And CASTIGLIONE has very similar numbers.
The PECOTA projections for the 2010 season arrived this week, as the god of the nerds, Nate Silver, unleashed a mighty flow of predictions upon his cult followers. I think PECOTA does a fine job of predicting individual players, but, if you think about it, the team records are based on the performance of these same projected players, which would be nice if the PECOOTERS functioned in a vacuum. Nevertheless, wacky managerial decisions (as just one example) have an undoubted effect on the actual outcomes of a game, meaning that the real baseball world is not a vacuum at all. As such, these projections can readily be thrown off by cold, hard reality. Therefore, I don’t really give much weight to these things; however, I do think they can be fun to look at in and of themselves. This year especially is proof positive of the sheer insanity that can happen when Silver starts poking around with numbers, and it begs the question of whether or not he is just messing with diehard baseball stats guys now.
You can view the complete projected records at Baseball Prospectus, where they are presented by division. Let’s just take a gander at a few of the things that make this year’s dose of fortune telling more ridiculous than Robert Downey Jr.‘s career arc:
* There are only three teams expected to eclipse 90 wins next season. Last season, PECOTA projected six. All three in 2010 (Rays, Red Sox and Yankees) are from the AL East.
* By PECOTA’s estimation, the Royals are going to have the lowest win total (66) next season. Unfortunately, this will be a year too late to grab Bryce Harper, which is just so fitting for this terrible, terrible franchise.
* Speaking of terrible teams, the Nationals are going to win 82 games. I will wager anybody that this doesn’t happen.
Reader Comments and Retorts
Go to end of page
Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.
1. Infinite Yost (Voxter)Facts are not important to people who write articles like this.
So I guess the Nationals are the new market inefficiency.
It'll be interesting to see the next few years. Personally, I wonder how much of PECOTA's prior success was in Silver's adjustments as opposed to Silver's baseline system.
The Yankees, Rays, Red Sox..etc. have built up the depth to survive the season with injuries. Teams like the A's, Nationals, etc. don't have the depth/resources to recover from injuries.
Really, they were never really that much better than any of the other systems to begin with, they just happen to have a cool looking presentation and a great hype machine.
And as someone who spends way more time trying to project team W-L than I probably should, I wouldn't kill them about their first few sets of team projections. Building depth charts and assigning playing time is a serious amount of effort and it's pretty much impossible to get right. So you do what you can, see what the feedback is from people who probably know more about the specific teams they follow than you do, and work on refining them as the season approaches. Any projections prior to mid-March are essentially meaningless IMO.
I guarantee that we'd be harder on Dan and Sean's corrections as well if they were charging us money to read their projections, while doing a PR blitz on how "scarily accurate" their projection systems are.
This is not to say they are going to win the division or compete for the wild card, but I bet a lot of Nats fans would be delighted with 82 or 73 wins.
Pretty sure that the runs scored issue is just rounding. As far as the AL wins not adding up to 1134, it looks like they're just saying the AL will go 1148-1120 and the NL will go 1282-1310, so it's probably just an AL/NL difference.
edit: Half Coke to GMoney.
Well, midyear trades generally can't improve a team by more than 2-3 wins at best, and that's essentially trading for an All-Star/MVP caliber player to replace a replacement level one. I'd think most mid year trades don't amount to more than a win or so.
As far as injuries, you just need to try and account for players who are poor health risks not playing as much and figure you're going to miss some teams badly no matter what you do.
Seriously. Let's see, nobody in the N.L. is going to win more than 86 games, and nobody in the A.L. is going to win more than 93. The Nats are going to go .500 (I wish), and the Angels are going to go... 77-85??
My friendly advice to anyone out there thinking of wagering money: throw these projections in the nearest garbage can, and don't look back.
In other words, don't get your panties in a bunch over the specificity.
Fixed that for you.
It does seem like these are version 0.1 of the PECOTA projections, so hopefully the final projections will be more useful. It's a bit of an embarrassing product for a pay site at this point.
No projection system will ever correctly predict the results of a baseball season. You're using PECOTA wrong if you're expecting it to mimic Nostradamus.
PECOTA is like a kid out there?
Everybody suspected as much.
Speaking of column totals not adding up properly and such, I always have to chuckle at like Danny Sheridan (who couldn't pick his nose) when he gives preseason odds, and the odds (converted to percentages) will often add up to like 200%...
Scarily accurate but not at all precise!
I do want to note that the "raw" player forecasts are not being affected by these issues - the projections in the book should, in fact, be fine. These are not in fact "PECOTA issues" per se, although since the depth charts is the only way we're presenting the PECOTAs right now (the spreadsheet is a reflection of the depth-chart adjusted forecasts) makes it impossible for anyone to know this.
But a good, conservative forecast of the spread of wins in the league is always going to be smaller than the actual spread in the league, because of the effects of luck/chance/whatever you want to call it. That's actually one of the big problems with last year's projected standings - the fact that we have a tight spread of projected wins means we've fixed things, not that things are broken.
So the pitcher projections in the book don't consider team defense?
What a difference a year makes...
That seems about right.
Mediocre relief in 2010 could be worth 8-10 more wins right there.
Sure, absolutely. I guess my next question is: What is the way to use PECOTA correctly? Is it just a fun tool designed to elicit arguments and discussion?
Lemme explain here. The BABIPs you're seeing right now aren't the BABIPs that PECOTA is spitting out. They're adjusted based upon the projected lineup in the Depth Charts and the projected defense for each. It's that depth chart based adjustment that's cauing the problems, not the projections that we're feeding in.
Eh, I could be wrong, but it doesn't FEEL that different. I'm sitting here looking at PECOTA projections that don't "feel" right, noodling with them on Excel and posting on Primer about why they are the way they are.
Use it however you want, but you're only going to make yourself angry if you're going in with the expectation that it's Jesus. Use it to place wagers in Vegas; use it to try and win your fantasy baseball league; regard it in the same way as Buster Olney's picks. Just remember that the more you expect from it the less you'll get from it.
It's like when you get a new LCD TV, and you turn it on. You're like, "Well, it'll only be good when I get HD TV." So you get HD. Then you're like, "Well, I don't have all the premium channels." So you get the premium channels. You keep moving the sticks further and further away and you can never be happy with what you have. Never mind that an LCD TV is a freaking amazing invention that required some extremely smart minds to create.
To be honest, I rarely use PECOTA. I just know that I couldn't come close to its accuracy if my life depended on it. And anyone this side of Tom Tango couldn't either on a year-in, year-out basis. No, PECOTA will not tell you when Jeter's herpes will clear up or when Mo Vaughn returns to 350 pounds. But pound-for-pound it's as good and in most cases better as anything out there. Sometimes you just need to step back and appreciate that. And this is Mr. #### the Status Quo talking here.
Now it is a fun tool like any of the other good projection systems.
I'm not the gambling type, but I believe that's called the vig.
If you have six teams projected to win 86 games each, one of them will win 92 games by dumb luck.
We have an expectation that somehow we'll be able to predict the outcome of sporting events, and we're puzzled every year when unforeseen things happen. I'm sure the BP people would be the first to admit they will get a ton of things wrong. It seems like the author is just figuring all this out now.
A good point. The liberal use of the phrase is ruining its original meaning, which is an important logical concept.
How do projection systems account for competition and schedule? I mean that in regarding single players and teams. How does a pitcher's projection change if he is in the AL East or the NL West? How does the projection of the Tampa Bay Rays change because their hitters and pitchers will have to face the Red Sox and Yankees for 36 games combined.
See, I don't have that expectation at all. I expect that no one will be able to predict how the coming baseball season will shake out with any great degree of precision. I mean, God bless them for trying to establish depth charts for all the teams and all that kind of stuff, but I doubt that would make their predictions much more accurate than sitting down three random Primates with bb-ref and telling them to take a few hours to hash out some 2010 final standings.
Crashburn suggests that you can take PECOTA and try to make money in Vegas with it; has anyone compared their predicted standings with the preseason over/unders? I'd be very surprised if there was any money to be made there. If I have the time, I'll try to look into that myself.
Then what exactly is BP trying to sell me with their "scarily accurate" projection system? Hint: if you're going to go around feeding unrealistic expectations, you don't get to whine about people having unrealistic expectations.
We have an expectation that somehow we'll be able to predict the outcome of sporting events...
Whaddya mean, "we"?
This is not to say they are going to win the division or compete for the wild card, but I bet a lot of Nats fans would be delighted with 82 or 73 wins.
Adam Dunn also impersonated an outfielder for half the season last year.
For 3 months, they had one of the worst defensive outfields I can remember.
I don't know if this is still true, but I seem to remember that the bulk of the AL-NL disparity over the last few years has come down to pitching. (Somebody . . . Tango? . . . I can't remember. Anyway, somebody put some research into it.) That might partly explain the discrepancy.
If it meant the "obvious" meaning, the phrase would be "begs for the question." Nobody "begs the food" or "begs the money."
Yes, as linked in post #4. And, no, there isn't much money to be made betting PECOTA across the board.
So what is the very obvious meaning? Is it to "avoid the question"? Or, is it to "raise the question"? Because both are used.
Not using "beg the question" both preserves the original meaning and avoids the potential ambiguity. Writers shouldn't try to use a phrase that they do not fully understand when there are phrases that expressly state the meaning they are trying to convey.
I think Dan's system produces reasonable results, but I'm never shocked or all that surprised at any of his projections. Gary Huckabay's Wilton system used to try to predict breakouts and so would have more variance in the projections; it made his system inherently less reliable overall, but also much more fun.
And mostly these systems are just for fun. I think any primate could scan a player's b-r log and come up with a projection that's no more or less reasonable than anyone else's.
The only time I'll really consult PECOTA or whatever is when we're discussing how surprising a player's year is, and then I'll look at PECOTA to see what a reasonable, objective system projected the player at.
Bull. "Raises the question..." (or "avoids the question" depending on what the writer really means) would facilitate fully understood communication. "Begs the question" in it's commonly misunderstood meanings, serves only to allow the writer to feel superior because he's using a fancier bit of phraseology.
You must be Registered and Logged In to post comments.
<< Back to main