The BeatTheBookie App is a web-based data application that grants me access to predictions generated by various models that have proven their performance over time on my blog. However, determining which model to utilize for a specific match can be challenging. Thankfully, the model performance dashboard offers the solution to this dilemma. By exploring the dashboard, I gain valuable insights into the theoretical performance of the different models in past scenarios. This information aids me in selecting the most suitable model for each match, increasing the chances of making informed betting decisions.
How to identify a good model?
For backtesting the different models, a simple 1-unit flat staking strategy is employed. When a predicted model’s odds offer value compared to the fair Bet365 odds, a theoretical bet is placed. The average profit and the number of bets provide an estimation of the profit generated by a model over a specific period of time.
The fair odds are utilized as they do not incorporate any margin. It is important to consider that a profitable model should yield a higher average profit than the average profit of the chosen bookmaker.
With the diverse filtering options available, the dashboard equips you with the tools to make well-informed decisions based on the performance history and predictive capabilities of the available models. By leveraging these features, you can assess the profitability and reliability of each model, enabling you to optimize your betting strategy.
The following filters are available on the dashboard:
You can assess the profitability of a model across different divisions or a specific group of divisions. By selecting multiple divisions, the “Performance per Division” tab presents a breakdown of the profit achieved in each division, allowing you to evaluate the model’s performance in various contexts.
Evaluating whether a model is consistently profitable over an extended period or if its success is limited to a single season is crucial. To facilitate this analysis, the “Performance per Season” tab allows you to compare the performance of each model across different seasons. This feature enables you to identify any seasonal patterns or fluctuations in a model’s profitability.
The dashboard offers the flexibility to delve into the details of a single model or conduct a comprehensive comparison of multiple models. By selecting a specific model, you can explore its performance metrics, historical data, and other relevant insights. Alternatively, by comparing multiple models simultaneously, you gain a broader perspective on their relative strengths and weaknesses, aiding in the identification of the most reliable predictive tool.
If you have further questions, feel free to leave a comment or contact me @Mo_Nbg.