Zig When They Zag

This post is about one element to choosing a set of statistical projections to base your fantasy baseball player valuations on. Obviously if you have reason to believe that you know which set of projections are the most accurate, that's the one you should use. But most people don't have any means of determining which of the dozens of projections on the market are the most accurate, and the only study I'm aware of that looked at this didn't find any overwhelmingly persuasive evidence in favor of one set or another.

So assuming that all the sets of projections are about equally accurate, you should use one that the other people in your league are NOT using. While your data won't be any more (or less) accurate than your opponents', you won't end up overbidding on the same players as everyone else, and the guys you identify as potential sleepers will tend to go cheaper than the 'sleepers' on everyone else's list.

Sometimes these 'tactical' suggestions are easiest to illustrate with simplified examples. Let's look at two players - Ted Lilly and Rich Hill. Imagine you're in a ten team league. You use one set of projections. The other nine people in your league all use the same set of projections as each other, but a different one from what you use. Let's say your projections show that Lilly is a $3 player and Hill is a $10 player. Your opponents' projections show the opposite - Hill is a $3 player and Lilly is a $10 player. Since we're assuming that both projection systems are equally accurate, Lilly and Hill are actually likely to perform equally over the course of the season. But in the draft, you're going to be able to get Hill on your team for $4 (a bargain according to your data) while someone else is going to have to pay at least $11, and likely much more for the same production from Lilly. Without having any more accuracy, your projections have helped you get the edge on your competition simply by being different than what everyone else was using.