I agree that this is an excellent article.
I have only one bone to pick with it. Your analysis that 8ms is only 8% of the stick to servo time
is exaggerated. The servo movement time in your analysis is 62% of your average time and really what a pilot
senses is the time at which he detects the model starting to react, the real delay he will sense is
the 30ms to the start of the servo movement and odd change as his eyes and brain process that.
So the 8ms reduction in the time is closer to a 26% improvement in perceived latency. Much more likely to be
significant than the 8% you claim.
Does this mean that we need to run out and get this. I don't really think so, at least not yet. One of
the primary processing miracles in our brains is the ability to adapt to latency, but we are not all
created equal here. Some people have a much better ability to adapt to latency. Interestingly these are
usually the top performers in activities like stunt flying. So ironically, low latency might be a bigger help
to below average performers than the elite fliers. Where low latency might help the elite flier is not in
doing practiced routine stunts but in reacting to the unexpected, such as me flying my PZ supercub into his airshow.