ZiPS creator Dan Szymborski

Posted on February 15, 2010 by


Everyone has an opinion on how their favorite players are going to perform. For most of us, it is just instinct, hunch of fanboy optimism. That is why it is important to look at unbiased, systematic approaches to projections. One of the best sets available are the ZiPS projections from Baseball Think Factory Editor-in-Chief Dan Szymborski. Recently Szymborski, an Orioles fan, agreed to a Q&A about his projections and feelings regarding the Mets.

Can you explain in basic terms how the ZiPS projection system works?

DS: The simplest way I can describe it is: ZiPS uses the last 4 years to establish a baseline for a player from their performance and from predictive models for things like BABIP. It then applies a regression model and calculates probabilities for age-related changes based on large groups of similar players from MLB histories. ZiPS uses the recent past for a player and tries to find similar players at roughly the same age.

All Mets fans want to know about David Wright’s power and ZiPS gives him 24 HR. Talk specifically about this projection, some of the historical comparisons and how you feel it will play out during the year.

DS: Wright had one of the more unusual seasons for a star as dropoffs of that magnitude in home runs is quite unusual. The biggest culprit seems to be his HR/FB rate, which dropped to a pretty low 7%, less than half of his output in recent years. One could blame Citi for that, but Citi wasn’t a big enough park to turn Wright into a less powerful flyball hitter than Anderson Hernandez and Wright didn’t hit homers on the road either. A number of great players have had mid-career homerun dropoffs (Schmidt, Robinson, A-Rod) for various reasons and most recovered. Even if Wright sticks to a more line-drive based approach, he should still add some home runs back. The only issue is if there’s some Super Secret Wrist Injury that the Mets didn’t want to make public. I remember everyone, including myself, overrating Brad Wilkerson after the Soriano trade because it didn’t come out until a year later just how messed up Wilkerson’s shoulder was.

Note: You can read more about Szymborski’s views on Wright at this ESPN Insider feature.

ZiPS likes Jason Bay well enough in 2010 but then projects a sharp playing time decrease the following year and a sharp drop in production. What causes that type of projection from a guy who has been pretty consistent in four of the past five seasons?

DS: He’s not the type of player that ages well and already, more of his offensive value has been tied up in home runs. Fenway kind of masked this somewhat by hiding the decline in Bay’s non-homer hitting.
One-trick ponies don’t age well.

Note: This is really not news Mets fans should want to hear. I disagree that Bay is a one-trick pony and after watching Daniel Murphy and Gary Sheffield in LF last year, I think Mets fans will be pleasantly surprised with Bay’s defense.

I admit, you stumped me with Marty Keough, the first comp for Angel Pagan. After looking Keough up, I am still a little confused. How is he a comp for Pagan? Also, how would you use Pagan if you were managing the Mets?

DS: Some of the difference is the era. If you neutralize Keough to 2009 Shea in B-R’s Neutralized Batting option, Keough was a mid-late 20s, 330/430ish centerfielder, who was generally average in most aspects, good speed, doesn’t steal as many bases as you’d expect. ZiPS just has him the closest – for players that aren’t either notably bad or great, there are a lot of players who could be the “top comp” with just the slightest hair of difference in the parameters.

If I were managing the Mets, Pagan would platoon with Francoeur in rightfield and be the primary Plan B in center. There’s no situation or role in which I wouldn’t play Angel Pagan before Gary Matthews, but it remains to be seen if the Mets agree with me or not.

How bad does Luis Castillo’s defense have to be to negate a .368 OBP and a 78 percent SB success rate? Where would you put replacing Castillo on the team’s list of things to do before the season starts?

DS: Castillo would have to be pretty awful to really be a problem. He’s not standing in the way of any preferable that the Mets actually have on the roster and while it’s frustrating that he’s locked up and has no power, he’s not really that bad a player. If he can stay in the 6-12 below-average range (I have him at -8) then he’s not a problem at all, just a place that the Mets could *possibly* do better at some point.

Daniel Murphy had a .496 SLG in Double-A in 2008 and slugged .427 last year with the Mets. His projection of a .407 SLG seems really low, especially given his 525 AB. Can we wager a Coke on this one?

DS: As long as it’s an actual Coke and not the generic coke, which might include Pepsi, I’m game.

Note: I am with Dan on beverage choices. When you order a Coke in a restaurant and the server asks you if Pepsi is okay, the only acceptable response is: Only if fake money is okay when I pay.

Relatively speaking, Jeff Francoeur has been a feast or famine type of guy. But ZiPS gives him a .269/.310/.425 line, which is pretty much what he did in 2009, when it was famine with the Braves and (relative, again) feast with the Mets. ODDIBE gives him a 1% chance of being EX, 4% VG, 11% AV, 26% FR and 59% PO. Can you explain what qualifies at each level and why he doesn’t have a better chance of being “good” given his “success” in 2005, 2007 and his time with the Mets in 2009?

DS: Those percentiles are the historical levels of the starters. “EX” means, based on recent history, a level of performance that would put a player in the top 20% of starters offensively at the position. The other classifications are the other 4 quintiles. For a rightfielder, the boundaries are roughly, in terms of OPS+ at 130, 114, 105, and 95. So ZiPS gives Francoeur a 4% chance at having what amounts to an OPS+ of 130 or greater and a 59% chance at being under an OPS+ of 95. I don’t go by OPS+, but it’s an easy metric which is widely used.

No one knows what to expect from John Maine, Jon Niese, Mike Pelfrey and Oliver Perez. ZiPS expects all to be average or worse pitchers. Subjectively, who from this group has the best chance of posting an ERA under 4.00?

DS: Perez does. Perez also has the best chance, in my book, of posting some insanely bad Brad Pennington-esque ERA. Considering I have ridiculously bad acid reflex, I’m not sure I would want Perez on my team, but it would be wrong to pretend he still didn’t have some of the tools to be a solid pitcher, even if it’s not exactly the best bet.

Note: Once tabbed as a potential closer for the early 90s Orioles, Pennington had a lifetime ERA of 7.02 in 75.2 IP over parts of five seasons in the majors.

ZiPS projects a strong rebound for Francisco Rodriguez, with K/9, K/BB and HR/IP marks all better than his 2008 season. Yet his ERA would be the 2nd worst of his career. What are your thoughts on this projection?

DS: It’s a pitcher thing. Take any pitcher with an established great track record and going forward, you have to expect them to be worse than that track record. If you mapped pitcher careers as an actuarial table, they’d look more pessimistic than that of the life expectancy of an 85-year-old.

Note: My mom is 87 and my dad will turn 90 this year. My wife’s grandmother is 98. Getting to 85 is hard but once you get there it’s a lot easier to make 86 than it is for someone who just turned 12. Or something like that. I guess that’s why I am optimistic that a return of the above peripherals for Rodriguez will mean a solid year as the team’s closer.

Right now Doug Davis, Jon Garland and Joel Pineiro are free agent pitchers. Which one projects better for the next few years? Should the Mets choose one of these guys or go for a lottery ticket with Erik Bedard or Ben Sheets?

DS: ZiPS has generally like Jon Garland better than either Davis or Pineiro. It’s probably a little late for this question with all 3 suddenly signed, however. They should still go after Bedard or Sheets if they can – even the most risk averse team should take a chance on a special talent if it comes available. You don’t want to have an entire rotation of pitchers like Bedard or Sheets, but it’s always good to take a flier.

Note: I left this in even if it no longer really applies to the Mets with the specific players. Plus I like that someone else thinks Garland was the better choice than Pineiro.

Any final thoughts on the Mets’ organization?

DS: The Mets are one of those organizations that has the core talent to succeed, but they’ve had a great deal of trouble with the complementary talent. While a contending team wants to avoid too much risk, players like Livan Hernandez or Tim Redding or Brian Schneider aren’t good enough that you can ignore the lack of upside. When you design your team in such a manner, you’ll get very few pleasant surprises and when you lose players the caliber of Reyes or Beltran, you need some pleasant surprises or you’re going to lose a lot of ground.


I would like to thank Dan for taking the time out to answer these questions.

Posted in: Perspectives