Barny, in my short time in the caper i've managed to get my own ratings together based on one of Shaun's spreadsheets using data from cyberhorse. I can't comment on how useful unitab's ratings are however in my testing i could've made a flat stakes profit betting my top rater in every race i've tested (both backtesting and forward testing) since 1 April 2012.

Also across that period on Good/Dead tracks a very high level of those winners were 100 raters on tatts. I can post the graph/details when i'm home on lunch break if you wish.

Ratings can be as good as the inputs you have at your disposal to put in to them, and I also think the calculations in which you use the detail can make quite a difference too. You also have to look at how many favourites lost on Saturday as well, typically tatts 100 raters will be 1st or 2nd favourite and so win quite regularly without making a POT due to being overbet. Quite a few of the winners on saturday (to my knowledge) weren't favourites and were paying $10+. Its not very often you'll find a 100 rater on tatts paying $10+.