Snips for clarity of musing to be posited below...

Not all processing of data yields useful information - and if it is bad
more often than it is good information you are better off ignoring it.

I honestly haven't had the time to dissect your
results at the level of granularity it would take to form a counter-model
I remember thinking that I disagreed with it on first-principles; i.e. the
way the model was set up using comparative climb rates in circling flight.
I tend to agree that thermals are "perishable" but knowing what others are
doing is valuable and may not actually result in a pilot doing the same
thing as the others. I'm not sure how you would use the statistics you
gathered to model some of those decisions...

Writing as someone without a dog in this (genuinely interesting) philosophical
contest, I'd guess from the exchange above that I'm not the only interested
side-liner to muse along the lines of, "Just because it (isn't/can't be)
easily measured doesn't mean it's not happening." That's not by way of
expressing an opinion on a preferred outcome, but by way of noting that "stuff
is happening" regardless of if or how easily it may - or may not - be measurable.

I've worked in manufacturing engineering for more than one high-tech
design/manufacturing company in which managerial actions were mostly driven by
apparently unquestioning belief in the concept, "If it (can't be
measured/doesn't show on the books), it's not (real/happening/action-worthy)."
Two such employers subsequently declared bankruptcy. Prior, in both
environments, I'd ultimately concluded that trying to make a case that
"unmeasurable effects are in fact real" (and further arguing that, in the
companies' cases, would sooner or later have significantly deleterious
effects/detectable/measurable outcomes), was a waste of personal energy. (For
those wondering, following those conclusions, on the line-support front I
opted to do what I thought was right/necessary, while avoiding to every extent
possible becoming ensnared in "stupid managerially-driven cluster-f*@%ing
wastes of personal time/energy"...slept better, too.)

The engineer in me seriously wonders if "determining when leeching takes
place" falls into the unmeasurable category...whether the goal is to determine
the presence of "acceptable" leeching or the "unacceptable kind."

Bob - still sleeping well - W.