In restrospect I should have boxed them. I had 5 to win by 9.9, then there was only 1.5 metres between 841. It was just that 1 was about $2.80 at that time so I wanted a bit of value..
As far as my method goes its pretty basic, and probably already been thrown out on this forum.
I only look at races with 8 dogs (actually 3 place dividends, late scratchings are alright.).
Each dog must have had at least 3 starts.
I work out the speed each dog ran in their last 3 starts, average it and convert to the distance of the current race. I try and convert as close to the decline in speed shown in the track records of all Aus tracks.

The dog with the highest speed is my top rater. For this dog to become a betting situation however it must pass these tests. It must have had the fastest m/s of all 24 races tested. The average race distance over the last 3 must also be within 10% of the current race distance.
At this stage I am still testing, and gathering stats to see how viable this is, and only betting $2 each race, but now thanks to you Moeee, I am also converting the top rater to odds and betting to return $100 (theory).

My basic idea is that the dog must have early speed to have the highest chance of winning. Box draw to me is irrelevant, I think too many people already account for this and is shown in the betting so the gain in odds accounts for anything lost in the race.
So far I have had:
21 betting opportunities
10 wins
$31.10 total returns
Also I had 3 top raters a week ago and put in another race to make up the quaddie which paid $300 and the daily double $40.

Once I work out the downloading of form from the net into excel properly, I will have more data to work with and will add more to it, but at this stage all looks good...