Hi Duritz,
I wasn't intentionally identifying the basis of your Handicapping, it was meant to throw up a different thought on selecting the base starting point.

I have not mentioned too much about my own methods of Handicapping, other then to say I use Times substantially but these are smoothed by an Age component.

Back to the original intention of my posting;

People assume Recent form will pave the way to riches an that they can see trends within those recent starts that points to a horse coming into form or out of form or holding steady.

You have to be careful with the "trends of 3-4 races" idea. It's very hard, statistically speaking, to determine whether a grouping is a "trend" (which implies a relationship between them, with performance changing in some direction) or just a random event. A person's bias is to see a "trend" where there is really only a random grouping. Again, the roulette example. If you look at a series of roulette spins, you would see "little trends" where the ball came down red three or four times in a row, or black three or four times in a row. These aren't trends. They are just a random grouping of similar outcomes.
Think about it like this. Let's say you've defined a Horse's performance envelope and divided it into thirds, with a good third, a middle third, and a poor third. We'll assume he has an equal chance on any given performance of falling into any of the three thirds. What kind of "trends" would you expect if, indeed, the Horse's performance is random? Let's say you look at a three-race series.

With totally RANDOM performance you would expect to see:

3 out of 27 where three performance are all in the same third.
12 out of 27 where two are in one third, and 1 is in an adjacent third.

That's 15 out of 27 possible combinations that would LOOK like a Horse was in some reasonably consistent form, whether good or bad. And that's with totally random performance! That's over 55% of all three-race combinations would suggest "form" to the statistically naive viewer. That's why you have to be so careful with what you call a "trend" or "form". Because that's not necessarily the correct description.

Here is a little portion of my Handicapping, Horses usually show a beginning pattern, and an ending pattern in their careers. With a big, steady-state chunk in the middle. I don't call these end-lines "form." They are a different pattern. "Form" implies that a group of races should have similar performance. The beginning/end times aren't similar lines, they ARE "trends." The lines change in a consistent direction. But a true "trend" isn't "form" either.
(Age Effect)
The early-career chunk screws up your "career" averages for quite a ways after the Horse levels out. If you don't remove those Starts from use, they will definitely bias you into thinking that form exists when it doesn't. Because near-term averages will be more predictive than full-career averages. The trick to accurate predictions is NOT to assume form is in play. Instead it is to study enough Horses to know WHEN the early-career pattern is likely to be over, and to take that into account when determining any sort of average/standard deviation for some performance statistic. Toss out those early-career lines, and "form" disappears and you also improve the accuracy of your predictions.

The end pattern is harder to deal with because the beginning of the end is much more difficult to identify. If an older Horse has a bad performance, is it just a randomly bad race, or is it the beginning of his tail-off? This is hard to judge because the age at which Horse's tail off varies hugely. There are all sorts of variables like number of races a Horse has run, frequency of running, genetic robustness, talent level, courses raced on, injury, and so on. Of course, a statistically astute player can look for signals. For example, let's say a Horse has three performances in a row where he is in the bottom third of his performance envelope. That would normally happen randomly about 1 time in 27. If it happens to a Horse that ALSO happens to be past a certain age, that might be taken as a good sign that the steady-state part of the career is over, and a new performance envelope is now in play for that Horse.

I really don't take Handicapping lightly, as I said before, you measure Handicapping by accuracy and wagering by ROI.