Bullish on Ryan Dempster

The 2011 season has been a high-priced train wreck for the Chicago Cubs. Despite having the sixth-highest opening day payroll among major league clubs, the North Siders own the second-worst record in the game at 27-40. The Cubs are highly unlikely to join the 2008 Seattle Mariners as baseball’s second team to rack up 100 losses with a payroll exceeding $100 million — that would require them to win at a .368 clip or less from here on out — but it figures to be a somber summer in Wrigleyville.

At first blush, Ryan Dempster looks like yet another well-paid underachiever. The 34-year-old righty, pulling down nearly $14 million, has a 5.48 ERA in 15 starts this season. Dempster is on less than half of ESPN players’ fantasy rosters (46 percent), and about two-thirds of Yahoo rosters (67 percent).

Is the starter-turned-reliever-turned-starter finished? Not hardly. The current version of Dempster might not be quite as dominant as the guy who averaged over four Wins Above Replacement per season from 2008-2010, but he’s still well worth owning in all leagues.

Dempster has the second-largest negative differential between his ERA and xFIP among qualified starting pitchers. His ERA sits perilously close to 5.50. But Dempster’s xFIP, which measures a pitcher’s performance based on strikeouts, walks and a normalized home run per fly ball rate and has been shown to be a more accurate predictor of future performance than ERA, is just 3.31. That’s actually in the top twenty among MLB starters.

While Dempster has struck out about 8.5 batters per nine innings and has issued 3.1 BB/9, figures in line with his previous work with the Cubs, he has been scorched by the Luck Dragons. The three “luck stats” over which pitchers exert less control — batting average on balls in play, left on base percentage and home run per fly ball rate — explain much of Dempster’s seeming decline.

Dempster has a .320 BABIP this season, compared to a .292 mark from 2008-2010 and the .288 major league average in 2011. He has stranded less than 67 percent of base runners, well under his 73-74% average over the previous three years and the 72-73% big league average this year. And batters have gone deep on 15.5% of fly balls hit against Dempster. His HR/FB rate was 10.1% from 2008-2010, and the MLB average in 2011 is slightly over eight percent.

As Mike Podhorzer noted in his “Bearish” post, Dempster isn’t fooling batters as much this year despite the high K rate. His 8.5% swinging strike rate is close to the MLB average and below his 11.2% average from 2008-2010.

So far, that hasn’t had much effect on Dempster’s ability to get K’s compared to past years, whether on a per-innings basis (8.2 K/9 from ’08 to ’10) or a per-batter faced basis (Dempster has struck out 21.7% of hitters faced this year, which is identical to his ’08 to ’10 rate). The reason is that the extra contact against Dempster has been mostly harmless. According to StatCorner, hitters have fouled off 19.5% of pitches that Dempster has thrown this season, compared to 16.1% from 2008-2010.

Whether Dempster can continue to get so many Ks with an average whiff rate is questionable — there’s a pretty strong relationship between strikeouts and swinging strikes, after all. Hitters are swinging at a lot of pitches that Dempster puts off the plate (32.5 outside swing percentage, 29.4% MLB average) and are making more contact on those outside swings (68.2%, 51.9% from ’08 to ’10) . But to this point, the result has just been more souvenirs for fans sitting down the base lines and, consequently, more strikes on the batter.

Dempster’s K rate will likely decline to an extent, and his BABIP could remain somewhat higher than average, considering that the Cubs have the third-worst team Ultimate Zone Rating in the majors. But even so, he’s a strong buy-low candidate. Grab him if he’s available on the waiver wire or try to work out a trade with a frustrated owner who may lump Dempster in with Chicago’s other eight-figure slackers.





A recent graduate of Duquesne University, David Golebiewski is a contributing writer for Fangraphs, The Pittsburgh Sports Report and Baseball Analytics. His work for Inside Edge Scouting Services has appeared on ESPN.com and Yahoo.com, and he was a fantasy baseball columnist for Rotoworld from 2009-2010. He recently contributed an article on Mike Stanton's slugging to The Hardball Times Annual 2012. Contact David at david.golebiewski@gmail.com and check out his work at Journalist For Hire.

9 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
JKB
12 years ago

A large negative differential between ERA and xFIP simply means that Dempster is a good pitcher on a bad team. You confirmed that with the Cub’s rotten UZR this year. You can add Garza to that list as well. The Cub’s defense is a far cry from the Ray’s defense, even though Pena is manning 1st base for them now.

Back to my point: Dempster is a good pitcher on a bad team. So, under what conditions should we expect his ERA to regress down towards his xFIP? I would guess two conditions: either the Cubs improve their defense and thus UZR, or Dempster gets traded to a better defensive team like the Rays. Bet Garza would love that.

Outside of those conditions, I don’t see a mechanism whereby ERA can regress down to xFIP, or any other fielding independent pitching metric. I actually have yet to see a study that provides evidence that ERA can regress to xFIP over the course of a season. The Hardball Times study you link to above is not a time series study, it shows that ERA on odd numbered days has a lower correlation to ERA on even numbered days than does xFIP, so it is more about estabilishing and comparing the internal consistency of the two metrics.

Do you know of any studies that have been able to provide evidence that ERA regresses to xFIP over the course of a season?

B N
12 years ago
Reply to  JKB

This. Though I would say that simply due to the greater variability in ERA, it will probably regress toward a more stable stat like FIP or xFIP (as the outliers in FIP and xFIP will be damped out quicker than the outliers in ERA).

Despite that, I don’t see why ERA should generally converge on FIP or xFIP. As you’ve stated, it’s not like pitchers with bad defense, bullpens, or parks are suddenly traded to a magic “league average” team where their performance will follow a semblance of DIPS. If that were true, we’d look at San Diego’s starters every year and exclaim: “Sheesh, look at these chumps playing over their heads! There’s no way their ERA will stay that low!” Instead, we recognize they have had good defenses, a huge park, and a great bullpen (which will all suppress their ERA below DIPS estimates).

So in my opinion, there’s really no need to run the study. We know that ERA for any given pitcher will not converge on its DIPS equivalent, regardless of the sample size, except for a subset of pitchers in the middle of your DIPS distribution (e.g. pitchers with league-average defenses, parks, and bullpen support or some league average combination of the three).

A much better question would be how we could incorporate all these factors to get a good non-DIPS projection of future pitching performance.