I prepare a selection file with ratings from a machine learning program that analyses Australian gallops racing.
The text file has headings "runner", "rating", "action" where "runner" is in required format for Australia of "entry number" "full stop" "space" "horse name" all enclosed in double quotes.
(See attached example file.)
The "action" is either "back" or "lay" and the "rating" is a number that I would like to use to modify the betting amount in some way: in the case of "back" a higher number could mean to increase the back bet size, in the case of "lay" to offer to lose a larger amount.
In effect the "rating" expresses how good or bad the horse is relative to the other runners in the race.
Initially I would like a vanilla trigger that makes back and lay bets at whatever is on offer in the win market at or past the advertised start time for the race.
(I have in place a Market Locator file that is working OK.)
I envisage starting with a test bank of 1000 with a back bet amount of 2% current bank multiplied by the "rating" figure up to a maximum bet size of some amount, say 40.
For lay bets I envisage to risk say 1% current bank times 1/rating with maximum at risk of 40.
Any advice, pitfalls, suggestions as to alternatives welcome!