SB Nation User Blog: danmerqury

Making Sense of the Fangraphs Game Graph: A Basic Look at Win Probability

danmerqury — Wed, 28 Oct 2009 06:03:06 -0000

This is a Fanpost aimed at those of us who may not be as stat-literate as others, but want to know what the Fangraphs game graph is all about. I’ll start from the very beginning, hopefully passing along useful knowledge about run expectancy and win probability, among other things. And if you already have a good handle on the principles, hey, it’s good review. Or something.

There’s a fundamental concept in this area of baseball research called base/out states. It’s simple, really. If we’re just talking about one inning, what’s the least amount of information you absolutely have to know to understand the situation? All you really need is the number of men on base and how many outs there are. It turns out that there are eight different ways to have men on base: no one on, a guy on first, a guy on second, a guy on third, guys on first and second, first and third, second and third, and bases loaded. Furthermore, all of these can occur with zero, one, or two outs, which leads to 24 total base/out states.

Most of the data that I'll talk about in this post comes from The Book: Playing the Percentages in Baseball (Tom Tango, Mitchel Lichtman, Andrew Dolphin). In The Book, the authors created a run expectancy table. They threw all of the games played from 1999-2002 into a massive database (over 85,000 innings!) and crunched the numbers. What they primarily looked at was each base/out state, and how many runs the average team scored starting from that state until the end of the inning. For example, they were able to show that, on average, a double that starts the inning makes a team score 0.634 more runs than they would have scored before the hit. And a strikeout with the bases loaded and no outs hurts the team to the tune of -0.767 runs. Over a full season, we can add up a player’s contributions (positive and negative), and come out with a number which expresses the total amount of offense that our example player produced. Fangraphs tracks this stat under the name RE24.

To take run expectancy a step further, Tom Tango decided to expand the context. Sure run expectancy is fine for an inning, but what if we wanted to look at a full ballgame? By crunching the numbers again, Tango created a set of tables that express the win probability for both teams at any point in the game. In other words, if team A was up by four runs in the bottom of the sixth, and team B had a man on first with two outs, there’s a section of the table that gives the probability of either team pulling out a win. As you can imagine, due to including the 24 base/out states in addition to the inning number and score difference between the two teams, it’s a really large set of tables.

It’s this win probability that appears on the game graphs generated by Fangraphs. Here’s an example. Click the graph to see the full page.

via www.fangraphs.com

Yeah, I chose a good one. Ignore the bar graph at the bottom, I’ll get to that in a bit. For now, focus on the crazy line graph. Remember, in win probability, the numbers assume perfectly average teams. So at the beginning of the game, both teams have exactly a 50% chance of winning. In this game, by the top of the second, after Justin Morneau’s grand slam, the Twins had an 87.6% chance of winning the game. The score was 7-2. And by the middle of the third inning, the Twins had a 98.9% chance of winning, assuming perfectly average teams. Of course, we know what happened. Most game graphs aren’t nearly this dramatic, but that’s the beauty of the model. When the graph takes crazy dives and turns, it’s visually obvious that it was a great game to watch.

Also, much like RE24, we can tally up a player’s contributions to win probability. This stat is known as Win Probability Added (WPA). In WPA, 0.500 corresponds to one full win, as the probability of winning changed by 0.500 (50% to 100%). In the aforementioned game, Gio Gonzalez was tagged with -0.587 WPA, more than enough to singlehandedly lose the game. Thankfully, Holliday and Cust combined for a positive 0.638, enough to overcome Gio’s poor pitching. The unique thing about WPA is that, unlike other stats, it’s not context-neutral. Almost every other statistic out there (batting average, OPS, wOBA, whatever) deals with players in a theoretical vacuum. But WPA deals with actual game occurrences. Give up a solo home run when your team is up by ten? Not a big deal, and WPA shows that. Give up a bases-loaded walk in a tie game in the bottom of the ninth? That’s far worse. WPA reflects that.

As an extension to this, Tom Tango created what he called a Leverage Index (LI). Leverage Index basically explains how critical a situation is. To put it in more technical terms, Leverage Index gives a numerical representation of the biggest possible swing in win probability, where 1.0 is the average. To use my previous examples, the LI in a game situation where you’re up by ten is very low. If you give up a home run, the win probability isn’t going to change much. But the LI in the bottom of the ninth of a tie game with the bases loaded is off the charts. The high LI shows that any little mistake could be devastating. Leverage Index is always denoted in a Fangraphs game graph by the bar graph underneath the regular win probability graph. In that crazy Twins game I mentioned earlier, the LI sinks to near-zero levels when the Twins are up 12-2. And yet, when the A’s came roaring back, the LI started to rise again. The last play of the game had a huge LI, as Michael Wuertz had runners on first and second with two outs in the bottom of the ninth of a one-run game.

I admit I’ve gotten into the habit of keeping the game graph open in another tab alongside MLB Gameday (and AN, of course) during a game. It’s pretty addicting to watch the graph unfold in real-time, and wonderfully cathartic to see a crazy game in graphical form.

Moneyball Revisited: How Much Did We Overachieve?

danmerqury — Tue, 06 Oct 2009 00:48:49 -0000

We all know about the size of our payroll. We all know that it makes us trade away, buy low, and otherwise obtain undervalued players. It's been seven years since the 2002 Moneyball draft, and our payroll hasn't gotten any bigger relative to the league. In part 2 of my apparent Jack an Idea From Another Website and Use It to Make a Cool Chart series, I found out how much we truly overachieved.

Click for a bigger version.

First things first: The idea came from Beyond the Box Score. Go there. They rock.

This chart simply shows how much the 2009 A's overachieved. To calculate the numbers, I pulled performance dollars from Fangraphs and subtracted their 2009 salary. Fangraphs converts their WAR numbers (wOBA, FIP, and UZR based) into performance dollars by finding out the average market value of a 1 WAR player and extrapolating. (More information here.) Last year, it came out to around $4.5 MM per win. All of the salary numbers came from Cot's Baseball Contracts. For the players who left or came to Oakland midway through the season, I calculated their prorated salary based on how long their stay with Oakland was. For the players that Cot's didn't have salary info for, I assumed the league minimum ($400,000).

Couple things:

The players that Oakland didn't have on their payroll for a full season are Matt Holliday, Adam Kennedy, Jack Hannahan, Orlando Cabrera, Scott Hairston, Craig Breslow, Russ Springer, Edgar Gonzalez, Brett Tomko, and Clay Mortensen. Some of these were trades and some were waiver pickups.
Included in the calculations are $1.5 MM that we paid St. Louis as part of the Holliday deal and $250,000 that we paid Orlando Cabrera as an "assignment bonus". Also, Jason Giambi got a full $4 MM from Oakland, even though he didn't play the whole year. I didn't include the $1.5 MM buyout for Giambi, as that's on the 2010 payroll.
The small line on both sides of the chart indicates $10 MM worth of production over salary.
I used a 20 innings pitched, 100 plate appearances cutoff, with two exceptions.
If you're getting paid almost 18% of the entire team payroll, you're on the chart, even though you only played in, say, 8 games. Also, Duchscherer's on the chart for the same reason.
Brett Anderson and Ryan Sweeney are really, really good.
Some good news: Out of the players that didn't produce as much as their salary, numbers 2, 3, 4, 5, and 7 are off the books in 2010. Number 6 is Scott Hairston, who has had injuries all year.
Out of these 39 players, we spent $59.3 MM. We got $165.1 MM worth of performance, which is good for 36.5 wins above market value. Not bad, Billy.

A Pitchf/x Primer, and why Trevor Cahill should start 2010 in Sacramento.

danmerqury — Mon, 21 Sep 2009 04:53:52 -0000

I don’t know how familiar the AN population is with the specifics of the wonderful Pitchf/x system, so I’ll start from the beginning. For those that do…I’m sorry. Skip ahead if you’d like.

In 2006, Sportsvision (the company behind the original yellow TV first down line and pretty much every piece of TV sports graphic wizardry out there) launched Pitchf/x. They installed multiple high-speed cameras in every MLB stadium that track every pitch, computing the ball’s trajectory and location to within a half an inch. Over time, the power of the system grew, and today, the cameras allow the Pitchf/x system to calculate a huge collection of variables about the trajectory of every pitch, definitely not limited to what you see on MLB Gameday.

Thankfully, data for every pitch is freely available from MLB.com. It’s easier to access data from a source like FanGraphs or Brooks Baseball, where they have tools and graphs already set up, but if anyone wanted to work with the raw data, you can download the spreadsheets too. I used FanGraphs for this FanPost, but it’s really up to you. In either case, the data generated by the Pitchf/x system gives us the greatest tool we’ve ever had for analyzing a pitcher’s pitches. For example, here’s a movement graph from Justin Verlander’s start on 8/8.

Keep in mind that this isn’t a location graph. That is, the dots don’t correspond to where they crossed the plate. Rather, this is a movement graph. The vertical axis corresponds to vertical movement—from the graph it looks like his fastball (the green dots) averaged about seven or eight inches of backspin. In other words, his fastball crosses the plate roughly eight inches higher than it would have if it was thrown without spin. The horizontal axis is very similar, showing horizontal movement. In these graphs, all movement numbers are seen from the batter’s point of view, so Verlander’s fastball comes in to right-handed batters by approximately six inches. It’s counterintuitive, but fastballs break just as much as curveballs do. The very best curveballs can break downwards by around ten inches (think Zito), but almost everyone throws a fastball that breaks upward anywhere from 6-11 inches. Upward vertical movement has to fight against gravity, however, so a hard fastball will fly far straighter than a spinless ball will, where downward breaking curveballs fall with gravity working for it.

The reason I used Verlander as an example is because his pitches are well-defined. He throws them repeatedly, with consistency. Take a look at this.

This is from Trevor Cahill’s 5/7 start, which, by all accounts, was a good one. He went seven innings, generated loads of groundballs, and only allowed one run. But yet, his movement chart is troubling. Unlike Verlander’s, Cahill’s pitches are not well-defined. Instead, the graph looks more like a smear of pitches, as opposed to tight bunches. And this is one of his good starts. When he’s bad…

Gio Gonzalez has, to put it lightly, control problems. He has trouble locating his pitches. But when he throws his fastball, the same pitch comes out every time. Cahill has trouble throwing the same pitch start after start. Here are all of Cahill's starts this year, from the beginning to 9/2. The date text is in green if he had a quality start, and it appears in red if he didn’t. (I changed the definition of "quality start" from 6.0 IP/3ER to 5.0 IP/3ER because he’s a rookie, and, well, he wouldn’t have too many quality starts if I left it as is.)

I see no consistency. I also don’t see much of a pattern. In an attempt to find one, I sorted his starts again, but instead of using earned runs as the deciding factor, I sorted them by the number of groundballs. He’s a sinkerballer, so I considered any start where he generated 50% groundballs or more as good. Anything under 50%, again, appears in red.

Now there’s some semblance of a pattern. I came to the rather obvious result that Cahill tends to generate more groundballs when his fastball has more sink. To illustrate the difference, I pooled all of his starts together, sorted by groundball rate.

In the starts where he generates less than 50% groundballs, his fastball has a nasty tendency to rise, losing it's characteristic sink. When his fastball is working, it tends to lose some of that positive vertical movement and causes batters to get poor contact. It looks like Cahill tends to overthrow his fastball at times, causing it to get more backspin than it should, which leads to the rise.

Cahill needs to get time in Sacramento to work on his consistency. If his fastball can become that consistent sinking force it needs be, he can succeed on the highest stage.

Poll

So, Cahill, 2010. Sacramento or Oakland?

Prospect Buzzkill '09!

danmerqury — Wed, 16 Sep 2009 00:39:48 -0000

This isn’t a fun post to read. It’s not even a fun post to write. That’s right, folks, it’s the latest installment of doom and gloom.

Okay, it’s not that bad.

I mainly wanted to write this post after seeing this article on Fangraphs. It’s something we should all read. I’m not preaching ultimate pessimism or anything, but, well, I’m just a little tired of seeing something like this.

C1: Donaldson, Josh
1B: Carter, Chris
2B: Weeks, Jemile
SS: Cardenas, Adrian
3B: Wallace, Brett
LF: Desme, Grant
CF: Brown, Corey
RF: Doolittle, Sean

It’s fun. It’s exciting. They’re shiny new toys. But we’ve got to be, at least a little bit, realistic. The Victor Wang SABR article that Fangraphs cites is a great read, but you can glean all of the necessary information from the tables he attaches. Not all prospects are going to succeed. Victor Wang analyzed Baseball America’s top 100 rankings from 1990 to 1999. He then put each prospect into categories according to Wins Above Bench (think WAR with a slightly different baseline). Prospects that averaged over 4 WAB per year he considered "stars". Those averaging between 2-4 WAB/year were "everyday" players. Those between 0-2 WAB/year were merely "contributers", and those below zero he deemed as "busts". I’m going to take a look and revisit Baseball America’s 2009 Top 100 Prospects preseason rankings, and give this an A’s related look.

Brett Anderson (7)
Pitchers in the top 10 busted 31% of the time. Only 8% went on to be solid regulars or better.

Trevor Cahill (11)
Pitchers ranked 11-25 busted 32% of the time. Interestingly, 15% became at least solid regular players.

Brett Wallace (40)
Hitters ranked 26-50 busted 35% of the time. 20% became everyday players or stars.

Michael Ynoa (54)
Aaron Cunningham (55)
Adrian Cardenas (74)
Pitchers ranked 51-75 busted 39% of the time. 8% were regulars or better.
Hitters ranked 51-75 busted 45% of the time, yet, 18% reached everyday status.

Chris Carter (76)
Gio Gonzalez (97)
Pitchers ranked 76-100 busted 43% of the time. Only 7% were regular players.
Hitters ranked 76-100 also busted 43% of the time. 13% were ever really any good.

Remember, no team is ever fully homegrown. Of course, these are preseason rankings, and things chance (hell of a season, Carter), but we can’t count on a lineup or rotation full of prospects. It’s just not going to happen. Remember Javier Herrera? Me neither.

Fangraphs article on Michael Wuertz, and how dominant he's been.

danmerqury — Mon, 17 Aug 2009 17:24:10 -0000

Fangraphs article on Michael Wuertz, and how dominant he's been.

That chart at the bottom of the link is definitely something. And on 25% of his sliders, he gets not just a strike, but a swinging strike? Wow.

In Defense of Jack Cust

danmerqury — Thu, 13 Aug 2009 23:47:48 -0000

Yeah, I know, a lot of the AN faithful already know a majority of what I’m trying to say, especially the justification of 2007-2008 Cust. But for those who don’t, and especially for newcomers to the AN scene, I’m going to lay out my case: Jack Cust is not only an above-average hitter, but a very welcome and valuable addition to the Oakland Athletics.

We all know how it started. In his first week as an Athletic, he almost averaged a homer a day. His OPS didn’t fall out of the 1.000’s for just shy of a month. Jack Cust finished his first regular major-league season with an awesome line of .256/.408/.504. To put that in perspective, the only other .900 OPS or better seasons we’ve had this decade are 2006 Frank Thomas, 2004 Erubiel Durazo, and 2000-2001 Jason Giambi.

A lot of the Cust criticism comes from his strikeout rate, which, let’s face it, is enormous. In 2007, he struckout in 41% of his at bats, easily toppling Ryan Howard, the only other player remotely close. Cust is a textbook example of a "three true outcomes" player, meaning that every time Cust comes up to bat, he either strikes out, walks, or hits a home run. Thankfully, in 2007, he led the league in walk percentage, at 21% of his plate appearances. It’s worth noting that with the exception of moving a runner from first to third, a walk is every bit as valuable as a single. Even with his anemic batting averages, Cust still managed to get on base 40% of the time. That’s a fine baseball player, folks.

Cust manages to walk more than virtually anyone else due to his phenomenal eye and plate discipline. In 2007, Cust was second in the league in outside-swing percentage. He swung on pitches outside the strike zone less than 14% of the time, behind only Luis Castillo. In total, he swung only on a third of all pitches. Of the balls he did hit into the field of play, he came in seventh in line drive percentage. Additionally, he led the league in home run to fly ball ratio, an indicator of power, coming ahead of Ryan Howard, Jim Thome, Carlos Pena, and Alex Rodriguez.

So what does it all mean? In 2007, Cust put together a fantastic offensive season. In 2008, his walk rate fell a little, as did his home run to fly ball ratio. He still managed an .851 OPS. In this decade, we’ve generally only had one player per year with an OPS that high, save for a few years in the early part of the decade (Giambi, Chavez, Tejada).

Of course, 2009 has been a down year, and most of the criticism that doesn’t stem from his strikeouts comes from 2009. I’ve heard "quad-A player". I’ve heard that the league has figured him out. It’s simply not true.

Keep in mind that before the season, Cust was told to cut down on strikeouts and swing more. It’s safe to say that the new contact-oriented Cust experiment was a failure. Thankfully, the experiment is over, as of the beginning of July. Take a look at these.

Since the around the beginning of July, almost all of Cust’s peripheral stats took a sharp turn upward. His average increased, and more importantly in Cust’s case, his on-base percentage took a huge upswing. This was buoyed by his walk percentage shooting through the roof. Along with his changing walk percentage, his strikeout percentage increased dramatically. Normally, this would be a worrying sign, but here it just signals that "three true outcomes" Cust is returning. And of the balls Cust did hit fairly, his groundball percentage decreased while his line drive percentage increased. These both triggered an increase in his batting average on balls in play (BABIP). These are all good signs.

Cust isn’t fully back yet, though. The power simply isn’t there. His slugging percentage actually decreased since July, while the number of fly balls he hit dipped down for about a month. It looks like Cust’s plate discipline and his approach to taking pitches has returned back to normal. He’s just not hitting the Custian homers we all know and love. He’s again being extremely selective about the pitches he swings at, and, thankfully, as evidenced by the large number of loud outs he’s been getting recently, it’s only a matter of time before the power returns.

Expect to see the return of Spartacust, folks. If not for the rest of the year, "three true outcomes" Cust should be back in full force for 2010, provided that the management doesn’t try any new experiments.

Visual Representation of the 2006 Oakland Athletics

danmerqury — Wed, 12 Aug 2009 01:56:20 -0000

It's nice to look back at brighter days. Although it probably wasn't the best team of the 2000's (that would either be 2001 or 2002, they both won 100+), it was the only team that got past the ALDS. After the jump is a visual diagram of the 2006 Oakland A's, showing when they came and left Oakland. Hint: Just three years later, we don't even have a fifth of this roster still with us.

via i20.photobucket.com

Orange marks a season with the A's, where gold highlights the 2006 ALDS-winning season. Brown indicates a season played elsewhere, not in Oakland.

In the case of a player playing for multiple teams in a season, I went with whichever team he played more games for in that year. Batters are sorted by games played, and pitchers are sorted by innings pitched.

I got the idea when I found a similar chart done by Craig Robinson on the 1986 World Series-winning Mets. Check his site out, guys, it's got a lot more cool and interesting visual charts.

http://www.flipflopflyin.com/flipflopflyball/index.html

Joe Posnanski's article on the death of Moneyball.

danmerqury — Thu, 06 Aug 2009 01:41:38 -0000

Joe Posnanski's article on the death of Moneyball.

No, wait! It's actually good (from the pen of Posnanski, how could it not be?). Instead of talking about how Moneyball never worked, the A's playoff failures, OBP is a sham, etc, he simply mentions how the rich got wiser and smarter.

Pitch Type Linear Weights: Our Bullpen

danmerqury — Fri, 05 Jun 2009 04:09:09 -0000

It's time for another look into the awesome tool that is Pitch Type Linear Weights. If you missed the first part, where I introduce the stat and talk about the rotation, you can go here.

I wanted to take some time to talk about one thing that Devo noticed in a comment in my previous post. The sample size is teeny. Far too small to be all that useful. At the time of writing my last post, most of the starters I looked at had around 40 IP. Most of the relievers on the roster have pitched around half of that, which leads to an even smaller sample size. Because of this, I'll include 2008's numbers to give a larger body of work to analyze. On to the numbers!

Brad Ziegler - I'll start with our AN fan favorite (Brad, if you're reading this, hi!). PitchFX shows a fastball, that awesome frisbee slider, and a changeup.

2008

wFB/C: 1.63

wSL/C: 3.46

wCH/C: 1.26

2009

wFB/C: -0.60

wSL/C: 5.30

wCH/C: 1.00

That slider truly is a great pitch, baffling hitters everywhere. His fastball was very nice last year as well, considering he threw it over 85% of the time. His changeup confuses hitters pretty well too. This year, he seems to be getting less performance out of his fastball, and a lot more with his slider. He's compensating by throwing his fastball a hair over 8% less in 2009 over 2008. I'll attribute this loss of performance on the league adjusting and realizing what a talent he is.

Andrew Bailey - Might as well continue with the good, as Bailey has shown himself to be a true gem in the bullpen. It's worth noting that he's thrown more innings than anyone else in our bullpen, as well as Dana Eveland. 34 innings in 24 games! PitchFX reports a fastball, cutter, and curveball. It also sees a slider and changeup, but those are thrown less than one percent of the time. Judging by the speed of these pitches, I'll assume they're misclassified cutters. I'll leave them out.

2009

wFB/C: 1.97

wCT/C: 0.79

wCB/C: 2.58

Thirty four career innings isn't exactly the best sample to really look at, but we can still see what Bailey's succeeded with so far. Bailey's done some pretty awesome things in his time in the big leagues, and it shows. He doesn't have a bad pitch. He throws his plus fastball sixty percent of the time, and it's working. The occasional cutter and curveball serve as a nice change of pace, and he's succeeding with those too. Hopefully he'll be a rock in our pen for a long time.

Michael Wuertz - Another candidate in the overused department, though not nearly to the extent that Bailey has been. PitchFX shows a fastball and slider, as well as the occasional changeup. I'll throw out the extremely rare curveball classification.

2008

wFB/C: -3.19

wSL/C: 0.76

wCH/C: 4.23

2009

wFB/C: -1.02

wSL/C: 2.74

wCH/C: 8.41

He's never been one to dominate with his fastball, but he has a great slider he can use to get hitters out. Throwing a changeup every once in a while really catches hitters off guard. It's kind of interesting to note that he threw his fastball over half of the time in his first year, with his slider making up a quarter of his pitches. Over the years, he slowly decreased the use of his fastball and began relying more on his slider. His fastball and slider are now more like 30% and 60%, respectively.

Santiago Casilla - Casilla is kind of an odd case, as he seems to go through stretches of amazing dominance, followed by stretches of bizarre hittability. PitchFX shows a fastball, slider, and changeup. In his first three seasons, PitchFX reported about a quarter of his sliders as curveballs, which may indicate a change in the movement of his slider.

2007

wFB/C: 0.25

wSL/C: 1.81

wCH/C: 0.48

2008

wFB/C: -0.14

wSL/C: -1.18

wCH/C: -1.61

2009

wFB/C: -1.83

wSL/C: 1.68

wCH/C: 3.30

If you go to Fangraphs, you can see the stats from all of his years, which I only bring up (and recommend) here, because Casilla is definitely an odd case. He's inconsistent. His slider and changeup numbers flip-flopped from positive to negative in 2008, only to bounce back this year. Granted, his first three years were his Jairo years, with extremely few innings pitched, but he's been inconsistent after that as well. The drop in production from 2007 to 2008 is particularly alarming. He looks like he's on the rebound from these numbers this year, but his FIP is up by two entire runs in 2009 from the low-4 FIPs he posted in 2007 and 2008. Not quite what we want to see from our once-possible future closer.

Russ Springer - Another of our somewhat unexpected free agent bullpen acquisitions. Springer's had less success this year than Wuertz, but hey, he taught Braden his cutter. Give the man his due. PitchFX shows a fastball, slider, cutter and curveball. It also reports a changeup, but I'll consider those as misclassified sliders or cutters.

His pitch type breakdown is actually quite interesting. In 2007, it looks like he cut down his slider and curveball and increased the frequency of his cutter dramatically. It seems to be that 2007 was the year he upgraded his cutter from an experiment to a pitch he used with regularity. 2008 brought the same amount of cutters, with more curveballs at the expense of his slider. This year, he's cut down on his curveballs again, increasing the frequency of his cutter yet again to 30%.

2007

wFB/C: 1.44

wSL/C: 3.36

wCT/C: 3.14

wCB/C: -1.10

2008

wFB/C: 1.75

wSL/C: 1.38

wCT/C: 0.79

wCB/C: -1.73

2009

wFB/C: -1.54

wSL/C: 4.64

wCT/C: -1.60

wCB/C: -5.63

This pitch breakdown gives us an interesting look at his linear weights. In 2007, his first year with his cutter in wide use, he dominated with it. His slider also got hitters out excellently. As he decreased the amount of sliders he threw, his cutter became less and less effective, and with that, his other pitches suffered. It would be interesting to see what would happen if he reverted to his 2007 pitch proportions.

This Fanpost is getting long enough already. In Part 3, I'll take a look at some other pitchers we have in the system. Thanks for reading, guys.

Pitch Type Linear Weights: Our Rotation

danmerqury — Fri, 22 May 2009 04:04:17 -0000

Fangraphs just unveiled a new tool from the the amazing minds at www.baseballanalysts.com: Pitch Type Linear Weights. Fangraphs has an excellent introduction to how it’s calculated here if you’re interested, but all you really need to know is that it allows us to look at pitcher’s pitch types and determine which type of pitches get hit around the most, as well as each pitcher's true "out" pitch. Of course, the first thing I did was look at our green and gold.

But first, I needed some sort of frame of reference. The numbers are weighted so that zero is a perfectly average pitch, negative numbers signify below average pitches, and positive numbers are better than average. They also give two numbers per pitch. Using a fastball as an example, wFB is just a counting stat. It's how many runs above average that fastball was over the whole season. Unfortunately, it depends on how many times that pitch was thrown, and for how many innings. What's better for comparison's sake is wFB/C, which extrapolates the performance over 100 pitches. That's the stat I'll be using from here on out. To put things in perspective, Johan Santana's changeup got a 3.01 wCH/C in his 2006 Cy Young-winning season. In other words, for every 100 times Santana threw that changeup, he averaged a hair more than three runs below average saved. That's a very good pitch.

For this analysis, all the stat names are based on a two-letter code for each pitch:

FB = Fastball
SL = Slider
CT = Cutter
CB = Curveball
CH = Changeup
SF = Split-Fingered Fastball
KN = Knuckleball

To avoid making this Fanpost too long, I'll look at the rotation first.

Dallas Braden - Braden's an interesting case, as this year marks his first year with real consistent success. PitchFX shows him with a fastball, slider, changeup, and new for this year, a cutter. It also shows a curveball, but last year he threw it 0.3% of the time. I'm going to assume these were misclassified sliders.

wFB/C: -1.49

wSL/C: 1.44

wCT/C: 1.35

wCH/C: 2.66

Immediately, his changeup stands out. That 2.66 wCH/C is pretty great. Additionally, that mark has improved from last year's 1.96, which also isn't exactly bad, by any means. Dallas's fastball isn't doing him any favors, and he throws it 50-60% of the time. Thankfully, his excellent variety of offspeed junk helps him succeed. His improvement in his slider's performance from 2008's 0.35 to this year's 1.44 is worth noting. It also looks like his cutter is a success. Good to see.

Brett Anderson - I'll start with the rookie with more success so far. PitchFX shows a fastball, slider, curveball, and changeup.

wFB/C: -1.58

wSL/C: 0.01

wCB/C: 0.78

wCH/C: 0.66

Already, it doesn't look as good as Braden. His only above-average pitches so far are his curveball and changeup. His slider is very much average, and his fastball is a hair worse than Braden's. Obviously, as a rookie, it's understandable to not have any dynamite pitches, but on the whole, Anderson's not doing bad at all.

Trevor Cahill - Now we get into the ugly. Cahill throws a fastball, slider, curveball, and changeup, with 0.1% of pitches registering as what PitchFX calls a split-fingered fastball. I'll throw it out.

wFB/C: -0.96

wSL/C: -10.53

wCB/C: -5.33

wCH/C: 1.07

Yikes. His only above-average pitch is his changeup, and his other three pitches range from bad to atrocious. Take his slider and curveball stats with a grain of salt: he's thrown them 4.1% and 5.0% of the time, respectively, leading to a small sample size, you know the rest. No way his slider will stay ten runs below average. Still, though, it obviously hasn't been the greatest start for Trevor.

Josh Outman - Na na na na na na na na, na na na na na na na na, Out Man! Josh has a fastball, slider, curveball, and changeup.

wFB/C: -1.19

wSL/C: 5.35

wCB/C: -3.41

wCH/C: 1.84

Look at that slider! That number is probably too high to be sustainable, but still, it's a great pitch. And he's thrown it almost 20% of the time, so this isn't some absurdly small sample size fluke. I don't know what Outman did, but his slider was a full run below average last year. In fact, Josh's best pitch last year, his fastball, was still a half a run below average. That low curveball score isn't as bad as it looks: he's only thrown it 4.2% of the time this year. Additionally, this seems to confirm fears that his speedy fastball is just too straight, without much movement. He's not having very much success with it at all.

Dana Eveland - Not quite the greatest year for Dana. PitchFX shows a fastball, slider, curveball, and changeup.

wFB/C: -1.12

wSL/C: -4.83

wCB/C: -0.61

wCH/C: 1.30

This is really kind of surprising. The slider, which he throws a quarter of the time, is atrocious. And it was his best pitch last year, at 1.76 wSL/C. Pretty much the only bright spot here is his changeup, which improved over last year by almost two runs per 100 games.

Sean Gallagher - I'm not going to show this year's stats after two games started. You don't want to see them, trust me. Gallagher throws a fastball, slider, curveball, and changeup.

2008

wFB/C: -0.67

wSL/C: -0.72

wCB/C: 0.43

wCH/C: 0.62

Oh fine.

2009

wFB/C: -0.47

wSL/C: 0.83

wCB/C: -6.78

wCH/C: -8.50

I'm still going to focus on last year's stats, because of Sean's mechanical issues that he's had this year. Keep in mind that these numbers were the result of 12 games with the Cubs and 11 games with the A's. Gallagher was pretty thoroughly average last year. His fastball and slider were both a tick below average, and his curveball and changeup were a bit above. With Sean's terrible 2009, we can only hope he regains last year's form and adds to it eventually.

Justin Duchscherer - Just for reference I'll look at Justin's amazing 2008 season. To, you know, make everyone else in our rotation look terrible. Or to make us all depressed. One of the two. PitchFX shows Justin with a fastball, slider, cutter, curveball, and changeup.

wFB/C: 0.74

wSL/C: 4.99

wCT/C: 2.45

wCB/C: 1.45

wCH/C: 5.82

Pretty amazing stuff, all across the board. His worst pitch, his 86-mph fastball, was still three-quarters of a run better than average! And that slider and change...wow. There's simply nothing to say except that we need Duke back.

I'll take a look at the bullpen next post. Thanks for reading, guys.