Using Bet Angel football ratings
You may have seen me post ratings from time to time and a debate raged on the forum about what exactly I was posting and what was it? So I thought I would take time to explain the rationale behind the ratings. You can find the ratings posted to this thread on a regular basis.
How did the ratings come around
Way back when I was trying to win the football pools I realised I needed to understand how to price a market. That was because it would enable me to figure out which matches were most likely to be draws. If I found a match where there were likely to be two goals and the teams were evenly matched then I knew the chance of a draw was much higher than other scenarios. So I set about working out how to price a match. The first step was to work out how many goals could be scored.
How many goals will be scored?
In the process of pricing up a match I’m interested in the number of goals I expect to see scored in a match. They are never an exact number, they will be 2.64 or 3.02 or something similar. What this is saying is if this match was played 100 times what would the average number of goals be per match. In the latter case just over 300 goals will be scored over 100 matches or an average of 3.02.
Football is highly variable and therefore you can never forecast an exact score or number of goals. Only an average number. You can say something like, no goals will happen 8% of the time, one goal or more x% of the time and so on. You then added that up to reach your top line number. So when you compare ratings to result the individual results may be somewhat out individually, but as a group and over the long term when will average out to be pretty accurate. Last season I ended up less than 0.1 goals out on average.
Where do they come from?
I managed to successfully work out how a match was priced back in the 80’s. In fact, putting up odds on football markets was one of the first things I did on Betfair.
Since then I’ve dived in and out of the pricing market to refine the model. I’ve religiously collected data every week on the main footy markets to do this. There are many people that can do this nowadays as it’s often an intellectual challenge that math students take up.
Personally, I’ve gathered data from 100,000’s of matches and can approach a rating from many angles. I’ll tend to take a number of metrics, create a price from that then compare that to previous matches to arrive at a final rating. There are many ways to do it. Ultimately many people more or less come to similar conclusions, but I still do them as I learn some new things all the time and find new angles, so it’s worth keeping my hand in.
Therefore you should expect the ratings mimic the odds in goals markets. I and others who price the markets will correct any large discrepancies and there isn’t much value before the off.
How should you use them?
The best way to look at the ratings is to judge clusters of matches that have certain characteristics. Obviously the lower scoring matches would suit strategies such as backing under 2.5 goals and trading out. Higher scoring matches will be less likely to be draws and throw off their own characteristics.
Because of the variability of footy it’s better to cluster matches together and take a general position on them. That way you even out some of the variability. So seeking strategies on two or three low scoring matches it’s much more likely to produce a positive result than just picking one. But that said the extreme’s of either will lead to a more likely result. Where they are most powerful is when you use them to spot a deviation to the rating.
Let’s say the market is forecasting 2.20 goals, quite a low scoring match. As the match gets underway, which you are watching or looking at the statistics, you notice that there have been a many more shots on target than should be expected from a match like this. Therefore you change position. Of course the opposite could be true, the match could be forecasting many goals but the tactics of the underdog are doing a good job of stifling activity.
If you look at the ratings from the champions league this week (see below) you can see how they often play out. The high scoring matches clustered towards the top and the lower scoring ones towards the bottom. Where there is a variation in this, you can usually see it exposed through the inplay statistics. If Mario Balotelli is playing then mark down the number of goals 😉
Summary
Ratings are a good way of assessing a generic strategy in the market. But always bear in mind individual results can be quite variable but tend to average out with time. Overall the market is fairly efficient before the start, but as the match gets underway you can quickly assess if the market is deviating from the judgement and profit from it, or change your original position.
Category: Football trading strategies, Trading strategies
