Speedy, I think the your "scanner dyno" results and these real-life dyno results are confirming something: that the ECU doesn't know what it's doing anymore. The reported gains from users are anywhere between 10hp and 35hp!
I'd estimate the gains from relieving the pressure drop in the intake path should amount to 15-20hp; anything more or less is likely coming from the air/fuel ratio changing. Either it'll lean it out (see above post) and you'll make more power than expected, or the car will get confused and make too large of a correction- and perhaps even pull timing- resulting in lower-than-expected gains. See, when you have the MAF after a bend, the correction needed at an RPM and load will not be the same as the same RPM but a different load. The ECU can't deal with this, but it tries, and these "tries" may be why power gains are so inconsistent.