FAQs

1) Alert Moves

Chart only compatible with web browser

The line / odds advised in the alert aren't available by the time I look on my bookmaker (even if I'm quick)

  • The alerts advised line should always be available at the time of the alert. The model uses the bet365 line at the time to generate a prediction and then an alert if the odds/line is good enough

  • However, we do experiences phases where the line/odds can move very quickly after we send the alert (within 30 seconds). This is likely to be what you may be referring to if you have asked this FAQ

  • The reason for these moves could be due to volume and/or betting strategy from our subscribers after the alert is sent

  • Please see chart above for a breakdown of this issue. This chart is setup to auto update so all customers can have visibility on the issue

Could one of the subscribers be using a bot and spoiling the prices for the rest of us?

  • This seems unlikely as the patches of quick line moves come and go and roughly correlate with human working / waking hours. If it was a bot we would expect all lines to move instantly 24/7

How long will these quick line moves last? How can we be sure they won't last forever?

  • The chart below shows the cyclical nature of these moves. It is possible that when a certain subscriber(s) is causing line moves, they may also be on the bookmaker's radar, possibly resulting in closure of their account. If any subscriber's betting strategy results in regular line moves, it is unlikely to ever be a sustainable strategy

  • This is supported by the chart whereby the highest proportions of alerts backed in usually last 1 or maximum of 2 weeks

What is the criteria of an alert 'being backed in' on the chart

  • An alert is classified as being 'backed in' on the chart if the line moved between time of alert and kick off. As the alert gets sent 2 mins before kick off, both of the following scenarios would be included in the chart: a) line moves 20 seconds later and many customers could end up missing the edge and b) line moves 1 min 50 seconds later and most/all customers would have time to utilise the edge

  • Therefore the chart represents a 'worst case scenario' in terms of forecasting number of alerts a customer will on average have time to utilise the edge. In reality you will be able to utilise the edge on many more alerts than the chart shows, correlating with how quick you are able to act

  • This is often reflected in the combination of thumbs up and thumbs down reactions to each alert

2) Variance

I've experienced a downturn in my results. Is everything okay with the service?

  • Short answer - yes! We (the developers) have automated daily reports to check every part of the pipeline is functioning as expected. We are very proactive from our side and the model evolves and improves over time, as is the nature of machine learning. If anything goes 'wrong', rest assured we will be right on top of it

  • Please see 'backtest' tab for realistic expectations of the service. Over the course of each year, we expect to perform around the 6-7% ROI mark, slightly improving over time as the model improves with more data

  • Battle Gold service started out with 4 months of very high performance followed by a slower month in August at +3.9% ROI. This is all well within the bands of expected results (-1% to 11% over a month would be considered 'normal')

  • Sticking to advised lines and prices is the best way to reduce personal variance