Ultimate Bullpen Guide: Arsenal Score, xBABIP & Rankings

Contributing to RotoGraphs’ Bullpen Reports last year brought me much joy: opening infinite Brooks Baseball player cards for sabermetric outcomes and watching glorious GIFs. Oh, Marcus Stroman’s Two-Seamer:


My last Bullpen Report from October looked at possible closers through outcomes and presented BABIP differentials (actual BABIP versus expected BABIP using Inside Edge data).

That’s your first tool in this presentation of, “Your Ultimate Bullpen Guide”: BABIP differential.

Tool number two: Arsenal Scores. Eno worked toward a Pitch Arsenal Score. I followed up with part deux. In short, I z-scored whiff-per-swings and grounders-per-flies (grounders-per-balls-in-play for this post) and then weighed each z-score by its effect on ERA, SIERA and xFIP (averaged the weights on each). The combined score was the pitch score. The combined pitch scores was the arsenal score. Why whiffs and grounders? Because they are obviously two of the most desired results verified by former MLB Pitcher Brian Bannister in a response to a pitch sequencing question on Twitter:   

In part deux, I referenced two of the larger contingencies: weighing pitches by usage and differentiating each pitch type. The following Arsenal Scores do that latter. I think Jeff Zimmerman and Eno (as well as myself) agree that we should not weigh the arsenal scores by usage because the pitchers can and hopefully will make adjustments steering to what is most effective.

Eno provided me with the different pitch type swinging-strike and grounder weights on ERA, SIERA and xFIP. Using the weights’ average to the three (expected) ERA’s, here are you Part Trois Relief Pitcher Arsenal Scores:

You can review each pitch score and arsenal score clustered by player in the first tab, titled “All.”


  • 100-pitch count was the minimum requirement for each pitch type. Otherwise, Zach Britton and his nasty Curve (96-count; 58+% whiff/swing rate)  would have the number one arsenal score (1.85).
  • Splitters were combined with Changeups and the only Knuckler, Steven Wright was added into Cutters.
  • I did not have the Cutter weights so I averaged the Slider and Fourseam weights, which wound up being .33 for whiffs/swings and .10 for GB/BIP.
  • Again the arsenal score is ranked solely by whiff/swings and GB/BIP – not weak contact, called strikes, how often they’re in the zone, etc. In conjunction with this matrix, I would recommend having Zone%, Swing% (swing prevention) and pitch-type outcomes like BABIP as well as homers-per-fly data handy. For example, K-rod still has a great Changeup, but it’s also over a standard deviation worse than the mean when it comes to HR/(FB+LD).

Tool number three: Rankings obviously. Last week, Zach Sanders furnished rankings and auction values with current Steamer projections.

My only beef with Steamer projections are the save totals, but at least they were updated from the last time I ran the rankings, which associated 28 saves to all the closers. In any case, here are your reliever-only rankings from Zach’s post (tab one) and FVARz’ed RotoChamp composite rankings (presented/tab two) for 12-team NFBC formats (108 Pitchers drafted):

…Which completes “Your Ultimate Bullpen Guide”: Arsenal Scores, xBABIP and Rankings.

We hoped you liked reading Ultimate Bullpen Guide: Arsenal Score, xBABIP & Rankings by Daniel Schwartz!

Please support FanGraphs by becoming a member. We publish thousands of articles a year, host multiple podcasts, and have an ever growing database of baseball stats.

FanGraphs does not have a paywall. With your membership, we can continue to offer the content you've come to rely on and add to our unique baseball coverage.

Support FanGraphs

Daniel Schwartz contributes for RotoGraphs when he's not selling industry leading thermal packaging. You can follow him on twitter @RotoBanter

newest oldest most voted

What is the definition of “Arsenal score”. I didn’t see it clearly defined (but might have missed it).