1871-2017 Defense Plus/Minus System

[Garlon]

I was able to take a copy of the baseball databank/Lahman DB and update it with LF/CF/RF data from 1871-1953 as well as defensive innings played for every position 1871-1953. Using the formulas from Bill James in The Fielding Bible (2006) I was able to calculate Plus/Minus defensive plays made at every position every season.

I have +/- Errors for every position and these categories and their Defensive Run Values according to The Fielding Bible formulas:

Catchers: PB +/-, CS +/- 0.62 each
Infielders: A +/-, DP +/- A = 0.76, DP = 0.15
Outfielders: PO +/-, A +/- 0.56 each

Here are the top defensive players at each position in the 1990's for each season:

NL, AL

C
1990: Pagnozzi, Parrish
1991: Pagnozzi, Rodriguez
1992: Manwaring, Rodriguez
1993: Wilkins/Manwaring, Karkovice
1994: Santiago, Steinbach
1995: Ausmus, Rodriguez/Steinbach
1996: Johnson, Rodriguez
1997: Johnson, Rodriguez
1998: Fabregas/Ausmus, Rodriguez
1999: Castillo, Rodriguez


1B
1990: Grace, Quintana
1991: Grace, Quintana/Stubbs
1992: Bagwell, Joyner
1993: Karros, Joyner
1994: Bagwell, Fielder (?)
1995: Bagwell, Joyner/Palmeiro
1996: Bagwell, Stahoviak
1997: Conine, King
1998: Helton, Segui
1999: Snow, McGriff


2B:
1990: Thompson/DeShields, Gallego
1991: Oquendo, Randolph
1992: Sandberg, Bordick
1993: Thompson, Fletcher
1994: Barberie, Alomar
1995: Sanchez, Gates
1996: Lemke, McLemore
1997: Lemke, John Valentin
1998: Veras, Grebeck
1999: Reese, Bush

3B:
1990: Hayes, Gaetti
1991: Caminiti, Pagliarulo
1992: Wallach, Ventura
1993: Caminiti, Boggs
1994: Williams, Cooper
1995: Williams/Caminiti, Fryman
1996: Castilla, Arias/Fryman
1997: Rolen, Cirillo
1998: Andrews, John Valentin
1999: Boone/Rolen/Ventura, John Valentin

SS:
1990: Larkin, Guillen
1991: Larkin, Vizquel/Guillen
1992: Smith, Listach/Gagne
1993: Sanchez, John Valentin
1994: Stocker/Clayton, Jose Valentin
1995: Stocker, Gagne
1996: Gagne/Sanchez, Gonzalez
1997: Clayton, Gil
1998: Grudzielanek/Perez, Cruz
1999: Benjamin, Sanchez

LF:
1990: Bonds, Rickey Henderson
1991: Gonzalez, Vaughn
1992: Gonzalez, Raines
1993: Gonzalez, Mack
1994: Gonzalez, Rickey Henderson
1995: Gonzalez, Cardova
1996: Gilkey, Phillips
1997: Hollandsworth, Higginson
1998: Bonds, Curtis
1999: Hidalgo/Jenkins, Erstad

CF:
1990: Dykstra, Webster/White
1991: Butler, White
1992: Sanders, Willie Wilson, White
1993: Carr/Lewis, Lance Johnson
1994: Grissom, White
1995: Grissom, Javier
1996: Santangelo, Nixon
1997: Andruw Jones, Braggs
1998: Andruw Jones, Hunter
1999: Andruw Jones, J. Jones/G. Anderson

RF:
1990: G. Wilson/L. Walker, Brunansky/Orsulak
1991: L. Walker, Sosa
1992: Justice/OÂ’Neill, Brunansky
1993: Merced/Justice, McLemore
1994: Whiten/Sosa, O,Neill
1995: Mondesi/Justice, Mieske
1996: Jordan, Green
1997: Andruw Jones, OÂ’Neill
1998: G. Williams, Lawton
1999: Cedeno, Dye

P:
1990: Maddux, Stieb
1991: Liebrandt, Swift/Abbott
1992: Maddux, Viola/Hibbard
1993: Maddux, Liebrandt
1994: Fassero, Fernandez
1995: Maddux, D. Martinez/K. Brown/Langston
1996: Maddux, Hershiser
1997: Rueter, Rogers
1998: Maddux, Rogers
1999: Maddux/Glavine, Rogers

[Garlon]

Here are the top 5 career at each position since 1901:

C: Rodriguez, Schalk, Bench, Boone, Carter

1B: Hernandez, Votto, Power, Tenney, Sisler

2B: Frisch, Mazeroski, Hubbard, Lajoie, Hudson

3B: Schmidt, Bell, Rolen, Wallach, Nettles

SS: Tinker, Smith, A. Fletcher, Doolan, Maranville

LF: R. White, Veach, R. Henderson, Cruz, L. Gonzalez

CF: Mays, Ashburn, A. Jones, Speaker, W. Davis

RF: Clemente, Barfield, Hooper, Ott, Aaron

P: Maddux, Mays, K. Rogers, T. John, Grimes

[Garlon]

Here are the top career defensive players from 1871-1900:

C: Holberg, Bennet, Snyder

1B: Anson, Werden, Connor

2B: McPhee, Bierbauer, Lowe

3B: Latham, Williamson, Cross

SS: G. Smith, Dahlen, Glasscock

LF: Selbach, Hornung, M. Mansell/York

CF: Welch, Griffin, McAleer

RF: Shaffer, Keeler, Fogarty

P: Bond, Clarkson, Galvin

[Garlon]

The career leaders for certain stats on baseball reference only goes from 1953-2017. Here is a comparison of their rankings to the plus/minus rankings.

C (since 1953)

BBR Total Zone Rating
1. Rodriguez
2. Molina
3. Sundberg
4. Boone
5. Carter
6. Ausmus
7. Bench
8. Crandall
9. Johnson
10. Dempsey

Plus/Minus Defensive Runs Saved
1. Rodriguez
2. Boone
3. Bench
4. Carter
5. Molina
6. Ausmus
7. Dempsey
8. Pena
9. Sundberg
10. Crandall

[Garlon]

The 1906-1910 Cubs won 536 games. Based on the league average of R/G, the team FIP, and the runs that they allowed in those seasons it appears that they saved 396 runs with their defense.

If we credit their errors saved as 0.25 runs each (this is mentioned in one context of The Fielding Bible III) and add this to their Defensive Runs Saved on their +/- plays, we get a total of 398 for those seasons. Without the errors, we get 345 from the +/- plays made. They made 215 fewer errors than expected in those seasons.

[Garlon]

The 1989 NL R/G was 3.94 and the league FIP was 3.60. The Cubs had a FIP of 3.59 and had 4381 IPouts. ((3.94/3.60)*3.59)*(4381/27) = 637, this is their expected Runs Allowed based on their FIP. They actually allowed 623 Runs that season, so we can say they had 14 runs saved on defense.

Their Plus/Minus defensive runs saved that season come to 12.

[Garlon]

I went ahead and calculated the FIP for every team 1871-2017. From this I was able to compare it to the league average runs allowed per 27 outs and determine how many runs above or below this a team was for a season. This is their expected Defensive Runs Saved for the season based on FIP. I then took a 3yr average of a team's expected DRS based on FIP and compared it to their 3yr average based on the PlusMinus system. The result was that the 3yr average was within 17 runs saved of the FIP expected. I did this because these are entirely independent ways of evaluating team defense and I wanted to see how well the PlusMinus system works. With regards to the difference, I think it may be small enough that a significant portion of those 17 runs allowed may just be to luck or a team having fewer/more opponents in scoring position against them, therefore their plays made above/below average may be worth slightly less or more than expected.

There are 2482 team seasons from 1901-2017 and 70% of them have a 3yr average difference within 20 DRS.

For example, if we look at the 1906-1908 Cubs their FIP DRS total for those seasons is 272 and their PlusMinus DRS is 265, a difference of only 7. This puts their average season difference between these at less than 3 DRS for those seasons.

This suggests that we have the data to properly model defense in OOTP.

[Garlon]

C (since 1953)
BBR Total Zone Rating Plus/Minus Defensive Runs Saved
1. Rodriguez 1. Rodriguez
2. Molina 2. Boone
3. Sundberg 3. Carter
4. Boone 4. Bench
5. Carter 5. Molina
6. Ausmus 6. Ausmus
7. Bench 7. Sundberg
8. Crandall 8. Pena
9. Johnson 9. Dempsey
10. Dempsey 10. Johnson
11. Yeager 11. Crandall
12. Dalrymple 12. Lollar
13. Karkovice 13. Wynegar
14. Posey 14. Lavalier
15. Perez 15. Wilson
16. Blanco 16. Blanco
17. O’Brien 17. Pagnozzi
18. Matheny 18. R. Hundley
19. Wynegar 19. Karkovice
20. Pagnozzi 20. O’Brien

[Garlon]

The baseball Reference site has Total Zone Career leaderboards by position from 1953-2017. I decided to compare their values with the values from the Plus/Minus system.

C (since 1953)
BBR Total Zone Rating vs. Plus/Minus Defensive Runs Saved
1. Rodriguez 1. Rodriguez
2. Molina 2. Boone
3. Sundberg 3. Carter
4. Boone 4. Bench
5. Carter 5. Molina
6. Ausmus 6. Ausmus
7. Bench 7. Sundberg
8. Crandall 8. Pena
9. Johnson 9. Dempsey
10. Dempsey 10. Johnson
11. Yeager 11. Crandall
12. Dalrymple 12. Lollar
13. Karkovice 13. Wynegar
14. Posey 14. Lavalier
15. Perez 15. Wilson
16. Blanco 16. Blanco
17. O’Brien 17. Pagnozzi
18. Matheny 18. R. Hundley
19. Wynegar 19. Karkovice
20. Pagnozzi 20. O’Brien

[Garlon]

1B (since 1953)
BBR Total Zone Rating Plus/Minus Defensive Runs Saved
1. Hernandez 1. Votto
2. Helton 2. Hernandez
3. Pujols 3. Power
4. Gonzalez 4. Helton
5. Olerud 5. Pujols
6. Scott 6. Gonzalez
7. Grace 7. Bagwell
8. O’Brien 8. Sexson
9. Power 9. Buckner
10. Murray 10. Evans
11. Martinez 11. Fain
12. Overbay 12. Lee
13. White 13. Rizzo
14. Joyner 14. Bream
15. Yastrzemski 15. Cash
16. King 16. Murray
17. Skowron 17. Karros
18. Rizzo 18. Palmeiro
19. Bream 19. Olerud
20. Hodges 20. Joyner

[Garlon]

2B (since 1953)
BBR Total Zone Rating Plus/Minus Defensive Runs Saved
1. Mazeroski 1. Mazeroski
2. White 2. Hubbard
3. Randolph 3. Hudson
4. Fox 4. Grich
5. Pedroia 5. Sandberg
6. Ellis 6. Green
7. Whitaker 7. Kinsler
8. Kinsler 8. Cano
9. Cano 9. Reese
10. Lemke 10. Knoop
11. Cano 11. Whitaker
12. Grich 12. Gallego
13. Polanco 13. Sizemore
14. Green 14. Stennett
15. Knoop 15. Randolph
16. LeMahieu 16. Gantner
17. Cruz 17. Phillips
18. Kennedy 18. Hill
19. Utley 19. Thompson
20. Hill/Sandberg 20. Ellis

[Garlon]

3B (since 1953)
BBR Total Zone Rating Plus/Minus Defensive Runs Saved
1. Robinson 1. Schmidt
2. Bell 2. Bell
3. Beltre 3. Rolen
4. C. Boyer 4. Wallach
5. Ventura 5. Nettles
6. Rolen 6. C. Boyer
7. Nettles 7. Robinson
8. Gaetti 8. Evans
9. Schmidt 9. Pendleton
10. Longoria 10. Gaetti
11. Pendleton 11. Ventura
12. Boggs 12. Longoria
13. Williams 13. Seager
14. Cirillo 14. Polanco
15. Aur. Rodriguez 15. Williams
16. McMullen 16. Aur. Rodriguez
17. Inge 17. Inge
18. Uribe 18. Machado
19. K. Boyer 19. K. Boyer
20. Chavez 20. Caminiti

[Garlon]

SS (since 1953)
BBR Total Zone Rating Plus/Minus Defensive Runs Saved
1. Smith 1. Smith
2. Belanger 2. Concepcion
3. Ripken 3. Belanger
4. Aparicio 4. Sanchez
5. Vizquel 5. McMillan
6. Tulowitzki 6. Guillen
7. Sanchez 7. Burleson
8. Guillen 8. Ryan
9. Simmons 9. Larkin
10. McMillan 10. Maxvill
11. Hansen 11. Simmons
12. Hardy 12. Speier
13. Gagne 13. Templeton
14. Trammell 14. Campaneris
15. Everett 15. Wilson
16. Brinkman 16. Russell
17. Wilson 17. Reynolds
18. Campaneris 18. Gagne
19. Bordick 19. Thon
20. Kubek 20. Perez

[Garlon]

LF (since 1953)
BBR Total Zone Rating Plus/Minus Defensive Runs Saved
1. Ba. Bonds 1. R. White
2. Yastrzemski 2. R. Henderson
3. Gordon 3. Cruz
4. G. Anderson 4. L. Gonzalez
5. Crawford 5. Ba. Bonds
6. W. Wilson 6. W. Wilson
7. L. Gonzalez 7. Cromartie
8. Gilkey 8. Payton
9. J. Cruz 9. Foster
10. Gant 10. Crawford
11. R. Henderson 11. Ward
12. Damon 12. Tresh
13. Rose 13. Rose
14. R. White 14. Herndon
15. Erstad 15. Cl. Jones
16. Rudi 16. Gordon
17. Surhoff 17. Irvin
18. Cromartie 18. Giles
19. Gardner 19. Parra
20. DeJesus 20. Gardner

[Garlon]

CF (since 1953)
BBR Total Zone Rating Plus/Minus Defensive Runs Saved
1. A. Jones 1. Mays
2. Mays 2. Ashburn
3. Blair 3. A. Jones
4. Piersall 4. Blair
5. Lofton 5. W. Davis
6. D. White 6. D. White
7. Beltran 7. Lemon
8. W. Davis 8. Flood
9. Flood 9. Maddox
10. Maddox 10. Pettis
11. Lemon 11. Dykstra
12. Cameron 12. Piersall
13. Erstad 13. Edmonds
14. Pettis 14. Busby
15. Dawson 15. Cameron
16. L. Johnson 16. W. Wilson
17. Edmonds 17. Berry
18. Geronimo 18. Javier
19. M. Upton 19. Landis
20. Agee 20. Taveras

[Garlon]

RF (since 1953)
BBR Total Zone Rating Plus/Minus Defensive Runs Saved
1. Clemente 1. Clemente
2. Kaline 2. Barfield
3. Barfield 3. Aaron
4. Jordan 4. Heyward
5. Suzuki 5. Callison
6. Heyward 6. Kaline
7. Sosa 7. Dw. Evans
8. Walker 8. Rios
9. Aaron 9. Oliva
10. Sanders 10. O’Neill
11. Dw. Evans 11. Jordan
12. V. Guerroro 12. N. Cruz
13. G. Wilson 13. Justice
14. Kearns 14. Bo. Bonds
15. Betts 15. Pence
16. J. Cruz 16. Sosa
17. Oliva 17. Victorino
18. Rios 18. Kearns
19. Bo. Bonds 19. Sanders
20. Drew 20. Betts

[actionjackson]

This is fantastic work Garlon. Keep it coming. Maybe someday we'll have something that is closer to the RL defensive capabilities of each historical player. Which Sanders is that in RF? Reggie? Can't be Deion. He wasn't in baseball long enough to accumulate that level of defensive counting stats.

...Sad that it only goes back to 1953, but you've got to work with what you've got.

I know TZR goes all the way back, but it must be Plus/Minus that's the fly in the ointment.

[Garlon]

That is Reggie Sanders. The Plus/Minus values I have calculated for every season 1871-2017, but I just posted those so you can see how close it is in agreement to BBR.

I took the 1906 season and custom edited every player's defensive ratings. I based this on the Plus/Minus values and a set of formulas for translating these to defensive ratings in OOTP with 120 on the 250 scale as the league average for every rating: I used 1905-1907 defensive data for players when creating the ratings. I moved some players to their traded teams and then turned off trades.

Here are the results 1906 season Real vs Simulated

Team FLD %

BRO .955 vs .953
BSN .947 vs .955
CHC .969 vs .969
CIN .959 vs .962
NYG .964 vs .961
PHI .956 vs .962
PIT .964 vs .962
STL .957 vs .954

BOS .949 vs .958
CHW .963 vs .969
CLE .967 vs .968
DET .959 vs .962
NYY .957 vs .957
PHA .956 vs .959
SLB .954 vs .956
WSH .955 vs .957

Team DefEff

BRO .694 vs .693
BSN .670 vs .689
CHC .736 vs .725
CIN .692 vs .704
NYG.696 vs .709
PHI .707 vs .710
PIT .701 vs .708
STL .693 vs .690

BOS .677 vs .686
CHW .705 vs .700
CLE .719 vs .715
DET.673 vs .688
NYY .693 vs .695
PHA .689 vs .703
SLB .708 vs .702
WSH .672 vs .684

The Cubs real 1905-1907 Team DefEff was .727, so I think this simulated value of .725 is really an excellent result.

BSN 1905-1907 had a .953 FLD, again this simulated result of .955 seems on target.

We have the Plus/Minus values and the formulas for properly modeling historical defense to this accuracy. Now we need to get it into the game.

[actionjackson]

Originally Posted by Garlon

That is Reggie Sanders. The Plus/Minus values I have calculated for every season 1871-2017, but I just posted those so you can see how close it is in agreement to BBR.

I took the 1906 season and custom edited every player's defensive ratings. I based this on the Plus/Minus values and a set of formulas for translating these to defensive ratings in OOTP with 120 on the 250 scale as the league average for every rating: I used 1905-1907 defensive data for players when creating the ratings. I moved some players to their traded teams and then turned off trades.

Here are the results 1906 season Real vs Simulated

Team FLD %

BRO .955 vs .953
BSN .947 vs .955
CHC .969 vs .969
CIN .959 vs .962
NYG .964 vs .961
PHI .956 vs .962
PIT .964 vs .962
STL .957 vs .954

BOS .949 vs .958
CHW .963 vs .969
CLE .967 vs .968
DET .959 vs .962
NYY .957 vs .957
PHA .956 vs .959
SLB .954 vs .956
WSH .955 vs .957

Team DefEff

BRO .694 vs .693
BSN .670 vs .689
CHC .736 vs .725
CIN .692 vs .704
NYG.696 vs .709
PHI .707 vs .710
PIT .701 vs .708
STL .693 vs .690

BOS .677 vs .686
CHW .705 vs .700
CLE .719 vs .715
DET.673 vs .688
NYY .693 vs .695
PHA .689 vs .703
SLB .708 vs .702
WSH .672 vs .684

The Cubs real 1905-1907 Team DefEff was .727, so I think this simulated value of .725 is really an excellent result.

BSN 1905-1907 had a .953 FLD, again this simulated result of .955 seems on target.

We have the Plus/Minus values and the formulas for properly modeling historical defense to this accuracy. Now we need to get it into the game.

.727 is off the charts for a team DefEff over three years. The Cubs' pitchers must've felt bulletproof out there in those years.

[Reed]

Thank you Garlon. I appreciate your various analysis. In fact, I started a historic league in 1901 using 20 teams and most of your other setting (using 3 yr recalc double weighted instead of 5 years) and it is going great. Just completed 1903. I do have to make some adjustments on opening day to make sure teams have enough pitchers and fielding positions are covered.