MLB Expert John Bollman - How to Avoid Fake Trends in 2022

Our MLB expert John Bollman is here to give you as much information as possible so you can bet with confidence during the 2022 season.

MLB and the union are starting to meet on a daily basis now, and I think we could see an agreement within the next two or three weeks. We might end up losing a week or two of the regular season, but it will be worth it in the long run to protect the players and the integrity of the game. For that reason, this will be my last analysis article before I get back to making pre-season plays. In this article, I don’t have anything that you might not already know. However, I want to get into misinformation, how to identify fake trends, and when to ignore small sample sizes.

There are tons of factors that go into handicapping a baseball game. Not every statistic is going to be on the side you like, especially for underdogs. For that reason, you have to know what information to consider and what information to ignore. When I was with the Blue Jays, I had trouble with this. I would always include as much information as possible for projects because I am able to comb through misinformation easily. However, I didn’t realize that by doing this I was actually providing more information to my bosses than they knew what to do with. And one day I asked one of my bosses “well, how do I know which statistics to give you then?” He said, “Give me everything I need, and nothing I don’t.” It’s such a simple answer but it makes sense. Give him all the essential statistics but leave out the secondary statistics. Those secondary statistics can cause your opinion to change even though they might not be as sturdy, and that is what I am going to explain.

Everybody knows your classic too many splits example. Well, that might be your most obvious example of using secondary statistics to sway your opinion. Instead of using an essential statistic like home record, you look at the home team during day games before 4pm when the Sun is in the East... you get the point. Adding too many splits just discredits the data. Every split you add decreases the quantity of data, therefore decreasing the quality. When the season is only 162 games, cutting the sample size in half can be extremely detrimental to a statistic. This causes misinformation. This misinformation causes you to sway your opinion on a game when it really shouldn’t.

Fake trends and small sample sizes are another form of misinformation. In a 162-game baseball season and especially during the first half of it, you only have so much data to go on. Let’s get into an example. Division rivals only play each other 19 times per season, so let’s say the Cubs and Cardinals are playing and we are halfway through the season. There have been 81 games played but only 9 games between the Cubs and the Cardinals. The Cubs have won 5 of the 9 games and the Cardinals are 4-2 at home, meaning the Cubs swept their one home series. Saying the Cubs have won 5 of the 9 games is not a trend. Saying the Cubs have won 3 of 3 games at home is. You might be thinking but one has a smaller sample size? Well, neither statistic is significant in terms of sample size, so it is up to us to find the one with a trend. “Significance” is a statistics term meaning the sample size is large enough to say that the true mean is represented by the data. The statistics definition of significance will rarely occur with the sample sizes in sports in general, so we have to be a little more lenient in finding trends. I won’t get into that anymore as to not confuse you, but with my statistics background it took a little getting used to to identify something like a 3-0 home record as a trend. Had the Cubs only gone 2-1 at home and the Cardinals 3-3 at home, there might not be a trend anywhere.

We have to be very careful of that, just because one team beats the other team one time more than the other doesn’t mean there is a trend. Same rule applies for a team that is 17-16 in their last 33 games or 3-2 in their last 5. We don’t know all the circumstances around those games and if one little thing went differently, it could be different. Maybe all 5 losses for the Cardinals were by just one run and the Cardinals didn’t get to pitch their two best starters in any of the 9 matchups. If that is the case, are the Cubs really better than the Cardinals this season or did they just happen to play them at the right time? Well, that’s where the real handicapping starts.

Now, with all the skills we have learned in the past couple articles we are finally ready to do some real handicapping. I am starting to see League Championship and World Series Championship lines with plenty of value, so we will either get into those or another division preview next week. Go back and re-read my previous articles on Overperforming/Evaluating teams and Explaining WAR, then we will get into some real handicapping using these same skills.