A Quick and Easy Way to Rank Starting Pitchers for Fantasy Baseball

There are different ways to rank players for fantasy baseball. These rankings will depend on the league settings and your personal beliefs as to what is the best method. Currently, there are two main methods that immediately come to mind. Here at FanGraphs, Zach Sanders has posted his method to determine rankings and dollar values using z-scores, which takes into account the standard deviation for that player’s statistics compared to the set of players in the sample. Others prefer a method called Standings Gain Points. With Standings Gain Points, you will need to have an expectation of the end-of-year standings in the different categories based on previous year’s data.

I wanted to find a quick-and-easy way to rank starting pitchers and I remembered reading a post at Tom Tango’s site about the usefulness of “strikeouts minus walks” for pitchers (K-BB). Similarly, some writers here at FanGraphs have begun to use K%-BB% in their posts. I decided to look at a few options to see which is the best “quick-and-easy” way to rank starting pitchers.

For each of the last three seasons, I used the z-scores method to determine the top 100 starting pitchers for comparison purposes. More specifically: I found the sum of the standard deviations for four traditional categories for starting pitchers: wins, strikeouts, ERA, and WHIP (saves were not included because I was just looking at starting pitchers). I’ll use Clayton Kershaw as an example. I divided Kershaw’s wins (21) by the standard deviation of all pitchers’ wins in this sample (3.5) to get the z-score for Kershaw in the wins category (6.0). I did the same for strikeouts (239/41.9=5.7).

For ERA and WHIP, I had to do a little more work. Again, using Kershaw as an example. I took the ERA of the group of pitchers (3.27), subtracted Kershaw’s ERA (1.77), multiplied by Kershaw’s innings pitched (198.3), then divided by nine. The result was 33. This gave me Kershaw’s number of runs saved (runs allowed below what a pitcher with a league average ERA would allow in that many innings). This may be confusing to some. That number, 33, is how many more earned runs Kershaw would have to allow to have a league average ERA. In 2014, Kershaw pitched 198.3 innings and allowed 39 earned runs for an ERA of 1.77. Had he allowed 33 more earned runs, he would have allowed 72 earned runs. Allowing 72 earned runs in 198.3 innings would give him an ERA of 3.27, which is the average of this group of pitchers I’m working with.

I did a similar thing for WHIP. I took the WHIP of the group of pitchers (1.18), subtracted Kershaw’s WHIP (0.86), then multiplied by Kershaw’s innings pitched (198.3). This gave me the number of base runners saved for Kershaw (64). This means had Kershaw allowed 64 more base runners (walks or hits) in the same number of innings pitched, his WHIP would have been league average.

Once I found runs saved and base runners saved for each pitcher, I found the standard deviation of the group of pitcher for each metric. I then divided that pitcher’s runs saved and base runners saved by the standard deviation of the group of pitchers to get the z-scores for ERA and WHIP. Because I was only dealing with starting pitchers, I did not use saves, but if relievers had been included, saves would be done the same way as wins and strikeouts. Once I found the z-scores for wins, strikeouts, ERA, and WHIP, I added them together to get a total number for each pitcher. The pitchers were ranked by this total number for fantasy purposes.

Once I had this total for each pitcher, I ran correlations with each pitcher’s total number based on z-scores and some potential “quick-and-easy” methods to rank these pitchers. I started off with four potential methods: raw strikeouts, K-BB, K%, and K%-BB%. The following shows the correlation for these four methods with the total number figured above (the sum of the z-scores for the four fantasy pitching categories for starting pitchers).

For 2014:

0.80    K-BB

0.72    Strikeouts

0.65    K%-BB%

0.58    K%

 

For 2013:

0.78    K-BB

0.72    Strikeouts

0.67    K%-BB%

0.56    K%

 

For 2012:

0.77    K-BB

0.71    Strikeouts

0.55    K%-BB%

0.44    K%

 

Of these four methods, K-BB has the highest correlation, followed by raw strikeouts. This makes sense because starting pitchers in fantasy baseball get some of their value from the innings they pitch. They need to pitch to get those wins and strikeouts. The other two options (K% and K%-BB%) don’t factor in playing time, so it’s not surprising that they don’t correlate as well as K-BB and raw strikeouts.

With this in mind, I took K% and multiplied by innings pitched for an additional metric, along with (K%-BB%)*IP for another. Below, I’ve included these two options.

For 2014:

0.83    (K%-BB%)*IP

0.80    K-BB

0.78    K%*IP

0.72    Strikeouts

0.65    K%-BB%

0.58    K%

 

For 2013:

0.81    (K%-BB%)*IP

0.78    K-BB

0.77    K%*IP

0.72    Strikeouts

0.67    K%-BB%

0.56    K%

 

For 2012:

0.80    (K%-BB%)*IP

0.77    K-BB

0.76    K%*IP

0.71    Strikeouts

0.55    K%-BB%

0.44    K%

 

As you can see, (K%-BB%)*IP comes out on top, but the more simple K-BB is close and the point is to find a “quick-and-easy” method. The next idea I had was to incorporate home runs allowed. Keeping it simple, I created K-BB-HR and compared it to the others.

For 2014:

0.83    (K%-BB%)*IP

0.82    K-BB-HR

0.80    K-BB

0.78    K%*IP

0.72    Strikeouts

0.65    K%-BB%

0.58    K%

 

For 2013:

0.81    (K%-BB%)*IP

0.81    K-BB-HR

0.78    K-BB

0.77    K%*IP

0.72    Strikeouts

0.67    K%-BB%

0.56    K%

 

For 2012:

0.80    K-BB-HR

0.80    (K%-BB%)*IP

0.77    K-BB

0.76    K%*IP

0.71    Strikeouts

0.55    K%-BB%

0.44    K%

 

This method (K-BB-HR) is right there with (K%-BB%)*IP. It’s not quite as simple as K-BB, but it is quite simple. Using K-BB is very simple and will get you close to the more complex methods to rank starting pitchers. If you want to take it one step farther, use K-BB-HR.

So, without further ado, here are the top 20 pitchers ranked by K-BB-HR using Steamer projections for 2015:

 

1. Clayton Kershaw (164)
2. Chris Sale (152)
3. Max Scherzer (151)
4. Felix Hernandez (141)
5. Stephen Strasburg (137)
6. Yu Darvish (136)
7. Madison Bumgarner (135)
8. Corey Kluber (132)
9. David Price (123)
10. Matt Harvey (121)
11. Zack Greinke (118)
12. Jon Lester (116)
13. Masahiro Tanaka (111)
14. Cole Hamels (110)
15. Adam Wainwright (106)
16. Johnny Cueto (106)
17. James Shields (105)
18. Jeff Samardzija (102)
19. Jordan Zimmermann (101)
20. Ian Kennedy (100)

 

That’s a pretty good-looking list and easy to figure using three basic statistics and subtraction.