This is once again, something I see all too often, people quoting scoring averages when assessing a particular match. Averages might be ok when analysing a particularly large sample size with scorelines of limited variance. But the problem is that even over a sample size of for example 15 matches, one irregular result can distort the average. Consider the following hypothetical example.Let’s say Chelsea have played 15 home matches with 45 total goals scored in those matches. That’s an average of 3 goals scored per match. The Over/Under 2.5 goals is paying even money. Looks tempting considering Chelsea’s home form in goal totals.
Now let’s say that two of those 15 matches were particularly high scoring, for example, a 4-2 and a 5-1 result, seeing a total of 12 combined goals scored. These two irregular results have severely distorted the average, which would be 2.53 in the other 13 matches.
What may have looked like an excellent opportunity to bet the Over 2.5 goals, suddenly seems less enticing.
So while averages can be helpful in large sample sizes, in general, it’s best to track occurrences. Look at how many times a team has gone over or under a particular goal total. In our Chelsea example, it could well be that while the average for the 15 home matches may be 3 goals per match, but that the actual number of times Chelsea home games have gone Over 2.5 goals could be less than 50% with a few high scoring matches giving an inaccurate and distorted impression. Perhaps Chelsea home games have only gone over 2.5 goals on 7 occasions.
Even if the average is 3 goals per match, would you feel good about betting the Over if the occurrence rate was so low? Probably not.