Archive for February, 2016

Hardball Retrospective – The “Original” 1906 Chicago Cubs

In “Hardball Retrospective: Evaluating Scouting and Development Outcomes for the Modern-Era Franchises”, I placed every ballplayer in the modern era (from 1901-present) on their original team. Consequently, Greg Luzinski is listed on the Phillies roster for the duration of his career while the Browns / Orioles declare Steve Finley and the Padres claim Derrek Lee. I calculated revised standings for every season based entirely on the performance of each team’s “original” players. I discuss every team’s “original” players and seasons at length along with organizational performance with respect to the Amateur Draft (or First-Year Player Draft), amateur free agent signings and other methods of player acquisition.  Season standings, WAR and Win Shares totals for the “original” teams are compared against the “actual” team results to assess each franchise’s scouting, development and general management skills.

Expanding on my research for the book, the following series of articles will reveal the finest single-season rosters for every Major League organization based on overall rankings in OWAR and OWS along with the general managers and scouting directors that constructed the teams. “Hardball Retrospective” is available in digital format on Amazon, Barnes and Noble, GooglePlay, iTunes and KoboBooks. The paperback edition is available on Amazon, Barnes and Noble and CreateSpace. Supplemental Statistics, Charts and Graphs along with a discussion forum are offered at TuataraSoftware.com.

Don Daglow (Intellivision World Series Major League Baseball, Earl Weaver Baseball, Tony LaRussa Baseball) contributed the foreword for Hardball Retrospective. The foreword and preview of my book are accessible here.

Terminology

OWAR – Wins Above Replacement for players on “original” teams

OWS – Win Shares for players on “original” teams

OPW% – Pythagorean Won-Loss record for the “original” teams

Assessment

The 1906 Chicago Cubs          OWAR: 58.8     OWS: 362     OPW%: .518

Based on the revised standings the “Original” 1906 Cubs finished fourth in a tight battle with the Giants, Cardinals and Pirates. Chicago topped the National League in OWS and OWAR.

Frank “The Peerless Leader” Chance supplied a .319 BA and led the circuit with 103 aces and 57 thefts. Frank “Wildfire” Schulte pilfered 25 bags and slashed a League-best 13 triples. Johnny Kling manufactured a .312 BA while the famous keystone combination of Johnny Evers and Joe Tinker collectively swiped 79 bases.

Hugh Duffy ranks twentieth in the All-Time Center Fielder rankings according to Bill James in “The New Bill James Historical Baseball Abstract.” Teammates listed in the “NBJHBA” top 100 rankings include Bill Dahlen (21st-SS), Chance (25th-1B), Evers (25th-2B), Tinker (33rd-SS), Bill Bradley (46th-3B), Kling (48th-C), Schulte (60th-RF) and Jim Delahanty (81st-2B).

LINEUP POS WAR WS
Joe Tinker SS 5.01 17.55
Johnny Evers 2B 4.95 19.46
Frank Chance 1B 7.36 33.26
Frank Schulte RF 3.33 23.94
Johnny Kling C 3.3 20.78
Jim Delahanty 3B 2.21 13.98
Bunk Congalton LF/RF 1.79 15.25
Davy Jones CF 0.69 12.35
BENCH POS WAR WS
Bill Dahlen SS 2.92 17.54
Larry Schlafly 2B 2.91 18.99
Frank Isbell 2B 2.19 25.91
Bill Bradley 3B 1.65 11.16
Tommy Raub C 0.66 2.84
George Moriarty 3B 0.14 5.97
Hugh Duffy -0.01 0
Tom Walsh C -0.01 0.01
Bill Phyle 3B -0.27 0.76
Germany Schaefer 2B -0.3 11.86
Malachi Kittridge C -0.55 0.58

Bob “Dusty” Rhoads (22-10, 1.80) delivered personal-bests in every major pitching category. Jack W. Taylor (20-12, 1.99) and “Tornado” Jake Weimer (20-12, 2.22) also surpassed the 20-win mark for the Cubbies. “Big” Ed Reulbach (19-4, 1.65) paced the Senior Circuit with a .826 winning percentage. Carl Lundgren added 17 victories and fashioned a 2.21 ERA.

ROTATION POS WAR WS
Jake Weimer SP 5.46 24.75
Bob Rhoads SP 4.77 23.12
Jack W. Taylor SP 4.67 25.42
Ed Reulbach SP 3.38 23.77
BULLPEN POS WAR WS
Carl Lundgren SP 2.07 18
Fred Glade SP 1.75 16.78
Fred Beebe SP 0.81 13.02
Big Jeff Pfeffer SP 0.67 16.13
Jack Doscher SP 0.35 1.11
Hub Knolls RP -0.16 0.38
Tom J. Hughes SP -1.84 3.32
Mal Eason SP -1.85 7.47

The “Original” 1906 Chicago Cubs roster

NAME POS WAR WS General Manager Scouting Director
Frank Chance 1B 7.36 33.26
Jake Weimer SP 5.46 24.75
Joe Tinker SS 5.01 17.55
Johnny Evers 2B 4.95 19.46
Bob Rhoads SP 4.77 23.12
Jack Taylor SP 4.67 25.42
Ed Reulbach SP 3.38 23.77
Frank Schulte RF 3.33 23.94
Johnny Kling C 3.3 20.78
Bill Dahlen SS 2.92 17.54
Larry Schlafly 2B 2.91 18.99
Jim Delahanty 3B 2.21 13.98
Frank Isbell 2B 2.19 25.91
Carl Lundgren SP 2.07 18
Bunk Congalton RF 1.79 15.25
Fred Glade SP 1.75 16.78
Bill Bradley 3B 1.65 11.16
Fred Beebe SP 0.81 13.02
Davy Jones CF 0.69 12.35
Big Jeff Pfeffer SP 0.67 16.13
Tommy Raub C 0.66 2.84
Jack Doscher SP 0.35 1.11
George Moriarty 3B 0.14 5.97
Hugh Duffy -0.01 0
Tom Walsh C -0.01 0.01
Hub Knolls RP -0.16 0.38
Bill Phyle 3B -0.27 0.76
Germany Schaefer 2B -0.3 11.86
Malachi Kittridge C -0.55 0.58
Tom Hughes SP -1.84 3.32
Mal Eason SP -1.85 7.47

Honorable Mention

The “Original” 1945 Cubs      OWAR: 50.4     OWS: 307     OPW%: .654

The Cubs (101-53) eclipsed the century mark in victories to secure the pennant and amassed a comfortable lead in OWAR and OWS. Phil Cavarretta (.355/6/97) merited 1945 National League MVP honors while topping the circuit in batting average and OBP (.449). “Smiling” Stan Hack scored 110 runs and supplied career-bests with a .323 BA and 99 bases on balls. Augie Galan (.307/9/92) coaxed 114 walks and registered 114 tallies. “Handy” Andy Pafko (.298/12/110) established personal-bests in RBI and triples (12). Hank Wyse (22-10, 2.68) completed 23 of 34 starts and Harry “The Cat” Brecheen (15-4, 2.52) contributed a league-best .789 winning percentage.

On Deck

The “Original” 1980 Astros

References and Resources

Baseball America – Executive Database

Baseball-Reference

James, Bill. The New Bill James Historical Baseball Abstract. New York, NY.: The Free Press, 2001. Print.

James, Bill, with Jim Henzler. Win Shares. Morton Grove, Ill.: STATS, 2002. Print.

Retrosheet – Transactions Database

Seamheads – Baseball Gauge

Sean Lahman Baseball Archive


Adjusted Mayberry Method Ratings for NL

The last couple years, I have calculated the Mayberry Method for the NL from the Baseball Forecaster Book from Ron Shandler. The Mayberry Method is based on speed, power, batting average, and role (PAs). A score from 0 to 5 is assigned for each of the three skill categories and added and multiplied to the role score. After the “raw” score is determined, it is multiplied by three reliability factors.

When I determine the total score for each player, I split them up into positions. Next, I take the 30th percentile score (widely considered as the replacement level) and subtract each score from the replacement-level score for that position.

For example, Christian Yelich had a 2 score for power, a 4 score for speed and batting average, and a 5 score for role. One would add (2+4+4+5)=15. Then I multiplied 15 by 5 (the role score) and got 75. Next I multiplied 75 by the three reliability scores which were 1.05, 1.05, and 1.1 to get 90.95 as the total score. Because, Yelich was an outfielder, his score was subtracted by the replacement level score for outfielder of 63. So, Yelich’s final score was 27.95, good for 25th overall. Without much further ado, here are the rankings.

The first column is the total score, the second column is the name and position, the third column the number equivalent of the position (Catcher 2, 1B 3, 2B 4 and so forth), the final column is the position adjusted Mayberry score. Notice four out of the five players are second basemen (the other is Starling Marte an elite power-speed-BA player). Note for fantasy players in NL-only leagues, second basemen make up seven out of the top-20 but are far-and-few-between lower than that.

88.935 Daniel Murphy 2B 4 52.635
88.935 Ben Zobrist 2B 4 52.635
113.135 Starling Marte OF 7 50.135
80.85 DJ LeMahieu 2B 4 44.55
78.65 Howie Kendrick 2B 4 42.35
97.02 Paul Goldschmidt 1B 3 42.02
73.5 Anthony Rendon 2B 4 37.2
86.82188 Nolan Arenado 3B 5 36.82188
96.8 Andrew McCutchen OF 7 33.8
82.5825 Todd Frazier 3B 5 32.5825
80.85 Jacob Realmuto C 2 31.85
68.07938 Joe Panik 2B 4 31.77938
81.675 Corey Seager 3B 5 31.675
86.515 Adrian Gonzalez 1B 3 31.515
94.5 A.J Pollock OF 7 31.5
79.86 Buster Posey C 2 30.86
93.7125 Ryan Bruan OF 7 30.7125
80.465 Matt Carpenter 3B 5 30.465
80.465 Matt Duffy 3B 5 30.465
66.70125 Brandon Phillips 2B 4 30.40125
93.17 Justin Upton OF 7 30.17
84.7 Anthony Rizzo 1B 3 29.7
92.4 Charlie Blackmon OF 7 29.4
90.95625 Ben Revere OF 7 27.95625
90.95625 Christian Yelich OF 7 27.95625
82.5825 Ian Desmond SS 6 26.5825
82.5825 Jimmy Rollins SS 6 26.5825
82.5825 Jean Segura SS 6 26.5825
88.935 Dexter Fowler OF 7 25.935
75.24563 Neil Walker 2B 5 25.24563
61.425 Josh Harrison 2B 4 25.125
80.85 Dee Gordan SS 6 24.85
60.5 Wilmer Flores 2B SS 4 24.2
78.82875 Freddie Freeman 1B 3 23.82875
86.82188 Edgar Inciarte OF 7 23.82188
72.765 Wellington Castillo C 2 23.765
86.625 Norichika Aoki OF 7 23.625
73.31625 Corey Spagenberg 2B 5 23.31625
78.82875 Brandon Crawford SS 6 22.82875
72.765 Yangervis Solerte 3B 5 22.765
77.175 Adam Lind 1B 3 22.175
84.8925 Jason Heyward OF 7 21.8925
84.8925 Seth Smith OF 7 21.8925
71.5 Kris Bryant 3B 5 21.5
69.3 Derek Norris C 2 20.3
69.8775 Martin Prado 3B 5 19.8775
82.6875 Khris Davis OF 7 19.6875
82.5825 Gregory Polanco OF 7 19.5825
82.5 Stephen Piscotty OF 7 19.5
68.25 Jonathan Lucroy C 2 19.25
80.85 Odubel Herrera OF 7 17.85
72.765 Pedro Alverez 1B 3 17.765
80.325 Bryce Harper OF 7 17.325
66.15 Justin Turner 3B 5 16.15
65.835 Yasmany Tomas 3B 5 15.835
78.75 Chris Coghlan OF 7 15.75
78.75 Yasiel Puig OF 7 15.75
71.6625 Chris Owings SS 6 15.6625
71.6625 Jose Reyes SS 6 15.6625
78.65 Jay Bruce OF 7 15.65
51.7275 Jace Peterson 2B 4 15.4275
51.7275 Javier Baez 2B 4 15.4275
70 Brandon Belt 1B 3 15
69.4575 Lucas Duda 1B 3 14.4575
77.175 Curtis Granderson OF 7 14.175
77.175 Marcell Ozuna OF 7 14.175
69.3 Eugenio Suarez SS 6 13.3
75.24 David Peralta OF 7 12.24
66.825 Joey Votto 1B 3 11.825
61.425 Maikel Franco 3B 5 11.425
66 Ryan Zimmerman 1B 3 11
66.70125 Adeiny Hechavarria SS 6 10.70125
60.5 Yunel Escobar 3B 5 10.5
73.205 Nick Markakis OF 7 10.205
72.765 Marlon Byrd OF 7 9.765
71.32125 Andre Ethier OF 7 8.32125
70.875 Matt Kemp OF 7 7.875
57.60563 Jacob Lamb 3B 5 7.605625
70.56 Hunter Pence OF 7 7.56
69.825 Randal Grichuk OF 7 6.825
62.37 Jung-Ho Kang SS 6 6.37
55.2825 Travis D’Arnaud C 2 6.2825
55.125 Yadier Molina C 2 6.125
68.4 Corey Dickerson OF 7 5.4
66.70125 Billy Hamilton OF 7 3.70125
40 Jedd Gyorko 2B 4 3.7
52.25 Nick Hundley C 2 3.25
53.24 Kolten Wong 2B 5 3.24
66.15 Denard Span OF 7 3.15
65.835 Michael Taylor OF 7 2.835
57.1725 Addison Russell SS 6 1.1725
49.37625 Miguel Montero C 2 0.37625
49.005 Kyle Schwarber C 2 0.005
36.3 Cesar Hernandez 2B 4 0
63 Carlos Gonzalez RF 7 0
63 Giancarlo Stanton OF 7 0
54.57375 Brandon Moss 1B OF 3 -0.42625
62.37 Michael Conforto OF 7 -0.63
55.125 Freddy Galvis SS 6 -0.875
55.125 Jordy Mercer SS 6 -0.875
60.6375 Joc Pederson OF 7 -2.3625
33.075 Kiki Hernandez 2B 4 -3.225
33 Danny Espinosa 2B 4 -3.3
33 Scooter Gennett 2B 4 -3.3
59.535 Matt Holiday OF 7 -3.465
44.55 Hector Olivera 3B 5 -5.45
49.5 Justin Bour 1B 3 -5.5
49.6125 Ruden Tejada SS 6 -6.3875
43.2 David Wright 3B 5 -6.8
55.2825 Jorge Soler OF 7 -7.7175
39.9 Yasmani Grandal C 2 -9.1
53.865 Cameron Maybin OF 7 -9.135
26.73 Jose Paraza 2B 4 -9.57
45.6 Zack Cozart SS 6 -10.4
37.8 Wilson Ramos C 2 -11.2
43.32 Wil Myers OF 1B 3 -11.68
45.36 Jayson Werth LF 7 -17.64
36.3 Ben Paulsen 1B 3 -18.7
29.7 Brandon Drury 3B 5 -20.3
29.62575 Derek Dietrich OF 3B 5 -20.3743
29.04 Cody Asche OF 3B 5 -20.96
38.115 Gregor Blanco OF 7 -24.885
36.6795 Aaron Altherr OF 7 -26.3205
32.67 Travis Janikowski OF 7 -30.33
28.728 Carl Crawford OF 7 -34.272
28.08 Michael Cuddyer OF 7 -34.92
26.46 Dominic Brown OF 7 -36.54

A More Appropriate Measure of Late-Inning Relievers

The issue that plagues the valuation of late-inning relievers is the generalized treatment of runs.

WAR is the most accepted player evaluation metric and wins are determined by run value. Run value is determined in a generalized sense; it’s too perilous and unwieldy to predict, or evaluate performance, based upon the sequencing of events.

However, late-inning relievers do not pitch in a general situation. Unlike many other players we know when they will perform. They are unique; they pitch in particular situations: the late innings of a baseball game.

They are not vulnerable to give up a home run in a large range of innings like a starting pitcher. They are vulnerable to giving up runs in the innings their role demands them to appear in; most notably the 7th, 8th, and 9th innings.

Therefore, reliever value should be measured by a more specific run value. This run value, and ultimately win value, cannot be measured in a general sense. Their valuation must account for the specific times they appear in a game.

I set out to do this with those principles in mind.

First, I used Baseball Reference’s Play Index to determine the amount of runs scored in between the 7th and 9th innings of all games in 2015. There were 13,448 7th, 8th, and 9th innings played last year. That is the equivalent of 1,494 full 9 inning baseball games. In sum, 5,968 runs were scored in the 7th, 8th, and 9th innings of baseball games in 2015. On average, that is 3.99 runs per “game”, where “game” signifies 9 innings of 7th-9th inning performance.

Second, also using Baseball Reference’s Play Index, I looked at the 300 pitchers with the most appearances in the 7th, 8th, and 9th innings. This does not represent every pitcher that pitched in the 7th, 8th, and 9th inning, but it gets us to Trevor Cahill, who pitched 16 innings in the 7th inning or later.

I then split this list of pitchers into two groups. Theoretically, the 90 best relievers in the league would be pitching in the 7th, 8th, and 9th innings (30 teams; 3 relievers each). Therefore, the first group is the first 91 pitchers with the most appearance (Tony Sipp and Blaine Boyer each appeared in 43 innings between the 7th and 9th innings, so there is one more than 90 in this case). The other 209 pitchers represent the “replacement” pool.

The average performance of the “replacement” pool was taken to determine the performance of a replacement player. Here is what that looks like:

This is the basis for the more nuanced portions of the calculation. 3.99 runs were scored in the 7th-9th inning of MLB games in 2015, on average. The first thing to do is calculate the Runs Per Win (RPW) in the “game” (the 7th-9th inning game).

Dave Cameron explains how RPW for pitchers is calculated in this post in the FanGraphs Glossary. You should read it in to become acclimated with the logic of the next step. The article notes that the WAR calculations at FanGraphs credit each pitcher with a unique RPW value, as the better or worse a pitcher is will lower or raise their RPW value. It then details the calculation recommended by Tom Tango to determine RPW value:

Runs Per Win = (Player Runs Against + Lg Runs Against)/2)*1.5

I’m using FIP for the Players Runs Against for this explanation, but you could simply use RA9 or ERA. The tables below include an ERA-based WAR calculation in addition to a FIP-based WAR calculation. That’s not the main point of this conversation though.

So, I’ll take the 3.83 FIP of the replacement-level pitcher and the 3.99 League Runs Against Average and plug it into that equation, which equates to 5.86 RPW for the replacement-level 7th-9th inning pitcher. This equation is applied to each individual pitcher. I’ll use Aroldis Chapman throughout the explanation to walk through the calculation.

Replacement Pitcher RPW = (3.83 + 3.99)/2)*1.5 = 5.86 RPW

Aroldis Chapman RPW = (1.95 + 3.99)/2)*1.5 = 4.45 RPW

Next, I made a calculation of runs above average for each pitcher and the average of the replacement pool. Again, the most important numbers in this calculation is the FIP of the individual pitchers and the 3.99 league average. These figure are plugged into the following calculation:

Runs Above/Below Average = (Lg Runs Against*(Player IP/9))-(Player FIP*(PlayerIP/9)

Replacement Pitching Runs Above/Below Average = (3.99*(26.2/9))-(3.83*(26.2/9) = .49 Runs Above Average

Aroldis Chapman Runs Above/Below Average = (3.99*(63.1/9))-(1.95*(63.1/9) = 14.33 Runs Above Average

The replacement pool was .49 runs above league average. The replacement pool averaged 26.2 innings pitched, or roughly three “games” per year. The replacement player would give up 11.48 runs a year over 26.2 innings based on a 3.83 FIP, which is .49 runs less than the 11.97 runs of the 3.99 league average over the same amount of innings. This calculation was done for each player. Chapman is given as an example above.

Finally, the Replacement Runs Above/Below Average is subtracted from Runs Above/Below Average for each individual player. The difference between the two is then divided by each player’s unique RPW value and the result is each pitcher’s WAR. For example, the difference between Chapman’s Runs Above Average and the Replacement Player’s Runs Above Average is 13.85. Chapman’s unique RPW is 4.45. This values Chapman at 3.11 WAR.

WAR = (Player Runs Above/Below Average — Replacement Runs Above/Below Average) / Player Unique RPW Value

(14.33-.49) = 13.84;

13.84/4.45 = 3.11 WAR

Before you glance at the tables below let me set out some more facts about the data:

  • The list of 300 pitchers does include starters who appeared in innings 7–9.
  • The list does not include every pitcher who appeared in innings 7–9 so the values in the chart are not exact. The exercise is meant to display the idea of an improved method to measure reliever value. My assumption would be that a more complete list would lead to an inferior measure of replacement.
  • The data is only looking at 7th-9th inning performance. It does not account for performance in extra innings, or performance prior to the 7th inning.
  • WAR is a counting stat, so WAR will be influenced by the amount of innings each player pitches.
  • The median calculated FIP WAR is .21 and the Average FIP WAR is .35. The 25th Percentile ranges from -1.67 to -.81. The 75th Percentile ranges from .71 to 3.2.
  • The median calculated ERA WAR is .26 and the Average ERA WAR is .5. The 25th Percentile ranges from -1.54 to -.27. The 75th Percentile ranges from 1.08 to 5.72.

 

 

 


Andruw Jones and Ken Griffey Jr.

Andruw Jones is likely to announce his retirement from Major League Baseball sometime in the very near future. Jones hasn’t been on the MLB radar since his last season in the big leagues back in 2012, when he played 94 games with the New York Yankees but hit just .197/.294/.408. He played 2013 and 2014 with the Tohoku Rakuten Golden Eagles in the Japan Pacific League and hit 26 and 24 home runs, while combining to hit .232/.393/.441. He’ll turn 39 years old in April, so he is likely to hang up his spikes after a 17-year Major League career.

In this column at FanGraphs, David Laurila made an apt comparison between Jones and Jim Edmonds with these numbers showing the similarity:

 

.254/.337/.486, 1933 hits, 434 HR, 10 Gold Gloves, 67.1 WAR—Andruw Jones

.284/.376/.527, 1949 hits, 393 HR, 8 Gold Gloves, 64.5 WAR—Jim Edmonds

 

It’s a good comparison. They were nearly equal in value in their careers and both hit many home runs and won numerous Gold Gloves.

Another interesting player to compare Jones to is more similar when you look at the arc of their careers. Both came up to the big leagues at the age of 19 and were very good players until the age of 30, then experienced a significant drop-off in value from that point on. That other player is Ken Griffey Jr. More on him later.

Andruw Jones came up with the Atlanta Braves in 1996, making his Major League debut on August 15th. He only hit .217/.265/.443 in 31 games in his rookie year but helped the Braves make it to the World Series. He hit two home runs in Game 1 against the Yankees, becoming the youngest player to ever hit a home run in the World Series. The Braves lost the series four games to two, but Jones hit .400/.500/.750 and made his presence known on a national stage.

Jones established himself in center field for the Braves in 1997 at the age of 20. He hit .231/.329/.416, which was below average for a hitter in an era of high offense (96 wRC+), but he was so good defensively that he was worth 3.7 Wins Above Replacement. The following year was the first in an impressive stretch of nine seasons from 1998 to 2006 during which Jones averaged 6.4 WAR per year. Not only did he excel on defense during this nine-year stretch, he averaged 35 home runs per season, 99 runs scored, 104 RBI, 12 steals, and a .270/.347/.513 batting line (119 wRC+). He was a five-time All-Star and won nine straight Gold Glove Awards (he would win a 10th in a row the next year). If Jones had played in the first part of the 20th century, his nickname might have been “Death to Flying Things.” Instead, he was just Andruw Jones. Jones’ best season was a 7.9 WAR year in 2005 when he hit .263/.347/.575 with 95 R, 51 HR, 128 RBI and finished second in the voting for National League MVP. This stretch was the essence of Andruw Jones—a power-hitting center fielder with 35 home runs a year and terrific defense. There were only four players in baseball worth more WAR during this nine-year stretch: Barry Bonds, Alex Rodriguez, Randy Johnson, and Pedro Martinez.

Jones was an above-average player again in 2007. He was worth 3.3 WAR thanks primarily to still excellent defense. His hitting dropped off considerably, though. After hitting .262/.355/.553 with a combined 92 home runs over the two previous seasons, Jones hit just .222/.311/.413 in 2007. His 26 home runs were his lowest total since 1999. This would be his last season in Atlanta and his last season with a WAR above 2.0. It was also his last excellent season on defense. Jones would play with four different teams over the final five years of his Major League career and hit .210/.316/.424. His once-great defense dropped off precipitously and he averaged just 0.6 WAR per season.

Those last five journeyman years for Jones could make it hard for people to remember how great he was in the first part of his career. Through the first seven years of his career, Andruw Jones was nearly the equal of Ken Griffey Jr. Both Jones and Junior reached the Major Leagues as 19-year-olds and were power-hitting center fielders. Griffey started winning Gold Glove Awards in his second year in the bigs and won nine Gold Gloves over the next 10 years. Jones won his first of 10 consecutive Gold Glove Awards in his third year in the Major Leagues. While both were considered good fielders, the truth was that Jones was significantly better than Junior for an extended period of time and held more of his defensive value in the latter years of his career. Jones was an elite fielder through his age-30 season, then became more of a slightly-below-league-average fielder in his last five years. Griffey, on the other hand, was rarely at the elite level as a fielder that Jones reached and when he declined, it was a significant decline to well-below-average defense in his late 30s.

Griffey was the better hitter, of course, but in terms of overall value, they were very close into their mid-20s. The chart below shows each player’s cumulative WAR by age. Griffey’s WAR advantage after each player’s first seven years was slim, just 38.2 to 36.5.

In a similar number of plate appearances, Jones and Griffey had a similar number of home runs, runs scored, and RBI. Griffey had a significant edge in batting average, on-base percentage, and slugging percentage. Jones was much better on defense. As mentioned above, they were very close in overall value.

Griffey took his game to another level in his age 26 and age 27 seasons, when he averaged 9.4 WAR per year while hitting 105 home runs and slugging .637. Jones averaged 5.2 WAR in his age 26 and 27 seasons, which is great — just not at the same level as Griffey.

The five-year stretch of seasons when Jones and Griffey were 26 through 30 years old makes up the bulk of the difference in career WAR between the two players. During this stretch of ages, Jones accumulated 27.7 WAR and Griffey had 35.6. Again, Griffey was a much better hitter, with a significant edge in average, on-base percentage, and slugging percentage, along with a large edge in runs, home runs, and RBI. Jones made up some of that difference with his still excellent defense.

This is not to say that Jones wasn’t an elite player. He was. Over the five-year stretch from age 26 to 30 (2003 to 2007), Andruw Jones was seventh in baseball in WAR.

If you expand the range to the first 12 years of his career, from 1996 to 2007, Andrus Jones was also seventh in baseball in WAR, behind Barry Bonds, Alex Rodriguez, Chipper Jones, Pedro Martinez, Randy Johnson, and Curt Schilling. In his first 12 seasons, Andruw Jones averaged 87 runs scored, 31 home runs, 93 RBI, and a .263/.342/.497 batting line with excellent defense.

And that was it. Those first 12 seasons make up nearly 96% of Jones’ career WAR even though he continued to play for another five years. He signed with the Dodgers as a free agent prior to the 2008 season and had the worst year of his career. He hit .158/.254/.249 and his defense went from excellent to average. His WAR for that season was -1.1. He rebounded on the hitting side over the next three seasons but was no longer the defensive stud he’d once been and became a part-time player. Over his last five seasons, he was worth just 2.9 WAR total.

Of course, Ken Griffey Jr. did not age well either. He was injured in 2001 at the age of 31 and finished with the lowest WAR of his career to that point (1.8). From 2002 to 2004, he played an average of just under 70 games per year and had 0.5 WAR per season. He continued to hit well (117 wRC+), but on defense he struggled. From 2004 to 2009, no outfielder in baseball with more than 2000 innings in the field had a worse Ultimate Zone Rating (UZR) than Griffey. He was even worse than Manny Ramirez and Adam Dunn.

The graph shown earlier reveals the similar arcs of the careers of Andruw Jones and Ken Griffey, Jr. They both were great players through the age of 30 and below average players from age 31 on. Griffey did have more truly elite seasons. He had three seasons with eight or more WAR, which were better than any season Jones had, but they were very close in the number of seasons with four or more WAR (Griffey had 10, Jones had 9).

It will be interesting to see what Hall of Fame voters think of Andruw Jones in five years. Admittedly, there was a 10-WAR difference between Jones and Griffey over the course of their careers, but they don’t seem all that different when you look at their similar career trajectories and their distribution of WAR, particularly in the number of great seasons they each had. Jones played 17 years, while Griffey played 22. But in Griffey’s final five years, his value was below replacement level. It didn’t seem like it because he hit a respectable-looking .247/.340/.444 and had nearly 500 hits and almost 100 home runs, but his defense was a killer that greatly affected his value.

Ken Griffey, Jr. was just voted into the Hall of Fame with 99.3% of the vote, the highest percentage ever. Jim Edmonds was on the same ballot and is now one-and-done with just 2.5% of the vote. How will Andruw Jones fare?


Squeezing a Little More Out of Ryan Vogelsong

The Pirates brought in Ryan Vogelsong this winter, and most Pirates fans believed it was a depth move, and that another move would follow and net them a better option. Recent comments by GM Neal Huntington hint that perhaps they are done, and that Pirate fans should resign themselves to seeing Vogelsong as the #5 starter, at least to start the season.

Vogelsong has not been good for a while. In 2011-12 he threw 369 innings for the Giants with a 3.68 FIP, adding 4.6 WAR.  Since then, 423.1 very mediocre IP with 4.33 FIP generating 0.8 WAR.  Steamer projects 109 IP at 4.38 FIP & 0.6 WAR.

With that in mind, I decided to do a little keyboard coaching to find a path to improvement.

The first unusual thing I noticed on Vogelsong’s Brooks Baseball card is that he throws five pitches in fairly equal proportion:

Let’s look at RHB first. In 2015, he was average, his wOBA against sitting right at .300. Here is the SLG against by pitch for the last five seasons:

His 4-seam, sinker, cutter, and curve all yielded good-to-excellent SLG, yet in 2015 the changeup got hammered, to the tune of .563 SLG (up from the .370 range), and he got less than 5% whiffs (down from 7% previously, and far below the league average of 11.9%). Yet he still uses it 5% of the time. The curve on the other hand, has been good, producing a SLG under .300 in four of five years, and GB% and SwStr% right around average. I’d suggest it’s time to ditch the change against RHB, and rely more on the curve.

Against LHB, he gets torched. Batters mashed a .383 wOBA against him in 2015 and a .346 wOBA in 2014. Here are the numbers by pitch:

His only pitches with decent SLG against in 2015 were the 4-seam (.390) and the cutter (.265), while the other three pitches all have SLG over .630. (The curve at least has a 16% whiff rate). Those three “bad” pitches are used over 60% of the time. While I’m sure he needs to mix those pitches them in sometimes to keep lefty hitters honest, they are simply getting destroyed. So, I suggest he could have some more success if he stopped trying to throw his sinker and changeup — or at least cut way back — and occasionally mixed in the curve as the pitch to keep them honest.

While Vogelsong’s problems likely require a solution more sophisticated than “throw your bad pitches less” (and undoubtedly his coaches have a better view on this than some guy behind a monitor looking at numbers), his recent results suggest he won’t be much above replacement level unless he changes *something* vs. LHB. Obviously, resurrecting his once-good changeup would be the preferred option, but failing that: ditch the changeup and stop throwing the sinker to lefties. Boost 4-seam and cutter usage against lefties, and mix in the curve to keep them honest. Maybe a 4-seam/cutter/curve mix would be enough to get him through the order twice, and if not, his results with his “good” pitches may be good enough for the bullpen.


Visualizing and Quantifying Strikes Zone Changes Over Time

This week the strike zone has been getting a lot of attention. If you’ve been paying any attention to baseball (and I’m sure you have since fantasy baseball leagues are starting to open up) there have been a few articles/releases suggesting that MLB may be considering raising the strike zone from the hollow beneath the kneecap to the top of the kneecap. It seems like a good idea since strikeout rates are on the rise, but was this a result of (1) pitchers getting better or (2) hitters getting worse or (3) have strikes been getting called differently? I’ll give you a hint; it’s neither of the first two suggestions, at least not directly. No, instead let’s focus on the strike zone and more specifically two things: (1) visualizing the strike zone from 2008 to 2015 and (2) using a standardized set of pitches look at how those pitches have been called over time.

Let’s go through the methods I used before we get to the plots. I used the pitchRx package in R to gather and store the data and used many of the functions included in the package. Next I went through the data and subset the PITCHf/x data by year since I was interested in looking at annual changes. Now due to a combination of time restraints and lack of computing power I didn’t run all of the pitches thrown in each year so I did some subsetting instead. I downloaded a CSV from the FanGraphs leaderboards of all qualified pitchers from 2008 to 2015. In each year I randomly selected 20 pitchers from the list of qualified starters to represent how the strike zone was called for that given year. Finally I ran the data through a general additive model (seen here) which was used to create the “heat maps” for the probability of called strikes in the plots below. I also tested the probability of five standard pitches being called strikes, but that is addressed a bit more later one so I won’t bore you with the details twice. Added note: if anyone actually wants a copy of the R code leave a comment below and I’ll get in contact with you.

Below I’ve included a GIF of the strike zone from 2008 to 2015 . If you watch it a few times you’ll begin to notice the gradual changes to the bottom of the strike zone, plus when it flips from 2015 to 2008 you can really notice the difference. It’s not surprising that there are inter-annual differences between the zones since I’m sure MLB makes a few minor tweaks every off-season and maybe there is a changing of the guard over time for the umps. I also need to apologize about the 2010 plot, the left (L) and right (R) are reversed and I can’t seem to switch them. We will just have to deal with that one plot being different. In all plots the label “L” refers to left-handed batters and “R” to right handed batters.

Now I wanted to find a way to quantify changes to how pitches were being called and I decided on using a set of standardized pitches. Below is a plot showing the locations I chose for my test pitches. I went with five different locations. The pitch right down the middle was my control of sorts, just to make sure things were getting called consistently over time. The remaining locations were the ones I was really interested about; three of those pitches were all located on the lower edge of the strike zone and the final pitch was located 0.2 feet or 2.4″ (the metric system would be more useful here, just sayin’) below the bottom edge of the strike zone. When I initially began this simulation I expected that the lowest pitch would be a second control pitch that would consistently be called a ball, but the results were pretty surprising. Also, I’d like to include that the strike zone to lefties is slightly shifted so that more outside pitches are called strikes.

OK so we are almost at the exciting conclusion. Using those standardized pitches from the plot above I used the general additive model to predict the probability of that pitch being called a strike in a given year. The results are summarized in the plot below. We can see that the pitch being thrown at coordinates 0, 2.5 (the one down the middle) the probability of being called a strike is basically 100% every year. Well that’s a good thing at least that call is consistent. The low pitch thrown down the middle on the bottom edge of the strike zone, coordinates 0, 1.7 (green line), has increasingly been called strike since 2008 to both right- and left-handed batters. Pitches down and in to righties increased pretty significantly this past season where the probability crept above 50%; to lefties that pitch is down and away and it’s been called pretty consistently since 2011 (red lines). Pitches thrown down and away to righties or down and in to lefties (coordinates 1, 1.7 — purple lines) haven’t changed all that much over the time period.

Now we get to what I think is the most interesting pitch. The low fastball down the middle (coordinates 0, 1.5) the one that should be out of the strike zone. This pitch is represented by the gold/yellow lines on the plots. In 2008 these pitches had a chance of being called a strike ~10% of the time to both righties and lefties. Over the last eight seasons that number has trended upwards and in the 2015 season settles in somewhere around 36-40%, which is not an insignificant proportion.

Based on this data it certainly appears as though MLB is justified into looking at raising the strike zone. Pitchers that live down in the zone have been given an increasing advantage in a relatively short amount of time. Hopefully this sheds some light onto the debate on whether or not to raise the strike zone in the coming seasons or maybe the umps will be able to make some adjustments for the upcoming season.


Pittsburgh’s Next Reclamation Project

During the past three seasons in Pittsburgh, Ray Searage has worked his magic to rejuvenate the careers of struggling pitchers. From increasing the usage of a two-seam fastball to induce ground balls and having a pitch framing expert behind the dish, the Pirates rotation has raised a few eyebrows. A few key pitchers that benefited from Searage were AJ BurnettFrancisco LirianoEdinson VolquezJA Happ, and many more. After seeing the success of these four pitchers, it becomes very difficult to doubt a pitching acquisition made by the Pirates. Therefore, who will be Searage’s next project?

On December 9, the Pirates agreed to send soon-to-be-free-agent second basemen Neil Walker to the Mets in exchange for veteran left-hander Jon Niese. Niece was drafted by the Mets in the 7th round in 2005 out of Defiance High School in western Ohio. Since 2010, he has been a consistent innings eater for the Mets rotation known for inducing a ton of ground balls. In 2012, Niese sported a 13-9 record with a 3.40 ERA and a career-high 2.6 WAR. From 2010 to 2014, Niese was consistently a 2 WAR pitcher, which would project as an average to above-average mid-rotation starter. However, in 2015, he struggled at times and posted a career low 0.9 WAR. Even though he posted a career high 55% ground ball percentage, he was not missing bats much with his 5.8 K/9. It’s safe to say that Niese is seeking a rebound in 2016 and he has come to the right place.

Heading into the 2016 season, I am very high on Jon Niese and believe he fits perfectly in a Pirates rotation managed by Ray Searage. Niese has a repertoire that includes a sinker, cutter, and four-seam fastball that will induce many ground balls. With their statistical findings on defensive alignments outlined in Big Data Baseball by Travis Sawchik, the Pirates could use Niese to their advantage. When looking into some of Niese’s pitch-usage data, I found his situation comparable to that of J.A. Happ. After acquiring a struggling Happ from Seattle during last summer’s trade deadline, Searage noticed a decrease in the usage of his fastball and encouraged him to be more aggressive. Happ adjusted his approach almost immediately and put up an impressive 7-2 record with a 1.85 ERA in 11 starts. So how does Jon Niese’s struggles compare to Happ’s? I found my answer after referring to brooksbaseball.net for pitch-usage data in his 2012 season and 2015 season. In 2012, Niese threw his four-seam fastball at 35.7 percent compared to 20.2 percent in 2015. This is a significant difference in a matter of only four seasons. I would also like to note that he reduced his cutter usage by almost 7 percent in that time span.

Upon his return to Pittsburgh, I am expecting Searage to take a similar approach with Niese as he did with Happ. Increasing the fastball usage and being more aggressive will only benefit Niese with an even better defense behind him. Steamers projects Niese to repeat at a 5.8 K/9 in 2016. However, Fans projections sees him returning to a 6.4 K/9. Let’s not forget that he is throwing to pitch-framing extraordinaire Francisco Cervelli, which may work in his favor to get more strikes. While I believe that he will be able to miss a few more bats than last year, his main strength will be pitching to contact and inducing ground balls into the many defensive alignments behind him.

While the Pirates’ projected rotation may seem a bit top-heavy at the moment, look for Niese to be a solid #3 behind Gerrit Cole and Francisco Liriano. By mid season, the Pirates rotation could be a force with the debuts of top prospect Tyler Glasnow and former second overall draft pick Jameson Taillon. In October, while many may disagree now, watch for the Pirates to be declared the winner of the offseason swap with the Mets.