It appears there are in fact at least a few things I've missed over the yrs, but I'm guessing that this quote appeared on the Bd, a concept foreign to me until what it is now, 6 mos ago? One thing I have learned during that time is not to be surprised both that you knew about this quote & were able to retrieve it on short notice.

I assume that an example of a "ridiculous" top would be the 1 War Emblem moved to on 3/17, and an example of a decent looking top would be the 4 Magic Weisner ran on 3/30. Aha, you say, 3yos in their early development were excluded from the sample.(as an aside, was there any mention of the size and/or the process used to select the sample?)

Although I did not say so in my original post, the excluded group, which I would expand to include still developing 4yos & in some specific instances very lightly raced 5yos, is what interests me for the same reason W. Sutton gave, namely that's where the money is, in this case in terms of 20-1 or higher horses who actually have a chance to and actually do win races. Despite the dramatic difference in their tops, my "impression" would be that both WE & MW would have a chance closer to 50% to either pair their tops or run within 3 pts of their tops( which would include a small fwd move & the so-called small bounce).

Equally important is what to expect in the next race after the pair. In both of the above examples, the horses moved fwd at high odds, so the question is whether a pair is a sign that a horse such as MW will, in the words of the TG analysis, make the fwd move needed "just to become relevant." The reason that question is important is that the best and perhaps one of the only logical ways to find live shots is to project that a horse has a percentage chance of running its best ever no. which is greater than its actual odds. My argument is that when a lightly raced horse pairs up, that is a sign that the chances that the horse will run a new and many times a dramatic new top are greatly increased, a situation which BTW is likely to be present in the Belmont.

Finally, if all horses who run significant new tops(5 pts) are included, and if small(within 2pts) bounces are excluded, the research I've seen is that highest percentage quoted by Robes is off by close to a factor of 3.