How Do Different Types of Pitchers Age?

In prior posts, I have calculated a statistic that gives us a good idea of a way to determine how to judge a hitter's ability.  I followed it up with an analysis of how hitters age in two different ways: elite vs. average and on-base vs. slugging.  In another prior post, I have discussed the creation of a statistic to determine how to judge a pitcher's ability.  Specifically, I look at three aspects of a pitcher that are essentially unaffected by anybody else.  An elite defense can't stop a walk from being a walk or a fly ball from being a fly ball.  Likewise, a terrible defense cannot both up a strikeout (with the immaterial occurrence of a dropped strike three, one of the weirdest baseball rules around).  Now, I will look at how pitchers age.  First, I will look at how elite pitchers age vs. average and some terrible pitchers.  Then, I will look at the three individual aspects that are used in the Expected Runs Against statistic: strikeout percentage, walk percentage, and groundball / fly ball ratio.

First, I took a group of 72 pitchers who pitched in at least 14 of the 17 years between age 24 and age 40 who pitched a reasonable amount of innings each year (I used a soft floor of 25 IP, but 100 PA against was a hard floor).  It's surprising that of every pitcher who pitched from 1995 - 2012, only 72 were eligible for this list (although I did eliminate pitchers who were strongly associated with PEDs: Roger Clemens, Andy Pettitte, Bartolo Colon.  I may have missed a few, but I did take out those three for sure.  I then took their career Expected Runs Against statistic to put them into three groups.  Group 1 consists of players with the best Expected Runs Against and Group 3 consists of players with the worst Expected Runs Against.  Here is a listing of each group:

Group 1 (Eq. RA of 3.16 - 4.23): Mariano Rivera, Greg Maddux, Kevin Brown, Billy Wagner, Derek Lowe, Pedro Martinez, Curt Schilling, Jon Lieber, Muke Mussina, Randy Johnson, Trevor Hoffman, John Smoltz, Orel Hershiser, Mike Timlin, David Wells, Rheal Cormier, John Franco, Rick Aguilera, John Burkett, Greg Swindell, Dennis Eckersley, Jimmy Key, Paul Assenmacher, Bob Wickman

Group 2 (Eq. RA of 4.27 - 4.67): Dennis Martinez, Mike Morgan, Terry Mulholland, Kevin Millwood, Rick Honeycutt, Dan Plesac, Scott Sanderson, Lee Smith, Buddy Groom, LaTroy Hawkins, Jamie Moyer, Tom Candiotti, Todd Jones, Alan Embree, Roberto Hernandez, David Cone, Norm Charlton, Tom Glavine, Danny Darwin, Mike Stanton, Tom Gordon, Michael Jackson, Guillermo Mota, Arthur Rhodes

Group 3 (Eq. RA of 4.68 - 5.59): Kenny Rogers, Chuck Finley, Livan Hernandez, Buddy Black, Woody Williams, Greg Harris, Aaron Sele, Darren Oliver, Mark Langston, Eddie Guardado, Steve Trachsel, David Weathers, Tim Belcher, Jesse Orosco, Tim Wakefield, Dave Burba, Jeff Nelson, Dennis Cook, Jamey Wright, Jose Mesa, Miguel Batista, Dave Stewart, Kent Mercker, Ron Villone

Basically, this is the control group.  Here is a graph showing how these three groups do as they age over time:

Just a few things to notice from this group.  First, there is a clear distinction between the three groups over time.  The lines never cross, showing that elite pitchers don't explode at the end of their career or anything like that.  The other thing to notice is that elite pitchers seem to peak between ages 29 and 32. Group 2 pitchers age a little later; ages 31-34 seem to be their best four years.  Group 3, the average and mediocre pitchers, peak much later; ages 34-36 seem to be their peak years.  Signing a 33 year old elite pitcher to a long term deal would, therefore, seem to be more risky than signing a 29 year old elite pitcher to a long term deal.  However, for a group 3 pitcher, a signing at age 33 would be a better idea than signing the same pitcher at age 29. 

The second statistic I will look at is career strikeout percentage.  I separated the 72 pitchers into three groups based on career strikeout percentage.  Group 1 has the best strikeout percentage, while group 3 has the lowest strikeout percentage.

Group 1 (19.10% - 33.20% K%): Billy Wagner, Randy Johnson, Pedro Martinez, Trevor Hoffman, Jeff Nelson, Curt Schilling, Lee Smith, Mariano Rivera, Arthur Rhodes, Dan Plesac, Paul Assenmacher, David Cone, John Smoltz, Jesse Orosco, Tom Gordon, Norm Charlton, Alan Embree, Roberto Hernandez, Michael Jackson, Eddie Guardado, Mark Langston, Mike Mussina, Guillermo Mota, Chuck Finley

Group 2 (15.30% - 19.10% K%): Rick Aguilera, Mike Stanton, Todd Jones, John Franco, Greg Harris, Dave Burba, Kevin Millwood, Ron Villone, Dennis Eckersley, Kevin Brown, Mike Timlin, Dennis Cook, Bob Wickman, Jon Lieber, Greg Swindell, Greg Maddux, David Weathers, Kent Mercker, Woody Williams, John Burkett, Tim Wakefield, Dave Stewart, David Wells, Danny Darwin

Group 3 (11.40% - 15.30% K%): Orel Hershiser, Jose Mesa, LaTroy Hawkins, Buddy Groom, Derek Lowe, Scott Sanderson, Tom Candiotti, Darren Oliver, Aaron Sele, Steve Trachsel, Rheal Cormier, Tim Belcher, Miguel Batista, Jimmy Key, Livan Hernandez, Jamie Moyer, Tom Glavine, Kenny Rogers, Dennis Martinez, Jamey Wright, Terry Mulholland, Buddy Black, Mike Morgan, Rick Honeycutt

And here is a graph showing how these three groups age over time:

High strikeout percentage seems to be a positive throughout a career, but it's certainly not the strongest of relations.  While from age 26 to age 37, group 1 had the lowest expected runs against, it was somewhat inconsistent.  In addition, groups 2 and 3 switched places very often.  From age 27-30, group 3 pitchers, the group with the lowest strikeout percentage, fared better than group 2, the middle of the road strikeout percentage pitchers.  So, while we can take a few things from this relationship, it's definitely not too strong.

Group 1 pitchers have a peak period from ages 31-32, with expected runs against being below 3.90.  Group 2 pitchers are consistently fairly good from ages 30-38, with no particular peak period in that timeframe.  Group 3 pitchers are all over the board.  At age 33, they have an expected runs against of about 4.50.  At age 34, it goes down to about 4.30.  The difference between those two years for group 3 is higher than the variance between any two years from ages 30-38 for group 2 pitchers.

Signing a group 1 pitcher at age 33 would likely involve an overpayment, based on this information.  Signing a group 2 pitcher at age 30 - 38 will likely at least give you some consistency, whereas signing a group 3 pitcher on the free agent market at any time is more risky than signing a group 2 pitcher.  There are some years that group 3 is actually better than group 2, but there are some years that group 2 is significantly better.  However, strikeout percentage doesn't seem to be a strong indicator of how a pitcher will do in coming years.

Next up, I look at walk percentage.  Again, I separated the pitchers into 3 groups with group 1 having the lowest walk percentage and group 2 having the highest walk percentage.

Group 1 (4.50% - 7.20% BB%): Jon Lieber, Greg Maddux, David Wells, Curt Schilling, Mike Mussina, Greg Swindell, Dennis Eckersley, Mariano Rivera, Scott Sanderson, Terry Mulholland, Jimmy Key, John Burkett, Rheal Cormier, Rick Aguilera, Jamie Moyer, Kevin Brown, Pedro Martinez, Danny Darwin, Dennis Martinez, Derek Lowe, Trevor Hoffman, John Smoltz, Rick Honeycutt, Buddy Black

Group 2 (7.30% - 9.00 BB%): Kevin Millwood, Mike Timlin, Woody Williams, LaTroy Hawkins, Tom Candiotti, Orel Hershiser, Livan Hernandez, Mike Morgan, Tom Glavine, Buddy Groom, Kenny Rogers, Billy Wagner, Tim Belcher, Aaron Sele, Tim Wakefield, Paul Assenmacher, Darren Oliver, Eddie Guardado, Steve Trachsel, Randy Johnson, Mike Stanton, Alan Embree, Dan Plesac, Lee Smith

Group 3 (9.10% - 12.60% BB%): Dennis Cook, Dave Stewart, Guillermo Mota, David Cone, John Franco, Bob Wickman, Michael Jackson, Todd Jones, Jose Mesa, Chuck Finley, Dave Burba, David Weathers, Roberto Hernandez, Arthur Rhodes, Mark Langston, Miguel Batista, Greg Harris, Kent Mercker, Jesse Orosco, Norm Charlton, Tom Gordon, Jamey Wright, Ron Villone, Jeff Nelson

Here is a graph showing how these three groups fared over time:

Walk percentage appears to be a much better way to judge a pitcher's ability over time than strikeout percentage.  Group 1 always has the lowest Expected RA and group 3 almost always has the highest Expected RA (and they always do after age 28, when most pitchers become free agents).  Not only that, but Group 1 shows a significant improvement over group 2 and group 2 shows a very noticeable improvement over group 3. Signing a pitcher with a low walk percentage at pretty much any point in a pitcher's career will generally be a better idea than signing a pitcher with a high walk percentage.

Pitchers in group 1 appear to peak from ages 31-34 (although the Expected RA at age 33 takes an unexpected leap, ages 31, 32, and 34 are lower than the rest of the year; the age 33 jump can likely be attributed to statistical noise.  Group 2 pitchers seem to peak from ages 34-36 and group 3 pitchers seem to peak from ages 29-32 (I'm attributing the drop in expected RA at age 35 to statistical noise).  However, At any point over a pitcher's career, it seems preferable to sign a group 1 pitcher.  Perhaps a good deal can be found on a group 2 pitcher in their mid-30s, but even then, a group 1 pitcher is preferable.

The final statistic to look at is Ground Ball / Fly Ball ratio.  Again, I split the 72 pitchers into 3 groups:

Group 1 (0.86 - 1.62 GB / FB Ratio): Derek Lowe, Kevin Brown, Orel Hershiser, Mike Morgan, Greg Maddux, John Franco, Bob Wickman, Jamey Wright, Rick Honeycutt, Mike Timlin, Dennis Martinez, Mariano Rivera, Rheal Cormier, Todd Jones, Greg Harris, Buddy Groom, Miguel Batista, Terry Mulholland, David Weathers, Norm Charlton, Tom Gordon, Tom Glavine, Jon Lieber, Kenny Rogers

Group 2 (0.70 - 0.86 GB / FB Ratio): Roberto Hernandez, Tom Candiotti, John Burkett, Chuck Finley, LaTroy Hawkins, Jimmy Key, Darren Oliver, John Smoltz, Jeff Nelson, Mark Langston, Steve Trachsel, Paul Assenmacher, Aaron Sele, Jose Mesa, Mike Mussina, Buddy Black, Randy Johnson, David Wells, Jamie Moyer, Billy Wagner, Livan Hernandez, Curt Schilling, Kevin Millwood, Dave Burba

Group 3 (0.43 - 0.70 GB / FB Ratio): Pedro Martinez, Guillermo Mota, Michael Jackson, Arthur Rhodes, Mike Stanton, Rick Aguilera, Tim Belcher, Dan Plesac, Ron Villone, Kent Mercker, Tim Wakefield, Jesse Orosco, Alan Embree, David Cone, Lee Smith, Greg Swindell, Dave Stewart, Woody Williams, Danny Darwin, Scott Sanderson, Trevor Hoffman, Dennis Cook, Dennis Eckersley, Eddie Guardado

And here is the graph showing the relationship between age and expected runs against using GB / FB ratio

 Wow, this looks like a mess.  Group 1 is consistently better than the other two groups from ages 24 - 38.  Groups 2 and 3 are close behind group 1, however, and like the strikeout percentage graph, group 3 and group 2 don't seem to tell us much.  Group 2 sometimes has a significantly lower expected RA than group 3 and sometimes has a higher expected RA than group 2.  In short, we may get some information from this graph, but the relationships don't look that strong.

Good ground ball - fly ball ratio pitchers don't seem to have a short peak; they seem to be consistently at their best from ages 29-35 with an expected run against stat of about 4.00.  Group 2 pitchers seem to peak from ages 30-32 and then again from ages 34-36.  Age 33 seems to be a bit of an outlier and is probably statistical noise.  Group 3 pitchers seem to be at their best from ages 31 - 36 with a lull in the middle (ages 33-34).  There doesn't seem to be much consistency for groups 2 and 3 when it comes to GB/FB ratio.  GB/FB ratio is very consistent from ages 29-35 for group 1, but it doesn't really seem to be a significant difference from the rest of the group.  Perhaps having a very good GB / FB ratio at age 29-32 may be something that could help a team determine if they want to sign a free agent long term, but it should not be the deciding factor.

So, in conclusion:

• When looking at the three major categories of stats that only the pitcher controls, walk percentage is by far the best in determining how good a pitcher is.  Not only are the best pitchers in that category consistently good, but they are also significantly better than the other groups.  At any time over a pitcher's career, walk percentage should be a top consideration when signing a free agent pitcher long term.
• Strikeout percentage and groundball / flyball ratio may tell us some, and the best of the best in these groups are better than the middle of the road and the mediocre pitchers in these statistics.  Still, using these ratios to determine whether signing a pitcher long term is very risky.
• Better pitchers also seem to peak earlier (in their late 20s and early 30s) as opposed to average pitchers, who peak in their mid and ever upper 30s.