Now is the time that a lot of managers, myself included, are starting to feel the first twinges of panic about our rosters. Just a sampling from my world: Kila Ka’aihue hasn’t hit since his big walkoff homer, and is starting to lose playing time to Wilson Betemit. Sean Rodriguez is still benched every other day despite the Longoria injury. MadBum is pitching like crap. The list goes on…
Here’s another, turning the clock back to last Sunday night: in my 20-team H2H points league (lwts-based), Lance Berkman is my starting 1B (actually, I’m playing him at UT to save his knees. Zing! That’s bad fantasy humor, but I’m incredibly amused by it). And Sunday, I was seriously thinking about benching him. Through his first eight games, Lance Berkman hit just .214/.290/.286. Meanwhile, Travis Hafner was sitting on my bench, and he was absolutely on fire, hitting .370/.433/.630 over the first week! Why not just swap out for the coming week?
So, I looked at projections. Here are the Hardball Times’ Oliver projections prior to the start of the season for the two hitters:
Oliver Preseason Projections:
Obviously, this favors Berkman by a mile. But that was pre-season. What about rest of season projections after that first week? Are they closer?
Oliver Projections through last Sunday’s games:
Umm…ok, not much difference. I guess it is only eight days. This is why I love updated projections, like Oliver at THT (updated Mondays, I believe), and ZiPS RoS here at FanGraphs–it helps me remember that eight games against years of performance really doesn’t mean much in terms of predicting the future! Even the first month of performance will cause only small changes in projection for most players–though catastrophic or spectacular performances will start to take their tolls come May or June.
But Hafner’s hot! Berkman’s ice cold! Just ride the streak, right?
Well…perhaps we’ll tell this story in more detail at another time, but the next time you start thinking about hot and cold streaks, go read The Book. In the chapter “When you’re hot, you’re hot,” Tom Tango and his co-authors almost literally bend over backwards trying to find some predictive quality to hot and cold streaks. They took some of the most unbelievable multi-game stretches they could find, and found zero–and I mean ZERO–ability to use hot/cold streaks to predict future short-term performance above and beyond the overall average projection for players.
Mostly, anyway…they did find a small effect for cold streaks, presumably because some players in terrible cold streaks are slumping because they have hidden injuries. Was that possible for Berkman? Sure: the guy is old, with bad knees, and someone who missed a big chunk of spring training with various small strains and pulls. Furthermore, perhaps Hafner really was healthy for the first time in years as was reported in the spring, and really was back in his old form once again! If so, those projections might have been severely underestimating Pronk, while Berkman’s projection may have been overestimating his injury-riddled body.
I stressed over this for a good hour. Ultimately, what it came down to was this: Berkman’s Cardinals had one more game this week than Hafner’s Indians. As we are a weekly league, that extra game might yield an extra start for Berkman that I can’t get with Hafner, which would give him ~five more PA’s to earn points. So I went with Berkman this week.
The results, through Thursday’s game, are the talk of fantasy:
Berkman: .471/.500/1.176, 4 HR in 4 games.
That’s a win for patience, folks. As I write this (Thursday night), Berkman is the #1 added player in Yahoo fantasy baseball, with 22k adds Thursday and 24k Wednesday. What a difference three games can make early in the season, eh?
Here’s the thing, though: in all likelihood, this insane streak by Berkman is already over. What we’ll probably see from Berkman the rest of the way is something pretty close to his most recent Oliver projection. Fantasy managers adding him right now are probably already too late to the party if they’re expecting elite production. And that’s the moral of the story here: be patient. Very few players make dramatic jumps in skill level from one year to the next, and a few weeks are far too short–and the sample sizes are far too small–to see any such changes in measureable performance anyway. Fantasy managers who spend their time rostering every “hot” player are usually just chasing noise. Will you get lucky sometimes and catch a Jose Bautista? Sure. But most of the time, you’re gaining nothing and may often be hurting your team by starting a “hot” Emilio Bonifacio.
(by the way, I still like Hafner to beat his projection this year…)
So, while I already knew that I shouldn’t, I’ve decided that I really need to not worry about slow starts. Instead, I’m just going to freak out about this Joe Mauer situation…
Justin is a lifelong Reds fan, and first played fantasy baseball on Prodigy with a 2400 baud modem. His favorite Excel function is the vlookup(). You can find him on twitter @jinazreds, even though he no longer lives in AZ.