Baseball Prospectus recently released its Player Empirical Comparison and Optimization Test Algorithm (PECOTA) projections for 2011. PECOTA is a complicated projection system created by Nate Silver and recently improved by Colin Wyers. It uses the statistics and characteristics (age, height, weight, position) of a given player and the statistics and characteristics of similar players to arrive at projections for that player.
Accessing the data requires a subscription, so I can't reveal too much, but I will give you a look at a few players. Rather than giving you statistical projections, I will present lists of players which were considered the closest comparisons to Tigers players. These comparisons are fun, but should probably not be taken all that seriously.
Austin Jackson
Roger Bernadina Felix Pie Adam Jones
Most systems are projecting a regression for Jackson in 2011 because of his high strikeout rate and probably unsustainable BABIP. PECOTA is no different and the first two comparisons - Bernadina and Pie - are not very flattering. Jones is a little better, although he's a different kind of hitter at this point (more power than Jackson)
Daniel Fields
Justin Upton B.J. Upton Adam Jones
These are certainly optimistic comparisons for a kid who has not played above single-A. I wouldn't complain if he turned into any one of those players. I think it's more of a best case scenario than something we can expect though.
Brandon Inge
Howard Johnson Ron Santo Al Smith
I don't know about you, but Brandon Inge does not remind me of Ron Santo.
Jacob Turner
Madison Bumgarner Don Drysdale Rick Porcello
As I said with Fields, it's pretty encouraging for a player coming out of single-A to compare with productive major leaguers, especially when one of them is Don Drysdale.
Rick Porcello
Mike Witt Bill Parsons Dave Rozema
Mike Witt makes sense because he debuted in the majors at a fairly young age and took some time before he increased his strikeout rate. I hope Porcello is not Parsons because his career would soon be over.
Max Scherzer
Roger Clemens Zack Greinke Len Barker
Roger Clemens!
Saturday, February 12, 2011
Wednesday, February 09, 2011
WERC for Everyone
My recent post about Weighted Component ERA (WERC) sparked more interest than expected, so I have created a database for WERC for every pitcher and put it into google spreadsheets. The statistics on the table are:
PA - Plate Appearances
IP = Innings Pitched
ERA=Earned Run Average
FIP = Fielding Independent Pitching
wOBAA = Weighted On Base Average Against
RAA = Runs Above Average
WERC = Weighted Component ERA
An explanation of wOBAA, RAA and WERC can be found in the earlier post. To be clear, WERC is not any kind of sabermetric breakthrough. It's just the linear weights version of Bill James' Component ERA (WERC). It's another option falling within the ERA to FIP spectrum. It's best use might be as an alternative to ERA for evaluating past performance.
Note: Some of the data for this table were taken from Baseball -Reference and FanGraphs
PA - Plate Appearances
IP = Innings Pitched
ERA=Earned Run Average
FIP = Fielding Independent Pitching
wOBAA = Weighted On Base Average Against
RAA = Runs Above Average
WERC = Weighted Component ERA
An explanation of wOBAA, RAA and WERC can be found in the earlier post. To be clear, WERC is not any kind of sabermetric breakthrough. It's just the linear weights version of Bill James' Component ERA (WERC). It's another option falling within the ERA to FIP spectrum. It's best use might be as an alternative to ERA for evaluating past performance.
Note: Some of the data for this table were taken from Baseball -Reference and FanGraphs
Monday, February 07, 2011
Tigers Annual Coming Soon
As I mentioned previously, the 2011 Maple Street Press Tigers Annual will be on sale at Michigan newstands on March 1. Kurt Mensching of Bless You Boys is the editor and head writer of the book and and he recently revealed the table of contents. The book can be pre-ordered now at Maple Street Press.
I wrote articles on Austin Jackson and 1930s and 40s slugger Rudy York, but that's only a small part of the book. Other authors include John Parent and Matt Snyder of Motor City Bengals, Detroit News writer Tom Gage, Ian Casselberry of Big League Stew, Michael McClary of Daily Fungo, and Bill Ferris of TigsTown among others. Even Tigers second baseman Will Rhymes contributed an article discussing what it's like to be a major league rookie.
It is 128 pages about the Tigers with beautiful photography and no advertisements. It includes previews of the upcoming season, detailed prospect reports, historical pieces and much more. It should be great spring training reading, so check it out now at Maple Street Press.
Sunday, February 06, 2011
Which Tigers Pitchers did the Best WERC in 2010?
Yesterday, I introduced yet another pitching statistic - Weighted Component ERA (WERC). which serves as a supplement to the Defense Independent Pitching Statistics (DIPS). The WERC metric is similar to Bill James' component ERA except that it's based on linear weights rather than runs created. If you read the previous post, you will learn how WERC is calculated and how it can be used. Today, I'll touch upon the main points and then use it to examine the Tigers starting staff.
Most readers of Tiger Tales are familiar with the Fielding Independent Pitching (FIP) statistic. FIP estimates what a pitcher's ERA should have been based on events over which he has the most control - walks, hit batsmen, strikeouts and home runs. Other DIPS (tERA, SIERA,etc.) do the same thing, but they add batted ball data (ground ball, fly ball, infield fly and line drive rates) to the equation.
All of the DIPS metrics remove results of batted balls (hits and outs recorded by fielders) and sequencing of events (e.g. stringing together hits and walks versus spreading them out throughout a game) from pitcher evaluation. This is done because it has been shown that pitchers have less (although not zero) control over these things than they do on DIPS components
WERC estimates what a pitcher's ERA should have been based on walks, hit batsmen strikeouts and home runs plus results of batted balls (hits and outs). So, it removes less potential noise than say FIP, but it still does not take sequencing of events into consideration. It is useful for at least two reasons:
Table 1: ERA, FIP and WERC for Tigers 2010, 2011 Starters
Some items which I found to be interesting are:
Note: Some of the data for this article were taken from Baseball -Reference and FanGraphs
Most readers of Tiger Tales are familiar with the Fielding Independent Pitching (FIP) statistic. FIP estimates what a pitcher's ERA should have been based on events over which he has the most control - walks, hit batsmen, strikeouts and home runs. Other DIPS (tERA, SIERA,etc.) do the same thing, but they add batted ball data (ground ball, fly ball, infield fly and line drive rates) to the equation.
All of the DIPS metrics remove results of batted balls (hits and outs recorded by fielders) and sequencing of events (e.g. stringing together hits and walks versus spreading them out throughout a game) from pitcher evaluation. This is done because it has been shown that pitchers have less (although not zero) control over these things than they do on DIPS components
WERC estimates what a pitcher's ERA should have been based on walks, hit batsmen strikeouts and home runs plus results of batted balls (hits and outs). So, it removes less potential noise than say FIP, but it still does not take sequencing of events into consideration. It is useful for at least two reasons:
- It can be used as an alternative to ERA for evaluating past pitcher performance. It is especially useful if you think that a pitcher was either lucky or unlucky in his sequencing (e.g extremely high or low left on base percentage).
- In cases where a pitcher's ERA is much different than his FIP, WERC can be useful in determining whether the difference was caused more by results of batted balls in play or something else such as sequencing.
Table 1: ERA, FIP and WERC for Tigers 2010, 2011 Starters
Some items which I found to be interesting are:
- The Tigers five projected starters for 2011 - Justin Verlander, Max Scherzer, Rick Porcello, Phil Coke and Brad Penny - ranked ahead of the two departing starters - Armando Galarraga and Jeremy Bondermam - on both FIP and WERC. That bodes well for 2011, but keep in mind the small sample sizes for Coke and Penny.
- Verlander ranked only 11th in the American League on ERA, but finished third in FIP and sixth in WERC. That is another positive sign because it's an indication that his ERA may be lower this year if he pitches similarly to 2010.
- Penny's WERC (4.10) was much higher than his ERA (3.23) and FIP (3.40) which says that he may have fared poorly on batted balls. His BABIP of .326 backs that up.
- Porcello's ERA (4.92) was much higher than either his FIP (4.31) or WERC (4.42). Because his FIP and WERC are so similar, the discrepancy between ERA and FIP would appear to be the result of sequencing rather than batted balls. His 65.9 LOB% tells us a similar story.
Note: Some of the data for this article were taken from Baseball -Reference and FanGraphs
Saturday, February 05, 2011
Filling the Gap Between ERA and FIP
(Note: This is a heavily sabermetric post rather than a Detroit Tigers post. I'll apply the results to the Tigers in a later post.)
The limitations of ERA are well known in the blogosophere. Two of the biggest issues are:
(1) ERA gives pitchers full responsibility for all hits allowed despite the fact that their control over batted balls is limited. For example, a pitcher with a strong defense behind him will give up less hits (and thus fewer runs) than if he had a poor defense behind him.
(2) ERA gives pitchers full responsibility for sequencing of events. That is, it assumes that they can control when they give up hits and walks. For example, if a pitcher pitches extraordinarily well with runners in scoring position in a given year, he will have a lower ERA than if he had a typical year in those situations.
In reality, pitchers have limited control over both the number of hits they allow and sequencing of events. Thus, Defense Independent Pitching Statistics (DIPS) such as FIP, xFIP, tERA and SIERA have been developed to remove some of the noise of ERA. DIPS are based on things that pitchers do control for the most part - walks, hit batsmen, strikeouts, home runs and types of batted balls (ground balls , fly balls, line drives, pop flies).
Because they are based on things that pitchers essentially control, the DIPS metrics are said to be better measures of true talent than ERA. As a result, they are also better than ERA at predicting future performance. However, they only measure a portion of a pitcher's talent and should be used as complements to ERA rather than as replacements.
While pitchers do have less control over results of batted balls (hits and outs) than they do over walks, strikeouts, homers and ground balls, they do have SOME influence on results of batted balls. Some pitchers are indeed better than others at preventing hits on balls in play. Pitchers also have some control over sequencing of events. Specifically, some pitchers are better than others at pitching with runners on base.
There is a big leap in going from ERA to FIP. Instead of removing hit prevention and sequencing in one step, it might be better to remove one factor at a time. Bill James did that with his Component ERA (ERC). Applying the runs created methodology to pitchers, he determined what a pitcher's ERA should have been based on walks, hit batsmen, strikeouts, homers AND hits allowed. The runs created model is not used much anymore though and linear weights are better, so I wanted to find a similar statistic based on linear weights.
J.T. Jordan at Hardball Times got us part of the way there. He used the Baseball -Reference data on batting against pitchers to calculate wOBA against (or wOBAA). wOBAA for pitchers is calculated the same as wOBA for hitters. The MLB leaders for 2010 are shown in Table 1 below.
Table 1: MLB wOBAA Leaders in 2010
One good feature of wOBA is that it can easily be translated into runs above average (wRAA or RAA). To calculate RAA, subtract league average wOBAA from a player's wOBAA, divide by 1.25 (that number changes from year to year but is usually between 1.15 and 1.25) and multiply by plate appearances. The 2010 leaders are listed in Table 2. Cy Young Award winner Felix Hernandez tops the list at 45 RAA. This means that he saved his team an estimated 45 runs compared to the average pitcher.
Table 2: MLB RAA Leaders in 2010
So, we are almost there. All we need to do is turn RAA into an ERA. Here are the steps:
(1) Calculate MLB average runs scored per nine innings (4.44 in 2010)
(2) Subtract a pitcher' runs above average per nine innings pitched from the league average:
4.44- 9 x RAA/IP
(3) About 93% of runs are earned, so multiply the result in step (2) by .93. The final result is a linear weights component ERA. I'll call it WERC.
Table 3 shows that King Felix led the majors with a 2.60 WERC in 2010.
Table 3: MLB WERC Leaders in 2010
WERC is useful because it gives us an intermediate step between ERA and FIP. For example, Braves right-hander Tim Hudson had a big discrepancy between ERA (2.83) and FIP (4.09)
His WERC was 3.03 which is a lot closer to his ERA. This tells us that a large amount of the difference between FIP and ERA was due to batted balls in play rather than sequencing. We could have figured this out by examining other numbers such as BABIP and LOB%, but it's more convenient to compare three stats on the ERA scale.
Another example is National League Cy Young Award winner Roy Halladay. Halladay's ERA (2.44) was lower than his FIP (3.01). However, his WERC was 3.05 which tells us that the discrepancy was not due to hits allowed but rather sequencing.
I will apply these statistics to Tigers pitchers in a later post.
The limitations of ERA are well known in the blogosophere. Two of the biggest issues are:
(1) ERA gives pitchers full responsibility for all hits allowed despite the fact that their control over batted balls is limited. For example, a pitcher with a strong defense behind him will give up less hits (and thus fewer runs) than if he had a poor defense behind him.
(2) ERA gives pitchers full responsibility for sequencing of events. That is, it assumes that they can control when they give up hits and walks. For example, if a pitcher pitches extraordinarily well with runners in scoring position in a given year, he will have a lower ERA than if he had a typical year in those situations.
In reality, pitchers have limited control over both the number of hits they allow and sequencing of events. Thus, Defense Independent Pitching Statistics (DIPS) such as FIP, xFIP, tERA and SIERA have been developed to remove some of the noise of ERA. DIPS are based on things that pitchers do control for the most part - walks, hit batsmen, strikeouts, home runs and types of batted balls (ground balls , fly balls, line drives, pop flies).
Because they are based on things that pitchers essentially control, the DIPS metrics are said to be better measures of true talent than ERA. As a result, they are also better than ERA at predicting future performance. However, they only measure a portion of a pitcher's talent and should be used as complements to ERA rather than as replacements.
While pitchers do have less control over results of batted balls (hits and outs) than they do over walks, strikeouts, homers and ground balls, they do have SOME influence on results of batted balls. Some pitchers are indeed better than others at preventing hits on balls in play. Pitchers also have some control over sequencing of events. Specifically, some pitchers are better than others at pitching with runners on base.
There is a big leap in going from ERA to FIP. Instead of removing hit prevention and sequencing in one step, it might be better to remove one factor at a time. Bill James did that with his Component ERA (ERC). Applying the runs created methodology to pitchers, he determined what a pitcher's ERA should have been based on walks, hit batsmen, strikeouts, homers AND hits allowed. The runs created model is not used much anymore though and linear weights are better, so I wanted to find a similar statistic based on linear weights.
J.T. Jordan at Hardball Times got us part of the way there. He used the Baseball -Reference data on batting against pitchers to calculate wOBA against (or wOBAA). wOBAA for pitchers is calculated the same as wOBA for hitters. The MLB leaders for 2010 are shown in Table 1 below.
Table 1: MLB wOBAA Leaders in 2010
One good feature of wOBA is that it can easily be translated into runs above average (wRAA or RAA). To calculate RAA, subtract league average wOBAA from a player's wOBAA, divide by 1.25 (that number changes from year to year but is usually between 1.15 and 1.25) and multiply by plate appearances. The 2010 leaders are listed in Table 2. Cy Young Award winner Felix Hernandez tops the list at 45 RAA. This means that he saved his team an estimated 45 runs compared to the average pitcher.
Table 2: MLB RAA Leaders in 2010
So, we are almost there. All we need to do is turn RAA into an ERA. Here are the steps:
(1) Calculate MLB average runs scored per nine innings (4.44 in 2010)
(2) Subtract a pitcher' runs above average per nine innings pitched from the league average:
4.44- 9 x RAA/IP
(3) About 93% of runs are earned, so multiply the result in step (2) by .93. The final result is a linear weights component ERA. I'll call it WERC.
Table 3 shows that King Felix led the majors with a 2.60 WERC in 2010.
Table 3: MLB WERC Leaders in 2010
WERC is useful because it gives us an intermediate step between ERA and FIP. For example, Braves right-hander Tim Hudson had a big discrepancy between ERA (2.83) and FIP (4.09)
His WERC was 3.03 which is a lot closer to his ERA. This tells us that a large amount of the difference between FIP and ERA was due to batted balls in play rather than sequencing. We could have figured this out by examining other numbers such as BABIP and LOB%, but it's more convenient to compare three stats on the ERA scale.
Another example is National League Cy Young Award winner Roy Halladay. Halladay's ERA (2.44) was lower than his FIP (3.01). However, his WERC was 3.05 which tells us that the discrepancy was not due to hits allowed but rather sequencing.
I will apply these statistics to Tigers pitchers in a later post.
Subscribe to:
Posts (Atom)
