
LTFT - off


This is the tune associated to the log above.
Attachment 160703

Just updating.

Reset to stock MAF Cal and restarted the tuning runs.
After the 3rd tuning drive in and applying the "* -1/2" from the STFT's everything is going the right direction
hz ranges I'm hitting on the reg are between 3375 - 6000
Percent err anywhere from 4.5 to 1 % as of this rev.
Noted that WB % variance similar to STFT % variance so far,
I think the WB starts to run away once the calibration table values are near 1% err reported by the STFT's.

Anyway good so far...