I even think the too large mids are on purpose as they increase the directivity at the critical presence region and enhance the euphonic sound power dip there, which can both be advantageous at the typical high end temple too high listening distance and too few absorption.
This is how I account for it as well.
A lot of folks here seem to assume that B&W engineers are idiots and couldn't figure out how to design a speaker that matches the Harman design goals. As in its a huge secret that a 7" midrange will get directional within its passband (shhhh don't tell anyone).
A lot of folks here also seem to assume that Abbey Road and Skywalker Sound engineers are also idiots who are willing to risk their brand and reputation (worth hundreds of millions of $) by using horrible sounding speakers so that they can get a little kickback from B&W.
And finally, a lot of folks here seem to think people who have listened to far more loudspeakers than they have, and own models that perfectly match the Harman goals of smoothness and directivity, are idiots because they obviously can't hear how bad the B&W's must sound.
Many people "think" they understand how loudspeakers "should" measure and they "think" the science around it is complete, just as it is for sota solid state devices. It's not.
You can't reliably predict how a loudspeaker will sound from eyeballing measurements. Even if you computer analyze them, your predictive value is still far from perfect. That's what the science shows, boys. Sorry if this is the first time you're hearing it. I went through my disappointment years ago (the papers have been out for a while).