The Average Starting Pitcher
Since this is the offseason I have some spare time with no Pirates games to watch so I decided that since I believe the Pirates biggest and perhaps only free agent splash this offseason should be a starting pitcher I decided to look into how the typical starter does. I’m hoping to make this a multi-step series and take a look at a bunch of stats from K/9, BB/9, WHIP, FIP, xFIP, WAR and so on but for now this post deals entirely with ERA. I wanted to see the defining marks between a 1 and 2 starter, a 2 and 3 and so on. So I took a look at the past 3 seasons of baseball that is 14,578 starts accumulated by 423 individuals.
I decided to attribute 2,970 starts to #1 pitchers, 2,968 starts to #2 pitchers and 2,880 starts to #3-#5 starters. There is some math behind this as in the course of a 162 game season a team would go through a rotation 32.4 times meaning the #1 and #2 starters would ideally make 33 starts and the #3-#5 would make 32 starts. Thirty teams, over 3 seasons playing 162 games should come out to 14,580 starts but since we were missing two (probably a game that was rained out and never made up) I took it from the pitcher number that ideally starts game 162 which is the #2 starter.
I then ordered the pitchers by the ERA they have put up as a starter over the course of the last three seasons. This means Shelby Miller, Brad Peacock, Aaron Thompson, Dellin BetancesCesar Ramos and Jeurys Familia are at the top of the list with a 0.00 ERA conversely Ryan Verdugo with a 32.40 ERA is at the bottom of the list. Now obviously these pitchers wouldn’t be considered the best and/or worst starters over the last 3 seasons because with the exception of Peacock who started 2 games everyone I mentioned has only made 1 start. However this still works for my purpose. I made the assumption that the best ERAs belong to number 1 starters, the next best #2 starters and so forth. To figure out the average ERA for a #1 starter I went down the list of ERAs gradually getting higher and adding the number of starts made by the pitcher. Once I reached 2,970 I stopped and moved on to a #2 starter and I repeated this process until I got to the end. Now of course this didn’t work out perfectly and I had 4 pitchers who had some starts in two categories so what I did was assign the ERA to each individual start and broke it apart accordingly. For example Roy Oswalt ended up being the borderline between a #1 and a #2 starter, he has made 64 starts over the last 3 seasons and I needed 28 of them to be added to the #1 starter group and the remaining 36 to the number two starter group. Essentially I broke him into two players one who made 28 starts at a 3.50 ERA and one who made 36 starts at a 3.50 ERA. I also split up the innings accordingly meaning I averaged his number of innings per start and then multiplied that by the number of starts in each category. I did this for the other 3 borderline pitchers as well.
So at the end what this gave me was 5 groups of starters separated by ERA. I then used the total ER allowed by each group and the total IP by each group to come up with an ERA for the group. It was a pretty straight forward process but as you can see the explanation of it can make it sound a little complicated. Below are the results of my breakdown.
#1 Starter: 3.06 ERA
#2 Starter: 3.74 ERA
#3 Starter: 4.12 ERA
#4 Starter: 4.51 ERA
#5 Starter: 5.47 ERA
Average Starter: 4.12 ERA
The ranges were as followed
#1 Starter: 0.00-3.50
#2 Starter: 3.50-3.96
#3 Starter: 3.96-4.29
#4 Starter: 4.29-4.80
#5 Starter: 4.80-32.40
To get rid of the extremes I took away the top 10% and bottom 10% of each category and the ranges and averages doing that were:
#1 Starter: 2.79-3.46 (Average of 3.09)
#2 Starter: 3.56-3.91 (Average of 3.74)
#3 Starter: 4.00-4.27 (Average of 4.11)
#4 Starter: 4.31-4.75 (Average 4.51)
#5 Starter: 4.90-6.37 (Average of 5.34)
I took away the top and bottom 10% just to make sure the extremes weren’t playing too much havoc with the averages. Also this now starts the #1 starters with Justin Verlander and ends the #5 starters with Sean O’Sullivan. O’Sullivan is probably not a common name to most but he has made 24 starts over the last 3 seasons so he isn’t just your typical cup of coffee starter who bombs in his 1 or 2 starts. The numbers as you can see for the most part stayed the same which validates my thought that the extremes really didn’t have much impact.
For reference sake I’m going to look at the Pirates six starters and compare their numbers to the above.. I’m going to use their ERA’s over the last 3 seasons, their ERA’s last season and their xFIP over the last 3 seasons and last year’s as well. The ERA’s should show us how they have compared and the xFIP should give us a fair estimate of how we could expect them to perform.
2010-2012: 4.62 ERA, 3.90 xFIP
2012: 3.51 ERA, 3.40 xFIP
2010-2012: 3.63 ERA, 3.80 xFIP
2012: 3.76 ERA, 4.09 xFIP (with Pirates only it was: 3.82 ERA, 4.42 xFIP)
2010-2012: 4.07 ERA, 4.23 xFIP
2012: 4.05 ERA, 4.17 xFIP
2010-2012: 3.96 ERA, 4.04 xFIP
2012: 3.89 ERA, 3.89 xFIP
2011-2012: 6.36 ERA, 4.49 xFIP
2012: 6.30 ERA, 3.61 xFIP
2012: 3.68 ERA, 4.66 xFIP
So now the next logical step is to ask where these Pirates pitchers rank on the number scale well before we can do that we must decide on one number to assign them. Using the xFIP and ERA numbers available I’m going to make some reasonable estimates and predict the following ERAs for the above 6 starters:
Remember these numbers are just approximations based on some educated guesses by me. Burnett, Rodriguez, McDonald and Karstens seem to fit them because that is essentially where they have consistently been. For Locke and McPherson their sample sizes are such that one can’t really get a good feel for them. I put the numbers in between their ERAs and xFIPs and assumed some struggles as young pitchers which I think should put them in 4.40-4.60 range. Basically what I feel comfortable considering either McPherson or Locke for next season is as a 4.50 ERA pitcher. I believe that is a reasonable number for either one to hit next year.
So now the question is what does all this give us?
Burnett: A strong #2 starter, maybe even a borderline #1
Rodriguez: An average to weak #2 starter
McDonald: An average #3 starter
Karstens: A strong #3 starter, maybe even borderline #2
Locke and McPherson: Average #4 starters
So that actually looks pretty good. Two #2s, two #3s and two #4s but alas we know it’s not that simple. Our top two pitchers Burnett and Rodriguez are getting up there in age and could regress. The chances of McDonald being an average #3 starter seem slim as over the course of the last two years he has alternated between looking like a decent #1 to looking like a #5. Karstens is solid and dependable when he is on the field but has durability issues. And Locke and McPherson are young unproven pitchers who could possibly step up and be solid #3s even or could bomb and struggle to hold on to good #5 status. The bottom line is the Pirates have the makings of a decent looking rotation when compared to the rest of major league baseball but there are just so many questions surrounding them that addressing that area this offseason needs to be considered.