BJ has a bit of a point regarding odds [see below]. I think I know where he is coming from.

Our ratings [or anyone else's] are never 100% correct. If they were there would be no horse racing, at least none to bet on.

If we are 25% to 35% correct with our ratings we are doing as reasonably well as anyone else's ratings.

Given BJ sample:
Imagine you rate the field.
Horse A) is rated as a $4 horse.
Horse B) is rated as a $5 horse.
A) on the books is $2.
B) on the books is $100.

I would take the ['value'] $100 offered on B) forever and a day, every time anyone was silly enough to offer those odds. Why? Elementary. Our ratings aren't foolproof and B will get up [very] often enough for us to quickly clean out any bookie stupid enough to offer that price.

I would take ['value'] $10 on B) forever and a day too and again quickly clean out the bookie for exactly the same reasons already outlined. However, if offered $7 on A) and $10 on B), I would back A) every time. Get the drift?

Backing prices wins money long term. Backing horses never will. Finding prices over what you rate, taking your ratings SR into account, is the name of the game.

If I rate 2 horses in a race at A$4 and B$6 to win and one of them is offering 25% above it's price, I will back the 'over' price even if it is B [or C if there is one] because my SR says I will be wrong about A and B [and C] often enough to make money by taking the best price, not what I think is the best horse.
Do that 100 times and your a winning punter in the end. Back horses and not prices and you can never win longterm.