Specific details concerning Tracy's cooling setup and performance are best obtained directly from him.

I can try to relate my take on the concepts that were discussed, though:

The area, cross sectional shape, location as far as practical from the center of the prop, and the location close behind the prop have combined to give good inlet characteristics under a wide range of conditions of airspeed and angle of attack.

The construction of the diffuser should result in smooth walls and avoid sudden changes in direction of the air flowing within it.  Divergence angles greater than 7 degrees are not necessarily a pressure recovery killer.

Other than the exact inlet cross sectional shape and area, the concepts just mentioned follow largely from common sense.

The concept that was surprising to me was that a large change in direction of air flow from the diffuser into the core right at the face of the core is not necessarily a bad thing.  Since the flow of the air through the core is intended to be turbulent, the turbulence resulting from the rapid direction change at the core face may actually be a good thing.  In Tracy's coolant exchanger arrangement, the core is almost at a right angle to the direction of air flow within most of the diffuser.  The critical part appears to be the shape of the interface between the core and the diffuser with the diffuser wall distance from the core becoming very small at the downstream end.  The goal is to provide consistent pressure across the whole face of the core and thus get uniform airflow through the whole core.  Achieving this goal will probably involve some experimentation and changes in the diffuser to core transition shape for a particular installation.  Making measurements of airflow through different parts of the core would be essential to asses the effect of changes made.  Although this all sounds daunting, the freedom resulting from the ability to change the direction of airflow at the core face may allow greater emphasis on the construction of a diffuser with good pressure recovery within the constraints of the under cowl space available.

That's just my interpretation of what I thought I heard.  Someone please correct me if I am mistaken.