You know I love projections. I love projections as much as - or more than - most anybody else. Projections are how we get some idea of what we can expect from the season ahead. But one of the problems with talking about projections in February and March is that we have so little to discuss that we end up taking them a little too seriously.
Look at this
Now look at this, and this, and this
None of this is new. Everybody who's ever played with PECOTA or another, actually legitimate projection system for five minutes knows that projections can and do often come out looking ridiculous. This is just a reminder to take them like a salted pretzel. On more than one occasion last spring, I said that the Cardinals could probably win the NL Central even after benching Albert Pujols. Now I look retarded. That's not the Cardinals' fault, or the Reds' fault. It's mine.
Of course, one need look no further than the Mariners to understand just how poorly these things can go. I didn't want to bring that up. But here I am, bringing it up like a sucker.
Projections are great. We'll spend all offseason making them, and linking them, and discussing them. Then the games start and we get to spend all year seeing the things we didn't expect. It's a lovely little cycle we have.
Note now that I'm back from errands: I wrote this very quickly and, looking back, I wish that I would've gone about it differently. This isn't about criticizing the projections. It's about criticizing how we interpret them, and talk about them. The projections, for example, didn't get the NL Central "wrong". I was wrong for acting too confident about the averaged-out outcomes.
0 recs | 17 comments
I'll always love the projections vs. results for the Blue Jays.
ThundaPC - September 28, 2010
Speaking of benching Pujols
When is there going to be another podcast?
Alexander Calloway - September 28, 2010
I don't think of it as screwing up, am I wrong about that perspective?
This outcome was also a possibility, not precisely how the M’s arrived to this point I guess. But it was a preseason possibility.
Kermit. - September 28, 2010
What I mean is I screwed up by being as confident as I was
That was my error.
Jeff Sullivan - September 28, 2010
Ok, that helps. I was taking this in the wrong direction from the get go.
Kermit. - September 28, 2010
Hey, everyone, look, we do disagree
If someone had projected Jose Bautista to hit 52 home runs this year, THEY screwed up, not all the people who didn’t.
Projections are not predictions of the future. They are the sum of the knowledge you can have at the time. Things change in the future, but there’s no way of knowing what those things will be when the projections are made. Looking at what happened in a single year and deciding that projections that missed “screwed up” is why people don’t trust projections to begin with, and they should. They’re not cyrstal balls, and we shouldn’t be surprised when they don’t act like them.
davidcameron - September 28, 2010
I don't actually think we disagree that much
this post was more about the confidence we put in projections than the projections themselves. I wrote it up quickly and if I worded it poorly, that’s my mistake and I apologize.
Jeff Sullivan - September 28, 2010
Right, we shouldn't be all that confident about knowing the future
We do the best we can, and then we throw our hands in the air and say “ehh, who knows, right?” That’s kind of the fun of it. If everyone played to their projections, it wouldn’t be any fun.
davidcameron - September 28, 2010
This is only tangentially related
but to your memory, what team outperformed or underperformed its preseason projections to the greatest degree?
I guess the easy answer would be ‘2001 Mariners’, but, well.
Jeff Sullivan - September 28, 2010
Maybe we're just homers, but the M's have to dominate that category
I suppose it depends on whose projections you start with – USA Today, Fangraphs/ZIPS, PECOTA, a bunch of fans.
But how could the ‘01, ’08 and ’10 Mariners not all be in the top 5? What do you do with certain projections that get a lot closer to the mark, like CHONE for the ’10 M’s?
I was going to say the 2009 Mets, but looking at it again, they just can’t compete with either ’01 or ’10.
marc w - September 28, 2010
CHONE wasn't very close
Sean had the M’s winning 79 games this year. He was a couple of wins more pessimistic than the rest of us, but most of that was about playing time allotments.
davidcameron - September 28, 2010
2010 Mariners, by quite a bit
The projection systems missed by 25 to 30 games. I think people knew that the 2001 team would be pretty decent, though obviously not win as many games as they did. No one thought this would be one of the worst teams in the league.
davidcameron - September 28, 2010
Should have taken the hint, Jr
Matthew - September 28, 2010
I'd love a look at a saber preview for 2001
My sense is that any 2001 projection would be off by ~25 games. Despite winning the ALDS, most questioned how the M’s would score runs that year. I can’t imagine any rigorous projection would have them at 90, and that’d be a gap of 26. MAYBE 87-88?
marc w - September 28, 2010
Average error for projection systems on the Mariners...
…was actually about 20 games (prorating out yesterday’s standings over 162 games). So not great, but a lot better than 30 games. Really not too far off the error for the Padres, which was about 17 games.
cwyers - September 29, 2010
Required Reading
Link
perfectstrat - September 28, 2010
I love how this post starts out by saying it's required reading.
yuniform - September 28, 2010
You must Login with your SB Nation account and be a member of Lookout Landing to post a comment.