Let’s say we are in the processing of testing ML strategy. We run the strategy automatically each day on horse racing win markets using Strategy Bot Executor tool. The tool reports profit/loss results for each market and total profit.
In the Output view we can see all messages the strategy reported, but is that really all we can use to analyze the strategy results?
By double clicking on the market in the Strategy Bot Executor view, we can reopen the market into the Bet Event view and see market prices snapshot from the moment when strategy was finished running on the market.
We can see placed bets, in the Market Bets view, profit or loss on the selection the strategy was executed on, or total profit / loss per market when our strategy executes bet placing on more market selections.
Clicking on the market context menu we can filter Output messages for the selected market by clicking on the menu item: Show Output Message.
What we can see in the Output view is time when strategy was executed and all messages, including data ML model used to chose dedicated selection, the action bot message, and race results, the winning horse was Elzaam Blue.
The strategy was exacted at 16:28:00 and it took 5 seconds to gather all required data ML model is using. The opening bet on Gold Wing was placed in 3 seconds later and because the strategy is using the trading bot the closing hedge bet was matched at 16:30:31.
That is all what we can read from output messages. This is actually not all, we have got quite a lot of other data we can analyzing as well.
Clicking on the market context menu we chose the menu: My Data Providers / Data Context Browser.
Depending on what data are produces and saved by your strategy you can find here a lot of data, by default it is added: Profit History, Last Traded Prices and Last Offered Back Prices in charts view.
The Profit History chart in my case shows that that action bot started to back/lay trading strategy, not in the right moment because the market went against my bet to maximal 10.8 loss bet position, but finally at 16:30:21 the market reverted and in another 9 seconds the bet position reach required profit of 3 ticks to hedge the bet in the profit of 2.54 Euro. Well yes, I was lucky even when the ML model chose the 3rd favourite, what we can see not only in the Bet Event view but also in additional data my strategy reports in the market data context My Selections Data.
When selecting the favourite for this race, Jaafel we can see that back/lay strategy would produce immediate profit as the Last Traded Prices went from 2.88 to 2.82, and at the same time when strategy closed bet position on my selection, the favourite was traded even better on 2.54.
Well yes, the favourite did not win, the winner Alzaam Blue went from 8.6 to 6.4, 57 seconds before the official race start time, then market prediction was not in favour for this horse.
What is your way to analyze your strategy results?