With the proliferation of the Internet, we’re now able to gather massive amounts of data on just about every professional sport, whether it is player performance, coaching, in-game events, betting lines, officiating, or even the weather. From the heaps of available data, we’re able to extract useful information that we can use with the aim of predicting future outcomes. However, one of the most common mistakes that novice bettors make is making bold and assertive claims from small samples of data. In this article, we’ll take a look at the importance of sample size and why a large sample size is almost always preferable to a small sample size.
Small Sample Size
Chances are, you’ve probably heard a person cite a statistic and the other person rebutting with the claim that it doesn’t matter due to a small sample size. What a small sample size means should be fairly easy to understand. Let’s illustrate this with an example. Let’s assume that the Golden State Warriors have covered against the spread in nine of their last 10 games, for a 9-1 record, or a .900 winning percentage. Does this information tell us much about how the Warriors’ will be likely to perform in the future? Maybe a bit, but it doesn’t hold too much weight. However, if the Warriors covered against the spread in 900 of their last 1,000 games, this is much more indicative of what will likely happen in the future. Taking this a step further, if the Warriors covered against the spread in 9,000 of their last 10,000 games, you can see that this holds more weight than our previous assumption. As a result, the size of the sample and the total number of bets helps us determine the level of confidence that we should be placing in our systems. In other words, the larger the sample size, the more a system can be trusted.
What Sample Size Do We Need?
Now that we know large sample sizes are much more preferable than small sample sizes, the question we should now ask is: What sample size do we need? How big of a sample size is enough? By using fundamental mathematics, we can determine the appropriate sample size when considering whether a betting system is worthy of wagering. This revolves around the concept of statistical significance. For example, if you have a betting system with a winning percentage of .570 and 2,000 or more games within your sample size, you can say with a 95% confidence rate that the results are true. As a result, this means that your betting system will perform better than 55% over the long-run.
However, a betting system with a sample size of only 200 games would require a winning percentage of .600 in order to have confidence in the betting system that it’ll win at a 55% rate or better. While this analysis is purely meant as a guide for the reader when building betting systems, the general rule of thumb is that with more results (larger sample sizes), a winning betting system is more likely to succeed over the long-run. Unless you’re only playing for the short-term, it’s extremely important to gain sample sizes so you can say with confidence that your betting system will result in profits over the long-run.