The Nats and PECOTA Projections

Oh poor Nats… they are just not as good as we all thought. At least that might be your first reaction if you look at the playoff odds report at Baseball Prospectus. They are projected to finish 87-75 (after rounding) despite winning 98 next year. And while they are projected to win the NL East, they have the lowest playoff percentage chance of any projected division leader at 67.9%, and the lowest chance of winning the World Series of any of them as well, at 7.6%. What gives?

Ok, before you go storm the offices of Baseball Prospectus (for which, I assume, you’d need some help from ENCOM) let’s keep a few things in mind. First of all, the Nats are forecast to win the NL East, not come in second. The Braves are forecast to be the second place team at only 82-80, which would give the Nats a comfortable lead. And all of this is based on PECOTA, which has some quirks that are worth noting. That doesn’t mean PECOTA is worth ignoring, its just important to know what the issues might be.

So let’s start with a few quotes from BP

PECOTA does not hate your favorite team. PECOTA is a collection of algorithms, written in computer code and run by an unfeeling machine

these are PECOTA’s estimates of what would happen if the season were played thousands of times. PECOTA is not infallible, and the season is played but once. It is a prediction, not destiny.

Alright, I got it. Programming, algorithms, computers, and predictions. Well, most of you reading this probably aren’t going to throw your hands up and scream “why don’t we not even play the season, NERD!” when you see this, so Spreadsheet those caveats are fine and good. But you’re probably still wondering why a 97 win team last year would be interprested as so “lucky” in 2012. They should be, after all, improved from last year with their roster moves (essential adding Haren and Soriano, and losing Morse, who didn’t have a good year last year).

Instead it is likely based on predictions of the players, so let’s get into some of the “bad” and why the computer might have doubts. I know, Colin Wyers, the computer doesn’t actually have doubts. Let’s start with the lineup, and where is may underestimate what all those pundits, who could see the Nats winning 100 games, are thinking:

  • It says Bryce Harper is going to hit .259/.324/.442. If I were you, I’d bet the over. This would be an incredible stat line for any 20 year old… human. But since Harper is superhuman, I’m guessing there’s nothing in the code to deal with that, and his age is hurting his predicted performance.
  • It’s only got Jayson Werth with 568 PAs, which is low. His injury last year and his poor performance hurts his recent track record, which is all PECOTA can base predictions on. He is still a strong player though (3.2 WARP), but he could be higher
  • Ian Desmond is going to regress. This is probably going to happen, but again, since PECOTA can only base things on a player’s history, it can’t just decide “oh 2012 was a breakout year” and forget what he did in 2010 and 2011. It has his OBP at .309, which wouldn’t be out of the question if he really did hit .269 (as they predict), but his projected SLG of .419 seems like more than just a regression, and again, I’d expect more, even if there is some regression from 2012.
  • It is perfectly reasonable to expect Danny Espinosa to hit .239/.305/.405 if you completely ignore his September injury and count his horrible April equally with the rest of his season, which is probably what a computer should do. I personally think his May 3-Sept 11 .269/.330/.451 is more indicative of what he can do. I think it’s more likely that he loses 2 months to shoulder surgery than has an OPS of .710.
  • PECOTA has Adam LaRoche hitting .251/.320/.446, which is significantly worse than last year. I don’t mind the lower AVG, or the power drop, but that OBP seems pretty damn low. He’s has a .320 OBP twice before, but never anything lower. So it’s predicting he’d have his worst OBP once again. This one, to me, is curious.
  • It does not expect spectacular bench contributions from this team, for which I can’t really fault it. Again, it goes back to track record.

On the pitching side, there are a few things that might make you think this computer is an unfeeling machine… that is out to get the Nats!

  • PECOTA sees significant regression from Gio Gonzalez, with a predicted ERA of 3.67, good enough for his worst since he’s been a full time starter. I’m guessing this has to do with his high walk total, and every year it underestimates his ERA. It makes sense, there just aren’t many starters who walk as many as Gio does and are as successful at preventing runs. Additional info: Ben Lindbergh noted in this article that some of his success in 2012 came from a great BABIP. To which I say “FEH!” – Gio had a better BABIP than the prior years, but not significantly so. I think this adds to his outlier status in that he gets guys out with alot of walks. It’s how the ball comes off that bat when he is pitching. At least that’s what I’m going with. 
  • Another guy who is forecast to regress significantly is Jordan Zimmermann. His 2.94 ERA is 2012 would be tough to repeat, but like Gio, his predicted 3.68 would be significantly worse than what’s he’s done recently. His career ERA is 3.47, and has been at 3.05 over the last two seasons, since returning from TJ surgery.
  • Ross Detwiler gets no love from this system. His ERA was an impressive 3.40 last year, but his FIP was only 4.09, and his WARP was a paltry 0.1. In other words, in Baseball Prospectus, Detwiler was a replacement level starter last year. So we shouldn’t be surprised that PECOTA sees his luck running out, giving him a whopping 4.71 ERA and an identical WARP of 0.1. There is certainly evidence to suggest that this is the proper assessment of the man, who only struck out 105 in 164 1/3 IP last year, while walking 52. A 2/1 K/BB coupled with a sub-6 K/9 isn’t a very good combination.
  • Ryan Mattheus has had an ERA of 2.84 in the majors, so it would seem like PECOTA is a little harsh to predict he will sit at 4.10 in 2013. But Mattheus’ two years in the majors have been at the ages of 27 and 28. While sustained success is certainly not unprecedented with someone that age who has a high level of talent and big injury history, it isn’t the norm. You would expect his predicted line to be pretty poor, even if he has turned a corner on health and success.

Now that I’ve explained away all of the PECOTA predictions, are the Nats really a team of outliers? Well, no, but it wouldn’t make LaRoche an outlier to age a little slower than the models suggest. Or for Desmond to have a breakout season in 2012, rather than just a lucky season. Gio and Harper could indeed be outliers, with the former able to walk more than most successful pitcher’s can and the latter just maturing with the bat faster than most hitters.

In all, it’s not outside of the realm of possibility that the Nats only win 87 games. Heck, they might only win 78. But most people are predicting 95 or more, and it’s because this team is so strong. The fact that they are mostly a young team, with a few guys who are just coming into their own, and a couple of key contributors that might buck statistical trends, means that while PECOTA is a handy guide, it is probably underestimating their performance.

One Response to The Nats and PECOTA Projections

  1. david juza says:

    The comments regarding gio are legit. I have done some analysis on pitchers with lower and higher babip. Guess who is the active career leader? Jared Weaver. Good but not otherworldly pedro esque starter. Gio has a career number below .300, no evidence to assume its going up. Maybe someday when his stuff declines but not at 28. Harper, toss the field comparisons out. Griffey, trout, mantle, cabrera, that is his year 2 peer group, not the field. Its like saying tigers odds of winning a tourney in 2003 are the same as steve strickers were because they both make the cut at similar percentages. Silly and highly unlikely.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

%d bloggers like this: