Stone Cold Retarded: ESPN's "Projection" System

I almost never read ESPN anymore. I won't spend too many words on why, it should be obvious. In short, their content is horrendous. Rob Neyer is the only baseball or football writer worth reading.

I was screwing around on the internet, avoiding cleaning myself this Sunday morning, and came across this article from Chad Finn of Boston.com. On the topic of the resurgence of Red Sox catcher Jason Varitek, Mr. Finn writes, "While he's apparently intent on having the last word with his skeptics -- hi there -- my willing suspension of disbelief calls it quits at his ESPN projections of 33 homers and 78 RBIs."

Finn's link is to Varitek's player page on ESPN.com. Under the 2009 Season Stats heading, there are four lines, "Season", "Career", "Last 7 Days", and "Projected". The first three are self-evident, but what does "Projected" mean, and how could they seriously project a 37 year old catcher to hit 33 homers?

When you think about it for a second, creating a player projection can take into account a million things including player aging patters, player type (i.e. speed, power, etc.), splits, and on and on and on. As you can see, getting numbers that even approximate what has and thus will happen on the field is very complicated.

There are a number of different projection systems available on the internet. Tom Tango who writes under the name TangoTiger and was recently hired by the Seattle Mariners created one called Marcel, SG at Replacement Level Yankees Blog created CAIRO, Baseball Prospectus has one by newly-renowned math whiz Nate Silver (also of the political projection blog fivethirtyeight.com) called PECOTA, Fan Graphs has one blah blah blah.

They are all different, but the one aspect that is the same within each is the complexity. Which brings me back to ESPN. After searching the site for an explanation of how they created their projections, I finally just looked at them and suddenly it was very obvious. Take a look at Varitek's numbers this year and his ESPN projected numbers for this year:

This year: .225 Average, .325 On-Base Percentage, .482 Slugging Percentage
Projection: .225 Average, .325 On-Base Percentage, .482 Slugging Percentage

Now take a look at Dustin Pedroia:

This year: .293 Average, .387 On-Base Percentage, .388 Slugging Percentage
Projection: .293 Average, .387 On-Base Percentage, .388 Slugging Percentage

You may be noticing a pattern here. Yes, ESPN's 'projection system' simply repeats the player's current stats. No accounting for playing time, injury, player history, splits, or really anything. This leads to some ridiculousness. For example, ESPN projects Raul Ibanez to hit 59 homers, 26 more than he's ever hit in his life. According to ESPN Jason Bartlett of the Rays will hit 23 homers, which doesn't sound like much until you realize he has hit 18 in his entire six year career. Tim Wakefield of the Red Sox will apparently win 21 games for Boston, and Kevin Slowey will win 23 for Minnesota.

None of these things are impossible, but all of them are extremely unlikely and together they're about as close to impossible as you can get without actually getting there.

So, what's my point? ESPN's projection system is sub-moronic. Good day.