Tuesday, April 14, 2009
Decision Making Based On Small Statistical Samples
Posted by Eriq Gardner at 9:03amIf your team isn’t doing so well in the standing so far, don’t fret. The season isn’t four percent up yet. Emilio Bonifacio will not finish the season as the game’s most valuable player. Take that to the bank. Conversely, Jimmy Rollins won’t finish the season with no steals and an average that would make Mario Mendoza blush.
As any statistics guru will tell you, small sample sets are the enemy of accuracy.
In pursuing trades and hitting the waiver wire, fantasy baseball managers may think they are making rational decisions based on available data and recycled scouting reports, but rationality is often applied on a rather arbitrary basis.
Here’s a couple of examples:
After hitting three home runs, knocking in 12 RBIs and stroking 12 hits in his first 30 at-bats of the season, Adam Lind’s ownership has risen from 26 percent to 81 percent in the course of a week in CBS Sportsline leagues. Adam Lind has all the makings of a “post-hype sleeper,” that term given to touted prospects who will only live up to their potential after we’ve forgotten about them — but so too does Kosuke Fukudome, who came into the league with just as much hype and is neck-and-neck with Lind as one of the most valuable players of the first week. Fukudome’s early success has only netted him a modest 41% ownership (up from 23), begging the question why fantasy baseball managers seem to strongly prefer Lind over Fukudome at the moment.
Anybody who owns Jason Motte is unlikely to be doing well in the category of ERA at the moment. Fantasy owners cursed Motte’s name after blowing a three-run lead in his first save opportunity of the year. Many have already dropped him, and even some intelligent fantasy gurus have proclaimed Motte’s era as closer a disaster and all but finished on the basis of one week of play! Call me a Motte apologist, but three strikeouts-to-no-walks in 2.1 innings following a fantastic spring training still makes him the best candidate to be the Cardinals closer. (Yes, this includes Chris Perez, who has a lot of hype but also has his fair share of control problems.)
Looking back at average draft position before the 2008 season, there were certainly high draft choices such as Carl Crawford, David Ortiz, and Erik Bedard who each turned from fantasy prince to fantasy frog last year. But the number of busts is modest compared to the number of players who came close to delivering their draft investment.
Right now is the time to be zigging while others are zagging, to take advantage of any big changes in market value. A week of baseball shouldn’t change the values we put on players. Making decisions on the basis of small sample sets is foolish. Making decisions that take advantage of the shifting perception of others can pay off.
Most teams in a fantasy league are destined to finish outside of first place. That’s a statistical fact. Transactions can certainly give teams a better chance at winning, but in the early going, the art of player assessment is overrated. Instead, focus on honestly evaluating your roster and leveraging the rash decision-making of others.
Eriq Gardner is a New York-based writer and founder of Fantasy Ball Junkie, a website for advanced fantasy baseball enthusiasts.
I agree about the use of small samples to make judgements about how good a player is really going to be. However, sometimes these decisions are highly rational even though we have little reason to expect the player to continue on the path. The reason is that some fantasy teams can often pick a guy up at no cost; would you ever say “no” to a free lottery ticket? Of course not.
The first guy to decide to take a flier on Ryan Ludwick last season got bank; if he dropped Adam Dunn to do it, it was a dumb decision. If he dropped a bench player to do it, it was a good one.