clock menu more-arrow no yes

Filed under:

Ranking Relievers by Their Skills Alone: Top 30

New, 2 comments

In this post, I unveil a purely mathematical listing of the best relievers in baseball in 2015, with no regard to their position in the bullpen, number of innings, or number of saves and holds. This ranking focuses solely on several key stats known to be good predictors of future success. The top 30 are covered today.

The ground ball king with a nasty sinker tops my statistical reliever rankings. Who else is in the top 30?
The ground ball king with a nasty sinker tops my statistical reliever rankings. Who else is in the top 30?
Tommy Gilligan-USA TODAY Sports

This is a post about relievers. Unlike most such posts, there will be little mention of saves or holds. This is about ranking relievers only on their "stuff". I used some basic statistical analysis to rank all the relievers in baseball in 2015, regardless of innings pitched (I set the cutoff at 10 innings minimum to eliminate extreme cases). The six stats I chose to include in my rankings are ones that I feel (based on other research) are the best predictors of successful pitchers.

The stats are: xFIP, K%-BB%, GB%, Hard%, and FBv. xFIP helps to predict a pitcher's ERA if HR/FB ratios and BABIPs were at league averages. K%-BB% helps balance a pitcher's strikeout ability and control. It allows for a high strikeout pitcher to walk more batters than a low strikeout pitcher. GB% is important because grounders lead to outs almost as often as fly balls, with no chance of a home run. Hard hit % is a measure of the quality of contact a pitcher allows. The lower this number, the better a pitcher is at preventing hitters from squaring up their pitches. Finally, FBv is simply a pitcher's average fastball velocity. Research has shown that high velocity pitchers have lower ERAs than ones with lower velocities. There is simply a larger margin of error for pitchers that throw hard.

The rankings are based on z-scores. Basically, z-scores are a statistical tool for comparing something to a larger population. A z-score is calculated by taking the value from a specific player, say his fastball velocity, and then subtracting the average fastball velocity of all relievers and then dividing all that by the standard deviation of fastball velocity for all relievers. It basically measures how far a given player in a given stat is from the league average. A higher z-score means he is farther from the average and that is a good thing. I have set up the z-score formulas so that a lower xFIP and lower Hard% result in higher z-scores, while the other three stats are simply higher=better.

Here are the league averages and standard deviations:

xFIP GB% SwStr% K%-BB% Hard % FBv
Avg 4.03 44.7% 10.8% 12.9% 28.3% 92.47
Std Dev 0.85 9.9% 2.8% 7.2% 5.5% 2.73

To make this all more manageable, I'm just going to post the top 30 today and then the next 30 later this week. I'm going to pick out a few interesting names to discuss, but I'll mainly let the table speak for itself. The table is sorted by total z-score, which is simply the sum of all six other z-scores.

Here it is:

Name zxFIP zGB% zSwStr% zK%-BB% zHard% zFBv Total z score
Zach Britton 2.69 3.46 2.01 1.78 1.59 1.26 12.79
Carter Capps 3.38 -0.39 5.21 4.22 -2.34 2.06 12.14
Adam Ottavino 1.98 1.86 0.69 2.57 2.26 1.18 10.54
Aroldis Chapman 1.82 -0.77 3.04 2.35 1.46 2.57 10.47
Tony Zych 1.27 0.92 1.15 1.85 2.40 1.37 8.96
Jeurys Familia 1.75 1.37 1.83 1.22 0.79 1.69 8.66
Andrew Miller 2.33 0.36 2.58 2.73 -0.27 0.67 8.39
Dellin Betances 1.91 0.30 1.54 2.01 0.65 1.66 8.08
Sam Dyson 1.42 2.43 0.58 0.46 1.97 1.22 8.07
Craig Kimbrel 1.85 0.14 1.76 1.99 0.49 1.77 7.99
Carson Smith 1.97 2.02 0.62 1.63 1.57 0.16 7.97
Jordan Walden 1.07 -0.24 3.47 0.86 0.94 0.63 6.73
Kenley Jansen 2.05 -0.96 2.08 3.20 0.02 0.01 6.40
Luke Gregerson 1.56 1.58 1.58 1.06 1.46 -1.13 6.11
Marc Rzepczynski 1.51 2.28 1.26 0.59 0.60 -0.25 5.98
Nate Jones 1.92 0.12 1.61 2.26 -1.85 1.88 5.95
Aaron Sanchez 0.82 2.31 -1.41 -0.11 2.86 1.40 5.87
Sergio Romo 1.90 0.01 2.11 1.89 1.70 -1.82 5.79
Alex Colome 1.07 -0.58 1.54 1.36 1.46 0.82 5.67
Hector Rondon 1.18 0.78 0.01 0.89 1.10 1.44 5.40
Wade Davis 1.13 -0.64 0.30 1.42 1.92 1.26 5.38
Brett Cecil 1.94 0.69 1.51 1.90 -0.63 -0.06 5.34
Mark Lowe 1.42 -0.44 1.19 1.38 0.60 1.11 5.24
Josh Osich 0.16 0.41 2.36 0.41 0.65 1.18 5.17
Blake Treinen 0.85 1.81 0.05 -0.15 1.21 1.40 5.17
Keone Kela 1.36 0.59 1.04 1.07 -0.18 1.15 5.03
Pedro Strop 1.03 0.66 2.15 0.89 -0.69 0.96 5.00
Ken Giles 1.16 0.01 1.44 1.10 -0.25 1.48 4.92
A.J. Ramos 0.93 -0.13 1.86 1.26 0.85 0.05 4.82
Mychal Givens 1.95 -0.53 0.51 2.01 0.09 0.67 4.69

The top of the list has two Tommy John victims in Capps and Ottavino. Ottavino might come back this year, but Capps will have to wait until 2017. Neither is likely to be relevant for fantasy this year. Sigh.

Zach Britton fans will be glad to see him at the top of this list. Obviously, his ground ball rate gives him a big boost (that 3.46 z Score for GB% is huge), but he's good enough in all the other categories as well to earn the top spot. Chapman, Familia, Miller, and Betances are all up there too, to no one's surprise.

This list is a great source of potential bullpen sleepers. Look at all the relievers with great skills behind shaky closers. I've talked about Tony Zych before, in my Mariners bullpen preview, so it's nice to see him high up. Both Sam Dyson and Keone Kela are in the top 30, so competition will be fierce if Shawn Tolleson struggles. That is a highly underrated bullpen. Carson Smith would be a great closer option if Kimbrel got hurt. I think he can pass Koji on the depth chart.

Alex Colome is my favorite Boxberger replacement option, even if Kevin Cash may not give him the job outright. He has the best skills in that 'pen. We all know Sergio Romo is breathing down Santiago Casilla's neck for that Giants closer job, so it's a distinct possibility he wins the job at some point. If Francisco Rodriguez falters or gets injured in Detroit, Mark Lowe would be a great pickup.

Nate Jones is a great holds option behind David Robertson in Chi-town and if something happens, he could end up closing there. Blake Treinen has a wicked sinker like Britton and could be the next man up if Papelbon gets hurt in Washington. Jordan Walden is underrated as a great holds option in St. Louis, if he can stay healthy. Josh Osich is at least the fourth option in San Fran and is only on this list because of his elite 17% SwStr% last season. I don't think he can maintain that over a full season, so I'm not all that interested in him. Marc Rzepczynski (hold on, I need to take a break after trying to spell that) is just a LOOGY and not worth anything in fantasy.

Mychal Givens is a very good reliever stuck behind three other good-to-great options in Baltimore (Brach, O'Day, Britton), so he probably doesn't have much fantasy value. Luke Gregerson actually has a better score than the guy replacing him in Houston (Giles). The back of that bullpen is also loaded.

Finally, it's nice to see the best closers in the league well-represented in this list, showing that it seems to do a good job of evaluating talent. Wade Davis, Kenley Jansen, Hector Rondon, Giles, A.J. Ramos, Craig Kimbrel, and the guys I mentioned earlier are all elite closers with great z-scores.

Tune in later this week for part 2 where we will look at the 30-60 ranks in this list and see if we can uncover any deeper bullpen sleepers or find out which established closers actually trail their fellow bullpen mates in raw skill and could be vulnerable. Tschus!