To date we have reported on the ROI of the trackers as follows:

1. Bet Tracker – all bets +EV at the time of entry are recorded.
2. Golf Tracker – 10-15% of the field is selected to mirror a reasonable member strategy.
3. Horse Tracker – all +EV horses are recorded from 6 bookmakers at set times
4. DDHH – all +EV bets after team news are recorded

These bets are straightforward to report because they have “back odds” which we can benchmark an ROI to. We have had difficulty reporting the ROI of tools because Tools only report “lines”, they don’t report back odds. We have a database with every closing line we’ve published for our tools. This is a dataset of hundreds of thousands of Closing Lines, covering:

• Match xG tool
• Game Centre (all goals lines)
• Player xG (AGS, FGS, 2+, 3+)
• Detailed games tool (corners + cards by team and half)
• Player Stats tool (Shots on Target, Assists, Shots, Passes, Player Cards)

This database is huge. Historically we have benchmarked some of these tools against bets put up on the bet tracker, but the sample size is low and it takes a painfully long time to aggregate enough data for analysis. We want to make this dataset available to users. Before it is available we are going to map all of the results against it. This is a sizeable project; we have tens of thousands of football players in the player xG tool with lines for FGS. Some started the game, some played no part and some came on after the first goal and were voided – and we have to go through every player and work out which category they belong to.

Once the results have been allocated we will release a data archive tool so that users can see how the different lines are performing, how accurate/inaccurate they are and the standard deviation against the norm. All lines won’t show 100% accuracy – for example, the odds of Newcastle being 5-0 up against Tottenham after 21 minutes was 66,000/1 according to the game centre. Were you to look at every team in their odds range to be 5-0 up after 21 minutes then the occurrence (1) has far exceeded the likelihood. However with larger datasets we should see the expectancies vs historical occurence to converge.

We are open to feedback via Support on features that the UX of this Data Archive tool should display. At the moment we want to apply Filters for:

• Date Range 
• Pre-match fav 1×2 odds (e.g. 1.3 – 1.7) 
• Betting Line (e.g. Win Both Halves)

The UX will display

• Win Occurrence / Lose Occurency
• Win Expectancy / Lose Expectancy
• Standard deviation
• p-value

On top of this we have some data on the exchange (for example FGS, AGS, 2+ and 3+ are recorded for all games) and now we have enough data we aim to have an ROI calculator to assess various strategies. The timeframe for this project is 2-6 months, and it may be released in stages.