Mets Geek » Blog Archive » A Real Relief

May 26, 2006

A Real Relief

Scribbled by: Matt Gelb @ 8:00 am | Filed under: Articles

Early returns on the official buzzword of 2006? Bullpen.

Omar Minaya and Willie Randolph have prided themselves on a deep and successful relief corps. The reluctantness to moving Aaron Heilman into the starting rotation and the recent proliferation of arms demonstrate the Mets commitment to the bullpen. Following a middle-of-the-pack performance by the relievers in 2005, Minaya made upgrading the bullpen an obvious priority in the offseason. Thursday’s performance aside, the returns are overwhelmingly good. Measured by conventional rate statistics, the success is easily seen.

NYM RELIEVERS
ERA - 3.47 (3rd - MLB)
WHIP - 1.21 (3rd - MLB)
K/BB - 2.98 (2nd - MLB)
K/9 - 9.35 (1st - MLB)

These rate statistics may be one of the finer methods of measuring the success of pitching, but there are better ways to evaluate the effectiveness of a bullpen. A reliever’s performance is one thing, but his contribution in high leverage situations, specifically when his team is ahead, is paramount to evaluating effectiveness. One such way is using Win Probability Added, a statistic I used often in charting game situations and individual contributions to the percentage of winning a single game.

Until recently, WPA was not a statistic kept (free of charge). The great people at Fan Graphs have been posting game charts on a daily basis beginning this season. Not only that, but Fan Graphs also keeps a season tally of WPA by team. By default, in crucial game situations, an out is worth more WPA than an out recorded by a reliever in the seventh inning of a 10-2 rout. Both lower the ERA and both can boost the same rate statistics listed above. This is why WPA can be a most valuable tool when measuring bullpen success.

Entering Thursday’s contests, the Mets bullpen has posted a combined 374.9 WPA, a close second in the major leagues only to the Rockies (384.6). Take a closer look at the numbers and immediately Jorge Julio’s -47.2 WPA jumps out. Consider him removed from the equation and realistically, the Mets have the most effective bullpen by WPA.

Now also consider this: The Mets bullpen has logged 160.2 innings, the most in the MLB. In this case, WPA is a counting statistic because the individual WPAs are added together. Obviously, the more innings pitched, the more opportunities to garner WPA.

Let’s break it down by general usage.

LONG RELIEF
Oliver - 41.2 WPA
Bell - -3.3
Fortunato - -11.5
Julio - -47.2

I lump Julio into the long relief category because many of his appearances came with the Mets either ahead or trailing by a large margin, thus less high leverage situations.

Immediately it is clear, the extra innings the bullpen has thrown, mostly due to the unsual high usage of long relievers, has not had a huge bearing on the total WPA. Oliver has been a good soldier for the Mets, this year’s Roberto Hernandez if you must.

SETUP RELIEF
Sanchez - 164.0
Heilman - 116.1
Bradford - 14.7
Feliciano - 8.5

WPA tells us that the relievers used the most often in critical situations have found great success in 2006. Sanchez and Heilman have been brilliant, but the WPA numbers are astounding. Bradford and Feliciano, or as I like to call them, “Feliciford,” are one reliever split into two. Feliciano has been shaky the past two days; this is one cause for concern. He has emerged as the primary lefty option in the middle innings and if he cannot straighten things out, the Mets may be forced to explore other options (Dave Williams?).

CLOSER
Wagner - 92.4

Wagner still has to catch Tom Gordon (140.6) to be the best closer in the NL East, but complaints against Wagner should be few and far between. He is right about where projections expect him.

Most fans would say hitting has saved the Mets on more counts than pitching, but WPA says otherwise. Mets hitters are a combined -53.3 WPA while the starters (228.4) and the relievers account for the positive. The bullpen has been used very often in the first phase of the season; this should be the primary concern for the pitching staff. With the recent additions of Alay Soler and Orlando Hernandez, one can only hope the bullpen faces less strain in the coming weeks and eventually months.

Matt Gelb, a journalism student at Syracuse University, infrequently writes columns for Metsgeek.com. He can be reached by e-mail at magelb@syr.edu.

15 Responses to “A Real Relief”

Comment posted by Richard on May 26, 2006 at 9:26 am (#40952)
Can someone give us the skinny on Dave Williams. What are his pluses and minuses? Does he have arm or injury issues? What are his core pitches and what can Rick Peterson do to help? Being lefty it usually takes more time to mature
Comment posted by sam on May 26, 2006 at 9:50 am (#40956)
according to ESPN insider, Williams, a lefty, throws 64% fastballs, 20% curves, and 13% change-ups, and mixes in a slider every once in a while at 3%.

he doesn’t throw his change-up to lefties, opting for 29% curveballs instead.

he pretty much works down and away to both lefties and righties, although he’ll come up and in (out of the strike zone) on lefties every once in a while, and up and away on righties.

not much, but i hope that helps.
Comment posted by Lets_Go_Mutts on May 26, 2006 at 10:12 am (#40962)
Just a note on usage of the pen, projecting innings for the year based on 46 games-
———–Proj.—Actual—–Max
———–2006—-2005—–2003-5
Wagner—-81—–77.2——86
Sanchez–102—–82——–82
Heilman—-95—-108——-108
Oliver——-82—–14———180

nobody else projects to pitch over 60 innings in relief, barring injury the only one that might be overused so far appears to be Sanchez and he is supposed to thrive on work. So I don’t see a problem unless the bottom of the rotation gets worse instead of better.
Comment posted by Danny on May 26, 2006 at 10:14 am (#40963)
Between Anderson Hernandez and Kaz Matsui, the Mets have more than -200 WPA at second base. That is just dreadful. Albert Pujols WPA is 425.5. That is just crazy. I recommend everybody click on the Fan Graphs link that Matt provided. It is very interesting (such as Jeter being 200 WPA, but A-Rod being -40 WPA).
Comment posted by Mike on May 26, 2006 at 10:32 am (#40967)
Speaking of Matsui, I posted this on the other thread a few minutes ago:

Matsui’s troubles aren’t related to a simple mid-May slump. He’s 30 years old, and after parts of 3 seasons in the majors, he’s not only been a poor hitter, but he’s been consistently *declining*. The following is part of yesterday’s Mets post on my blog (to read the whole thing, click on my name):

Since landing at JFK two years ago, his batting average has gone from .272 to .255 to .227. His walks per 10 ABs? 9—>5—>4.5. Isolated slugging? 124 to 97 to 82.

When a 30+ year-old guy has over 900 plate appearences in MLB, and he’s declining so significantly across the board, he’s doomed. There’s NO REASON to keep penciling him in. None.
Comment posted by Brian S. on May 26, 2006 at 11:32 am (#40974)
So in Dave Williams, we got a sub .500 pitcher with a mid 4’s era who played for the Pirates? Sound familiar? Omar basically got a left handed Kris Benson for free. I did not see any Benson lovers post after his stellar 2 inning performance earlier this week. His ERA is now 4.62 and he is striking out a whopping 4.6 guys per 9.
Comment posted by peeder on May 26, 2006 at 12:25 pm (#40980)
WPA isn’t so hot a stat to use for this, because it becomes more volatile as it approaches the 9th inning from what I can see.
Comment posted by SF Metfan on May 26, 2006 at 1:16 pm (#40983)
It does get more volatile later in games (close games, that is) because WPA is a stat that measures “clutch” performances. A strikeout in the 1st inning, according to WPA, is less important to the game’s overall outcome than a strikeout in the 9th inning of a one run game. Naturally, Sanchez and Heilman will have impressive WPA numbers because they only pitch late in games, and in close ones. Therefore, WPA cannot be unequivocally compared from player to player. You can compare closers to closers, setup men to setup men, and so forth, but comparing Sanchez to Pedro doesn’t really tell you much.

It’s a great stat, and perfect to throw at anyone that thinks A-Rod was deserving of the MVP last year.
Comment posted by dptydwg420 on May 26, 2006 at 1:20 pm (#40984)
So in a 1-0 nothing game…does a strikeout in the 7th with say 2 guys on get more WPA than a strikeout in the 9th with nobody on?
Comment posted by Matt Gelb on May 26, 2006 at 1:47 pm (#40988)
dptydwg:

All situations are with one-run home team lead

Home WP when visitor has two outs and runners at second and third in the top of the 7th: .756
Home WP after third out is recorded (start of bot 7th): .824
DIFFERENCE: .068
….

Home WP at start of the 9th, no one on: .867
Home WP with one out, no one on in top 9th: .925
DIFFERENCE: .058

So the verdict? Strikeout in the 7th with runners at second and third is more valuable than strikeout for the first out in the ninth.

This makes the most sense, obviously.
Comment posted by Matt Gelb on May 26, 2006 at 1:48 pm (#40989)
Oh, and I forgot to add this.

It’s a win expectancy finder you can play around with yourself.
Comment posted by Mike on May 26, 2006 at 2:30 pm (#40994)
Anyone point me to a website that explains how WPA is calculated? I’m not familiar with it.
Comment posted by dptydwg420 on May 26, 2006 at 3:09 pm (#41001)
Thanks a lot Matt :)

I’m actually surprised the differential is that close, but I guess since it’s only a 1-run lead, every out is very important. At first thought I was thinking getting an out having guys on in a late inning other than the 9th with those runners being able to score on just a single would be worth much more than the .01 spread.

I guess maybe you can also look at in terms of % added rather than just the pure point spread in WPA. That would give even more weighting to the 7th inning example vs. 9th inning one since the WP in the 7th inning was .756 vs. .867 in the 9th.

That’s for the Win Expectancy finder, very cool :)
Comment posted by Dan in Gettysburg on May 27, 2006 at 10:24 am (#41203)
Anyone point me to a website that explains how WPA is calculated? I’m not familiar with it.

http://www.hardballtimes.com/main/article/the-one-about-win-probability
Comment posted by badgercta on June 14, 2006 at 10:22 am (#47701)
翻译公司