Michael

The patterns outputted are not apples and oranges. We apply rigid time and surface criteria, which is not to say there isn't variability among the qualifying patterns. That I surmise is what you mean by apples and oranges. Pardon the nitpicking but let's be a bit more precise in the criticism.

JB, two posts up, brings up two specific instances, (b), where TP may be very significant assuming the horse in question fits. And I would also agree that TP can be very useful for newcomers in terms of quickly identifying better and worse lines.

The extremes are most transparent, but unfortunately they most often occur for poorer lines, i.e. the off and x categories add up to 75% or above. And looking at the respective horse's line, we too can project it won't run, as well. In other words the TP, while transparent, may not be particularly illuminating. But maybe it is, as a disqualifier. It's another piece of data and always has to be used in the context of the field in question.

Good lines beget good TPs--that is as JB notes, two posts up, the top and pair categories add up to a high probability--60% or above--that the horse will run well. Whether the horses are fast enough given the conditions is another question.

One thing to note is that the good TPs invariably show lower good probabilities (top + pair) than the bad TPs show poor probabilities (off + x) for obvious reasons. A horse can't keep running new tops or pairing tops forever. They're mortal.

Hope this helps.