Hello Russell,
It's good to hear that the "book" doesn't lie.
At 86% (I assume you picked that figure off the "% Throttle" from the datalogger) you might see a voltage overhead issue with old batteries on a cool day.
If it ever happens, you would notice a drop-off in power in the hourglass and clover.
There is no need to concern yourself, don't change a thing; after a whole season, please tell us if you've you find that 86% works for you, or if you think that you have ever run short of battery voltage at the end of a flight.
Then we will all learn a bit more.

Even though the average current is the same, you must remember that the transistors are being turned on and off rapidly to make that average current flow.
If the average current is the same, and the ON time is less, then the current peaks are higher.
The peak currents are what strongly affect transistor heating, rather than the average.

take care,
  Dean P.

Well, the trouble is Dean, I won't be using this setup on this plane as the 12 x 6 three blade prop really does work the best on my heavy plane.

One thing I learnt with IC is that heavy planes tend to work better with higher pitch props. (Or more diameter, or both.)

(I'm seeing the same with electric.)

When I get my lighter plane, the lower pitch props will probably work well. ☺

Russell:
First off CONGRATS on your recent NATs performance – WELL DONE
 
Thank you also for sharing your data on your experiences.  The comments added about power spikes reminded me of something.  I did a square wave sheet plotter in Excel – I was trying to figure out how to match-up kv, cell counts, PWM rates & RPM to create some kind of thermal modeling scheme.  I got as far as the square wave chart, but could not figure out how to make the leap from there to HEAT so I let it slide.  However I dusted it off to see how it looked with your data in it.  The input data looks like this:
 
 
Base: 12x6 Robinson                            Alt: 11x5.5 Ressinger
KV=930                                             KV=930
Lipo Cells = 4                                     Lipo Cells = 4
PWM Rate = 8                                    PWM Rate = 8
Efficiency = 90%                                Efficiency = 90%
Max RPM = 12,388                              Max RPM = 12,388
Base RPM = 9,200                               Alt RPM = 10,700
% of Max = 74.27%         % of Max = 86.38%
 
At first the RPMs & throttle percentages were way off, but I reasoned I needed a motor efficiency in there.  Some trial and error showed that 90% made everything come together.  I do not think any outrunner claims that kind of efficiency, so I probably need to re-think the cell voltages some too.  However – this works!  Indeed the throttle percentage (% of Max) matches up right on top of your data!
 
The resulting square waves are shown below.  The black lines are for your baseline (12x6 Barry Robinson) the red lines re for the 11x5.5 Alan Ressinger prop.  The solid lines are for the steady state RPM, the dashed lines show how much ON & OFF time is required for the RMS average to generate the steady State RPM.
 
Unfortunately, there is nothing here that really links up to the TEMPERATURE you recorded.  Maybe it would be possible to work backward from your recorded temperatures and try to link that to the differences in square waves?

Finally, I know you were also working with the APC 13x4.5 & 13x5.5; do you have any comparable data from those props?  I would be curious to see where those measure up with respect to RPM & heat - gotta hunch they would be just about on-top of the 12x6 BR..?

Hi Russell
I can run the plot at any PWM but chose to show the chart with 8 because it is the largest (widest) steps and the easiest to see.  I also set my Castles for the "16" setting and it seems to run smoother.