Sports Statistics 101: How to Spot a Bad Stat

Last week in this blog I said I was going to do an article on Sabermetrics ... well I lied, not intentionally but I realized that Sabermetrics is much less a 101 topic and much more Advanced Stats territory. Instead I thought up a better topic for a 101 blog and that is how to spot a "bad stat". The problem with keeping numbers for everything is that sometimes, numbers tell a story that isn't true. These are bad stats and I think the best example of this in sports is a pitchers Win-Loss record.

On the surface you may look at how many wins a pitcher has and go "well now how is that a bad stat? The job of the team is to win games and if a pitcher wins a lot then he is the best at his job". Baseball writers felt like this for years and looking at the winner of the Cy Young Award for the best pitcher in the Majors, that shows. For much of the 80s and 90s, winning 20 games as a pitcher was a prerequisite to winning the Cy Young. Many great pitchers had great seasons and lost the Cy Young Award. Guys like Dave Stieb, Kevin Appier, Phil Nekro and many others feel victim to the "you didn't win as many games as the other guy" argument despite their stats being much better than people who won the award (who I will not flame in this article because they are all great pitchers who do not observe hate). The Cy Young needed a hero and that hero was "King" Felix Hernandez. In 2010, Felix Hernandez was 13-12 with a 2.27 ERA and a 3.04 FIP, he was next level but David Price went 19-6 (with a 2.72 ERA and 3.42 FIP) that season and C.C Sabathia went 21-7 (with a 3.18 ERA and a 3.54 FIP), surely our boy Felix was going to lose out despite his stats lapping the field but then something miraculous happened, the writers ignored the win-loss record and went with Felix Hernandez. Since that point, stats have seemed to be the most important part of voting for the Cy Young award.

So why is pitching wins a bad stat? It's quite simple, in the season that Felix Hernandez was clearly the best pitcher in the league, the year where he completed 6 games, gave up an opponents batting average of .212 and had a better ERA than anyone in baseball, Felix won 13 games. How does that happen, simple, the Seattle Mariners that season SUCKED. The Mariners scored 513 runs that year, 70 less runs than the next closest team to them, had almost 100 less hits than anyone else, had the least home runs, worst SLG% and worst OPS in baseball. The Mariners offense was awful, so awful that they couldn't take advantage of dominant pitching. Despite giving up an average of just a little over 2 runs per game, that level of pitching was only enough to get the anemic Mariners offense 13 wins that year. Pitching wins are not a reflection of a pitchers ability but are more so a reflection of the teams ability. A lot of times these correlate because if a pitcher is really good and a team doesn't completely suck, that pitcher will probably win a lot of games. This is still correlation however not causation, ultimately no matter how great a pitcher is he will never be able to win games on his own.

And that is the difference between a good stat and a bad stat. A good stat is the kind of stat that a player has complete (or a large majority of) control over. These stats are the stats that tell you how good a player is in their position however there are many stats out there like pitcher wins that end up being less about the player in question and are more a measure of team success. Though team success is the ultimate goal of all players, one player alone cannot dictate team success, especially in baseball when your best players only hit 4-5 times a game and your best pitchers have no way of producing offense. So the next time you look at stats be sure to see if you can figure out if the stats you're seeing are describing how good the player is or how good the team they play on is.