HI all!

Some time ago, I mentioned that my rating works on a large number of tracks almost embarrasingly well, yet it fails miserably on others.
I just learned to live with that, but now I have a question, why is it that the favs are winnig on one track in a disproportional manner, but fail on others.

Following on from Wesmip's idea I ran a programme to find what happens to the favs on different tracks under different track conditions.
(from 2000 to 2006)

Here are two glaring examples of what happens.

Gilgandra:

Track........Races ..Winners..StrikeR%.....Ret.........ROI.....Track_ cond
Gilgandra.....8...........4.........50.00%.....$.. 8.60.....7.50%.......fast
Gilgandra....70.........35.........50.00%.....$93. 20...33.14%......good
Gilgandra....32.........16.........50.00%.....$41. 40...29.38%......dead
Gilgandra.....4...........1..........25.00%....$.. 2.40..-40.00%.....heavy
--------------------------------------------------------------
While a similar class of track, Kempsey, produced this

Track......Races.. Winners....StrikeR%.....Ret...........ROI......Tra ck_cond
Kempsey....68.........17..........25.00%....$39.40 .....-42.06%.....good
Kempsey....39..........8...........20.51%....$21.6 0.....-44.62%.....dead
Kempsey.....8...........1..........12.50%....$..2. 20.....-72.50%....heavy

(there were no records under some track conditions)

There are quite a number of tracks, where under, fast or good conditions the loss is minimal or actually there is a slight win despite the TAB take.
All returns are Supertab VIC.
Since this covers more than 6 years and there is very little variation from year to year, it's a fairly reliable indicator of the track behavior, BUT WHY?, after all, a race is a race and the same horses are running around.
It would suggest that there is some hope, in following the favs, given some judicious thought as to where?.
Anyway, if anyone has an idea why this is so? you are most welcome, I'm sure.
Cheers