The Danger of Relying on Small Sample Sizes


If you’ve spent any time on the internet in the past 24 hours or so, you have probably seen various baseball writers and fans making hilarious* jokes about the Astros going 162-0, or Bryce Harper hitting 324 home runs.

*At least they were hilarious, until everyone with a keyboard had repeated the same thing, thinking they were the first person to come up with such a clever line.

Mandatory Credit: Kim Klement-USA TODAY Sports

With one game in the books, we know that the Royals will not be finishing the season 162-0. We also know James Shields won’t be going the whole year without giving up a run. And, we know that Jeff Francoeur will not play the entire season without getting a hit. Of course, you already knew each of those things was not going to happen (although the third scenario may have seemed more plausible than the others). As you are well aware, trends in early April don’t necessarily result in season-long trends.

Every year, there are players that get off to insane starts. Whether that insanity is of the positive or negative variety, we’re bound to see stat lines that are at an extreme. After the first 8 games in 2012, Alex Gordon was hitting .133/.235/.233. Fans and media alike were starting to wonder if 2011 was nothing more than a fluke, and if Gordon had started to regress to his disappointing numbers from earlier in his career.

Gordon went on to finish the season with a slash line of .294/.368/.455.

In Eric Hosmer’s first 18 games of the 2012 season, he had an OPS of .786 with 5 home runs. While the team record was terrible after the 12 game losing streak, some fans may have been thinking that Hosmer was on his way to another solid season following his impressive rookie campaign.

Hosmer went on to finish the season with an OPS of .663 and only 14 home runs.

Despite the numerous examples like the two above, there are still fans and media members who get worked into a tizzy over the smallest of sample sizes. A slow first week can lead to a player being labeled a bust or a disappointment, while a hot start can result in articles proclaiming him to be the league MVP favorite.

Obviously, expectations are factored into this as well. When a player performs at or near their expected performance level early on, it can be safe to predict that their numbers will continue. It’s also easier to project if the player has been in the league for a longer time. If an established player like Chris Getz starts the season with a .375/.450/.500 line, we should expect a serious regression to something closer to his career numbers. If Billy Butler goes hitless in 8 of his first 10 games, we should expect him to eventually get it turned around and start hitting at his ability level. These players have a recognizable track record, so a start that deviates too much from their career averages shouldn’t be expected to continue. Breakout seasons like Toronto’s Jose Bautista had in 2010 are very rare.

However, with the younger Royals hitters, we don’t have an established pattern of success, or lack thereof, to compare against. Hosmer has had one good season and one bad season, so anything is possible for 2013. Similarly, Mike Moustakas showed flashes of dominance in the first half of last year before fading down the stretch. Because of this, some people may be tempted to jump to conclusions based upon how well the young guys come out of the gate.

I’m begging you: Don’t fall into the trap. DON’T DO IT.

Setting ridiculously high expectations after a great first week is only setting yourself up for disappointment. Likewise, seeing one bad week and expecting utter failure from a player, particularly a player like Moose or Hosmer, is foolish. These players are going to have to make adjustments to pitchers, and then make adjustments to the pitchers’ adjustments. Give them time to get comfortable, and give the league time to get comfortable with them.

Besides, changing your outlook on a player’s season after just a handful of games is pure craziness. It’s like buying a house, then trying to sell immediately after a light switch stops working. The house might have some electrical issues, but shouldn’t you at least try and fix the switch first?

It is possible that a player’s poor start will be followed by an equally poor 5 months. It’s also possible that a great start will result in a great 2013 season. But with the volatility of early season statistics, it’s usually not wise to make any sweeping declarations about a young player after a week or two. You may end up being right, but the best course of action is waiting until sufficient data is available prior to evaluating a player.