The Dirty Projectors

| | Comments (18)
billjames.jpg

The week's version of baseball limbo is upon us. The winter meetings don't begin for a few days, the free agent pool is a tepid puddle of middle infield mediocrity. What can possibly fill the precious baseball moments of our day and fuel the empty speculation of trade rumours? Why projections of course! The cherry pickers delight.

If you're new to the geek game, you may not be familiar with the assorted projection models and their banal acronyms. CHONE, Marcel, ZiPS, PECOTA, and Bill James are the best known, each with their own faults and strong points. Based on past performances, linear weighting, age regression and magic 8 balls; these systems spit out incredible reams of information that can be taken or left at your convenience. Sometimes the results are bang on, sometimes they're laughably off. Sometimes they make little or no sense, as the playing time they predict simply won't or can't happen.

Projections can serve as guides or indicators of potential future success. They are helpful in attempting to forecast how worsening the offense of your favorite team might be. Pitcher projectors are more tricky, as pitchers are fickle flakes whose flights of fancy cause frequent fluctuations in their FIPs and Fly Ball rates. But in the wrong hands, projections are dangerous. Reading too much into any given projection starts the average diehard down a dangerous line of thinking:

No team in their right mind would trade Stud Corner Outfielder X for 2 years of Proven Starter Y! Look at his Bill James Branded Projection brought to your by BIS, BATS, and The Boston Red Sox Baseball Club! SCOX is a sure thing! SURE THING! I don't care if we're getting Sandy Koufax in his prime, my team benefits from cheap production and I have new face on which to hang all my hopes and dreams! Gimmie the youth! Gimmie the youth!

You may ask "Drew, what's the difference between these systems and the community projections that seem to be in vogue these days?" At which point I will tell you community projections are a fancy way of saying "homertastic orgy of hopefully guesses, offset by spiteful lowballing and simmering self-hatred." As stated above, the bulk of these systems are based on elaborate algorithms and other things I don't understand. That doesn't mean they're better, it just means they're different. And better. But worse. Come o'er the jump for the projector breakdown.

Bill James - The most famous sabrmetrician of all* offers his self-branded system that is heavy on the offense, low on the sanity. Available in his annual handbook and freely on Fangraphs, Bill James juiced numbers make him the Larry Flint of prospect porn. Brand Bill never met a prospect they didn't like. Rake in the hilariously hitter-friendly Pacific Coast League? You're good for 20 Petco dingers! Knock a few line drives that didn't touch green? Let's just go ahead and boost your on base percentage by 50 points.

* - There is no such thing as a famous statistician. If you asked every person in an average, sold out baseball stadium who Bill James is, you would get a lot of blank stares. If you told them he was a tweedy, bearded eccentric that changed the way people look at the game, more than half would punch your face and demand you bring them a tallboy in repayment of their wasted time.

LOLine for 2010: Chris Davis. Davis had a tough time in Texas last year, striking out at record rates before being demoted mid-summer. But Davis had a solid cameo in 2008 and possess enough pop to get another chance. Bill James is bullish on Davis, projecting him for a 150 point OPS boost.

CHONE - Created by Angels blogger Sean Smith of Angels All The Way, this system is named after popular Haloman Chone Figgans. With Figgans testing free agency, Smith should adapt a new moniker for his system. I recommend SCIOSCIA. The projections will vastly overrate the contribution of every player, especially those with scrapitude or gritasticty. Any questioning of the SCIOSCIA system will be met by instant derision and a long, cold, befuddled stare.

LOLine for 2010: Ichiro. I'm all for realism and get why one would be conservative when projecting a 36 year old player. But this is ICHIRO we're talking about. CHONE only has him down for 188 hits. Uh, look CHONE. Certain things happen during a baseball season. The Mets disappoint, the Twins overachieve, and Ichiro gets 200 hits. It's really that simple.

ZiPS - Created by DaveDAN Szymborski of Baseball Think Factory, the ZiPS projections are quite elaborate, including defense and ODDIBE, or Odds of Important Baseball Events. Szymborski is well respected throughout the SABR world, and his system is thought to be pretty sound.

LOLine for 2010: Jeremy Guthrie. According to ZiPS, Jeremy Guthrie will only surrender 29 home runs this year. That's less than thirty! Jeremy Guthrie is a lot of things; solid pitcher is not among them. Guthrie gave up three home runs pouring milk on his cereal this morning. Waiting in line at the DMZ last week he served up two gopher balls and later allowed a walkoff tot to his massage therapist.

Marcel - Created by world-renowned King of The Nerds Tom Tango, this system is named after the pet monkey from early seasons of Friends; as in Marcel is so easy "a monkey could do it." Look at that, a dated sitcom reference AND a dismissive shot at the plebes unable to compute this elaborate math on their wristwatch. Respected or otherwise, we're moving on.

LOLine for 2010: Courtney Cox. Cox's mildly disastrous CougarTown is projected to finish this season but may fail to be extended into next year. Turns out shows "mainly about older women sleeping with younger men" might have a hard time finding an audience. Or being remotely watchable.

PECOTA - The grandaddy of them all, available on Baseball Prospectus for a small subscription fee. An incredible variety of information is provided, with age curves and future probabilities mapped out as far as 6 years into the future. Why bother even playing the games? Creator Nate Silver bases the projections on career trajectories of similar players and phone poll results conducted by Fox News, CNN, and John Kruk's autodialer.

LOLine from 2009: Raul Ibanez. The Ibanez signing left many heads scratching last offseason. An aging player moving to a new league had some (not all) wondering aloud if Ruben Amaro was crazy. Ibanez posted on incredible 34 home runs with a steller .899 OPS. What did PECOTA think before the season? 9 home runs. Nine. Though the robot army did nail the potential OPS (.892 predicted), Silver and friends only estimated 198 plate appearances. No word why, though Ken Rosenthal did give PECOTA a public lashing in Visual Basic for insinuating Ibanez's steroid-addled body would break down. For shame PECOTA, for shame.

Update! In my haste to have a little fun at the expense of the machines that'll soon control our lives, I overlooked a key detail in PECTOA's projection: it was a "remainder of season" outlook from August, not a pre-season projection. I thank everyone who pointed it out and hope it didn't take too much time away from trolling "Best Of The Decade" lists demanding Firefly's inclusion. Thanks!

Image courtesy of The Seattle Times


PREVIOUS: People Just Like Saying "Kenesaw Mountain Landis"   |   NEXT: Billy Begrudges Boston Beef, Begins Braves Bullpen Bidness

18 Comments

There is only one projection system that matters, and that is the OMGReds bobblehead projection system, or OMGBOBBLE as I like to call it.

Talking about baseball and the human condition through hipster puns

Ichiro sucks. Nice ground ball to short you managed to beat out again, asshole.

Full disclosure: I couldn't tell you which ipod commercial features a Dirty Projectors song.

Slighty off-topic, there is a adult dating site out there that caters to cougars and the dudes who want to have aging cougar claws sunk into their naughty bits.

I can only imagine the recent Grady Sizemore kerfuffle has erupted said website into all sorts of hot flashes.

I watched an episode of Cougar Town; it was atrocious.

Joey was way better.

BTW, I refuse to pay money to BP until they come out with the "Calculate My Rec League Softball PECOTA Comparables" app.

Ichiro! is impossible to project by formula, because there hasn't been anybody remotely like him in the last 90 years or so. PECOTA's 2009 weighted mean projection for him was .291/.337/.358 with something like 132 hits. Nate Silver readily admits that they just can never get that one right.

Thanks, Nate Silver! PECOTA helped me win finish in second place in the Walkoff Walk fantasy league.

They correctly projected a great 2009 from Jayson Werth and Russell Branyan, while wisely advising me to steer clear of Clinton Portis because of his injury issues.

Other difficulties in projecting Ichiro: he's an alien powered by Farthammer's irrepressible hatred.

I only trust the BONILLAs.

I knew I should of listened to Nate Silver. John McCain was a complete failure on the Gashouse Anthem Gang.

These projection models need to get off my lawn.

Fact: the throw to third that got Terrence Long that is shown on every single Ichiro highlight is not impressive. Ichiro was shallow and Terrence Long is slow.

Fact: Rick Ankiel does throw like a gay Tri-Lambda. He will be an asset prowing right field in Great American Ballpark.

I saw about 90 seconds of Cougar Town at some point, and while nothing therein dissuaded me from the conventional thinking that the show is awful, there was no denying that Courtney Cox looked fantastic.

Fact: Ichiro's throw to third that got Terrence Long was impressive because it was a hit-and-run play. Even if Ichiro was playing shallow, it took quite a throw to nail a runner with decent speed and a head start by such a wide margin.

Leave a comment