Factoring Bullpens Into DFS Decisions
About three or four times each week I create projections and player rankings for DFS contests. As an example, you can find my breakdown of tonight’s DFS action here ($). My projections start with the rest-of-season ZiPS projections, which I break down to a per-game basis. I then adjust each projection for matchup and ball park.
For hitters, the matchup adjustment is based on the wOBA allowed by the starting pitcher they will be facing against hitters of the same handedness since the start of the 2014 season. When small sample sizes apply, I manually set the matchup adjustment. But it recently occurred to me that adjustments for matchup should probably include the strength of the opponent’s bullpen. To date, starters are pitching an average of 5.88 innings per game this year, meaning my matchup adjustment is ignoring roughly a third of each hitter’s plate appearances.
Unfortunately, our team leaderboards will not split up bullpen performance against left and right handed batters, but we should still be able to look at a bullpen’s overall results so far and identify additional matchup advantages. To examine this, I simply took each bullpen’s performance to date and calculated how many fantasy points they are allowing to opposing hitters per batter faced. For the sake of this post, I used the DraftKings scoring system.
Team | Pts/BF | Factor |
---|---|---|
White Sox | 1.43 | 1.14 |
Rangers | 1.38 | 1.10 |
Phillies | 1.37 | 1.09 |
Rockies | 1.36 | 1.09 |
Athletics | 1.36 | 1.09 |
Braves | 1.35 | 1.08 |
Red Sox | 1.33 | 1.06 |
Diamondbacks | 1.32 | 1.06 |
Twins | 1.32 | 1.05 |
Nationals | 1.31 | 1.04 |
Reds | 1.31 | 1.04 |
Mariners | 1.30 | 1.04 |
Indians | 1.29 | 1.02 |
Padres | 1.28 | 1.02 |
Marlins | 1.28 | 1.02 |
Rays | 1.27 | 1.01 |
Tigers | 1.27 | 1.01 |
Cubs | 1.25 | 1.00 |
Brewers | 1.25 | 0.99 |
Angels | 1.22 | 0.98 |
Pirates | 1.22 | 0.97 |
Giants | 1.21 | 0.96 |
Dodgers | 1.17 | 0.93 |
Orioles | 1.16 | 0.93 |
Blue Jays | 1.16 | 0.93 |
Yankees | 1.16 | 0.92 |
Mets | 1.15 | 0.92 |
Cardinals | 1.13 | 0.90 |
Royals | 1.01 | 0.81 |
Astros | 0.99 | 0.79 |
This doesn’t exactly line up with traditional stats like ERA, but it’s fairly close. Five of the six worst bullpens in ERA are among the six worst bullpens in terms of fantasy points allowed per batter faced.
Let’s apply this to the 10-game evening slate tonight. First I’ll show you the matchup factors I created to adjust hitter projections based solely on the quality of the opposing starter. Then we’ll compare that to how those factors differ if we include the strength of an opponent’s bullpen.
Name | Throws | Opp | wOBA v. L | Factor | Name | Throws | Opp | wOBA v. R | Factor | |
---|---|---|---|---|---|---|---|---|---|---|
Jeremy Guthrie | R | Mil | 0.379 | 1.20 | CC Sabathia | L | Mia | 0.355 | 1.16 | |
Colby Lewis | R | LAD | 0.368 | 1.16 | Justin Verlander | R | Cin | 0.35 | 1.15 | |
Ryan Vogelsong | R | Sea | 0.363 | 1.15 | Danny Salazar | R | ChC | 0.335 | 1.10 | |
Jimmy Nelson | R | KC | 0.362 | 1.14 | R.A. Dickey | R | NYM | 0.33 | 1.08 | |
Mike Leake | R | Det | 0.348 | 1.10 | Colby Lewis | R | LAD | 0.329 | 1.08 | |
Clay Buchholz | R | Atl | 0.328 | 1.04 | Clay Buchholz | R | Atl | 0.322 | 1.05 | |
Michael Montgomery | L | SF | 0.317 | 1.00 | Michael Montgomery | L | SF | 0.32 | 1.05 | |
Mat Latos | R | NYY | 0.312 | 0.98 | Bartolo Colon | R | Tor | 0.317 | 1.04 | |
Jeff Samardzija | R | Pit | 0.312 | 0.98 | Mat Latos | R | NYY | 0.306 | 1.00 | |
Bartolo Colon | R | Tor | 0.311 | 0.98 | Ryan Vogelsong | R | Sea | 0.294 | 0.96 | |
Doug Fister | R | TB | 0.310 | 0.98 | Jimmy Nelson | R | KC | 0.294 | 0.96 | |
R.A. Dickey | R | NYM | 0.308 | 0.97 | Mike Leake | R | Det | 0.293 | 0.96 | |
Gerrit Cole | R | CWS | 0.306 | 0.97 | Jeff Samardzija | R | Pit | 0.291 | 0.95 | |
Justin Verlander | R | Cin | 0.305 | 0.96 | Chris Archer | R | Was | 0.29 | 0.95 | |
Danny Salazar | R | ChC | 0.303 | 0.96 | Doug Fister | R | TB | 0.288 | 0.94 | |
Jason Hammel | R | Cle | 0.299 | 0.94 | Gerrit Cole | R | CWS | 0.284 | 0.93 | |
Shelby Miller | R | Bos | 0.298 | 0.94 | Zack Greinke | R | Tex | 0.284 | 0.93 | |
Zack Greinke | R | Tex | 0.276 | 0.87 | Shelby Miller | R | Bos | 0.282 | 0.92 | |
Chris Archer | R | Was | 0.268 | 0.85 | Jason Hammel | R | Cle | 0.276 | 0.90 | |
CC Sabathia | L | Mia | 0.265 | 0.84 | Jeremy Guthrie | R | Mil | 0.273 | 0.89 |
Alright, now let’s compare the chart above to a matchup factor that inlcudes a bullpen component. For simplicity’s sake, I’ve weighted the total matchup component two-thirds for the starter and one-third for the bullpen. I could get much more specific about the weights based on each starter’s average length of start as well as expected game flow based on Vegas lines, but we’ll keep it simple for this post.
Name | Team | Opp | SP Factor | BP Factor | Total | Name | Team | Opp | SP Factor | BP Factor | Total | |
---|---|---|---|---|---|---|---|---|---|---|---|---|
Colby Lewis | Tex | LAD | 1.16 | 1.10 | 1.14 | Justin Verlander | Det | Cin | 1.15 | 1.01 | 1.10 | |
Jimmy Nelson | Mil | KC | 1.14 | 0.99 | 1.09 | Colby Lewis | Tex | LAD | 1.08 | 1.10 | 1.09 | |
Ryan Vogelsong | SF | Sea | 1.15 | 0.96 | 1.09 | CC Sabathia | NYY | Mia | 1.16 | 0.92 | 1.08 | |
Mike Leake | Cin | Det | 1.10 | 1.04 | 1.08 | Danny Salazar | Cle | ChC | 1.10 | 1.02 | 1.07 | |
Jeremy Guthrie | KC | Mil | 1.20 | 0.81 | 1.07 | Clay Buchholz | Bos | Atl | 1.05 | 1.06 | 1.06 | |
Clay Buchholz | Bos | Atl | 1.04 | 1.06 | 1.04 | Michael Montgomery | Sea | SF | 1.05 | 1.04 | 1.04 | |
Jeff Samardzija | CWS | Pit | 0.98 | 1.14 | 1.04 | R.A. Dickey | Tor | NYM | 1.08 | 0.93 | 1.03 | |
Michael Montgomery | Sea | SF | 1.00 | 1.04 | 1.01 | Jeff Samardzija | CWS | Pit | 0.95 | 1.14 | 1.01 | |
Doug Fister | Was | TB | 0.98 | 1.04 | 1.00 | Mat Latos | Mia | NYY | 1.00 | 1.02 | 1.01 | |
Mat Latos | Mia | NYY | 0.98 | 1.02 | 1.00 | Bartolo Colon | NYM | Tor | 1.04 | 0.92 | 1.00 | |
Shelby Miller | Atl | Bos | 0.94 | 1.08 | 0.99 | Mike Leake | Cin | Det | 0.96 | 1.04 | 0.99 | |
Danny Salazar | Cle | ChC | 0.96 | 1.02 | 0.98 | Doug Fister | Was | TB | 0.94 | 1.04 | 0.98 | |
Justin Verlander | Det | Cin | 0.96 | 1.01 | 0.98 | Shelby Miller | Atl | Bos | 0.92 | 1.08 | 0.97 | |
Gerrit Cole | Pit | CWS | 0.97 | 0.97 | 0.97 | Jimmy Nelson | Mil | KC | 0.96 | 0.99 | 0.97 | |
Jason Hammel | ChC | Cle | 0.94 | 1.00 | 0.96 | Chris Archer | TB | Was | 0.95 | 1.01 | 0.97 | |
Bartolo Colon | NYM | Tor | 0.98 | 0.92 | 0.96 | Ryan Vogelsong | SF | Sea | 0.96 | 0.96 | 0.96 | |
R.A. Dickey | Tor | NYM | 0.97 | 0.93 | 0.96 | Gerrit Cole | Pit | CWS | 0.93 | 0.97 | 0.94 | |
Chris Archer | TB | Was | 0.85 | 1.01 | 0.90 | Jason Hammel | ChC | Cle | 0.90 | 1.00 | 0.93 | |
Zack Greinke | LAD | Tex | 0.87 | 0.93 | 0.89 | Zack Greinke | LAD | Tex | 0.93 | 0.93 | 0.93 | |
CC Sabathia | NYY | Mia | 0.84 | 0.92 | 0.86 | Jeremy Guthrie | KC | Mil | 0.89 | 0.81 | 0.86 |
As you can see, adding a bullpen factor does result in some noticeable changes. The Dodgers go from having one of the best matchup adjustment factors to having the best factor for their left-handed hitters and second best for their-right-handed hitters. Even if Anthony Ranaudo ends up starting instead of Colby Lewis for Texas, those Dodgers factors will remain atop each list. On the flip side, targeting some bargain Brewers that hit from the left side seemed like the value play of the day when just using Jeremy Guthrie to adjust their projections. But after considering the strong Royals bullpen, they’re less of a consideration.
I realize there are a lot of factors that I’m not considering here. Depending on the game situation, hitters may run into a LOOGY or ROOGY who will present a tough matchup in that plate appearance. And game flow will have an impact on the quality of relievers that enter a game. But on a general basis, having an idea about which bullpens are on the extremes can be useful when making decisions. If nothing else, it can serve as a tie-breaker.
You can find more of Brett's work on TheFantasyFix.com or follow him on Twitter @TheRealTAL.
Yeah I think that bullpens can serve as a tiebreaker, but as you stated in your last paragraph, there are a lot of different factors that mitigate a basic bullpen analysis. Another one is that relievers themselves are extremely volatile from year to year and bullpen ranks seem to fluctuate wildly.
I think that in 50-50 type games it makes more sense to look at bullpens, in general if you’re going to win a tournament you’re going to stack a team that’ll get into the pen early and end up facing generic mop up man (or even better Jeff Franceour).
Yeah I’m with you on all accounts.