Back to Basics in Machine Learning Football Betting
Nowadays advanced technology provides us with football predictions and other sport betting tips after the analysis of big data combined with the trends and the advice of the best tipsters from all over the world. Predominantly, sports clubs, managers, professional betting tipsters, free prediction site owners are paying particular attention to machine learning techniques in order to better understand and formulate strategies necessary for accurate predictions.
It is obvious that machine learning can support potentially transformative advances in a range of areas and the social and economic opportunities which follow are significant. As we have mentioned above, in betting, machine learning is helping to build better predictive algorithms to bookmakers, teams and professional punters and offering new insights into more accurate predictive models.
Recently, we had the pleasure of chatting with Ola Lidmark Eriksson, a founder of betbot.soccer. Also, we thought it would be interesting to ask him some questions about football betting.
Let’s find out how can you turn raw data into winning football predictions? How can you take advantage of both data sources and the Python programming language from free prediction site?
This short interview shows how Python developer built a free soccer prediction model that correctly predicts the outcome of the games. He also gives you some expert advice and free betting tips.
Python developer, System architect and Machine learning expert. Football analytics wannabe and data scientist. Founder of Football Analytics Sweden and creator of several machine learning and sports-related services. Expert on the TV-show Fotbollslabbet (The Football Lab)
What tools do you recommend for successful football betting?
- For me, being a developer I build my own tools. A simplification of how I make my picks can be found at betbot.soccer. And for me, as I see it building your own models is a key if you ever over time can be successful in beating the market.
- For data sources, I think http://www.football-data.co.uk/ and Betfair’s Exchange API are good starting points. For building models on top of those sources, I think using Python as a programming language is a good way forward. On GitHub, there is a good example of how to use the Betfair API: https://github.com/betfair/API-NG-sample-code/blob/master/python/ApiNgDemoJsonRpc-python3.py
- And I have personally published a small lib that can be used to easily get data from football data with python: https://github.com/olalidmark/football-data
What have you found most challenging about predicting the football matches on your free prediction site?
- I have written a lot of what my experiences are about predictions on Medium, starting with this article: https://chatbotnewsdaily.com/what-i-learned-about-big-data-and-machine-learning-from-trying-to-predict-football-matches-2f81d019bea0
- To sum it up I would say that making statistically based predictions with a correct error-margin is quite easy once you have access to a large good quality database. The real challenge is to find the anomalies – when trends are to break.
If you had one tip you could give to our readers about football betting in general, what would it be?
- If it sounds too good to be true – it probably is. Beating the market over time is really hard so for me, I’d say that any strategy that not is proven over at least 1000 picks really aren’t proven at all.
If there is one thing you could change about football what would it be?
- Change to measure the time to ball-in-play.
Who do you support? (Football club)
- UC Sampdoria