Machine learning bet bot

We have a database which predicts horseracing in Australia and we currently have a positive tipping strategy with a bank $ amount allocated to each bet. We need a script/bot which works exactly like a human bettor/gambler. The system must start with an allocated bet per day and decide 1. which races and prospects are the good value or probability chances 2. The strategy for the day. 3. Bet amounts and adjust during the day based on win/loss and strategy. The bot must be dynamic with the aim of returning max profit daily This is similar to our current positive strategy except we want this to work race by race sequentially from the beginning of the day to the end. You will need a broad understanding of what types of models We have a database which predicts horseracing in Australia and we currently have a positive tipping strategy with a bank $ amount allocated to each bet. We need a script which works exactly like a human bettor/gambler. The system must start with an allocated bet per day and decide 1. which races and prospects are the good value or probability chances 2. The strategy for the day. This is similar to our current positive strategy except we want this to work race by race sequentially from the beginning of the day to the end. so the process would be 1. Bank is pre-determined 2. Script decides possible bets from our current calibration of bet 3. Script then works and suggests bet type and amount as the day progresses, not releasing a defined amount at the beginning of the day but instead acting like a bettor and adjusting future bets to try and capitalise the profit at the end of the day You will need a broad understanding of what types of models like LSTM and RNN or others which will suit this process and you will also need to be extremely versed in mathematics probability.