Steven Bisig-USA TODAY Sports

Steven Bisig-USA TODAY Sports

When Tom Wilhemsen went lost his job as the closer in Seattle, I picked Danny Farquhar as the replacement ninth inning guy for the Mariners. Wouldn’t you know it: the Mariners agreed. This preamble isn’t a flagrant plea for praise and compliments — though I would take them if offered — it’s for the “why”. I picked Danny Farquhar because of his arsenal.

Yoervis Medina had strikeouts and velocity, and Jack Moore once showed that those things matter. So he was interesting. But Yoervis Medina is a fastball/slider guy, and sliders have platoon splits, and lo and behold Medina’s strikeout rate drops and his walk rate doubles when he’s facing a lefty. If you don’t have pin-point command of that slider — in which case you can make the pitch tickle the inside corner — it’s a pitch that breaks into a lefty’s bat. So that’s why I went Farquhar.

Can we formalize this approach somehow? Does it hold up in the prism of history?

Let’s try it this way. Over at FanGraphs, we have 513 qualified relievers since BIS started supplying the site with pitch classifications. What happens if we take the top 10% by each pitch type and see how many saves they average? In other words, do heavy slider users average fewer saves than heavy changeup users? Let’s run this before we see if there are other ways to get at this problem.

Pitch Average Saves Since 2002
Slider 24.6
Curveball 19.9
Changeup 26.2
Splitfinger 30.2

This might be problematic. Because these guys might still have other pitches. Just because they throw the slider a lot doesn’t mean they don’t have a changeup that could bust platoons. Huston Street is in our slider sample — he throws a ton of sliders — but he has a changeup he throws more than 10% of the time that keeps lefties honest.

So what we really want are people that throw 95+% fastballs and sliders, maybe? So rank by FB+SL, FB+CH, and FB+CB? Let’s try that.

This one went so spectacularly wrong right off the bat that I thought I’d just show you the most fastball-slider-est relievers since 2002:

Aroldis Chapman 40.1% 12.5% 2.51 65 85.1% 14.9% 100.0%
Manuel Corpas 16.6% 6.9% 4.07 34 62.2% 37.7% 99.9%
Chris Perez 22.6% 10.1% 3.26 124 73.1% 26.6% 99.7%
Darren O’Day 22.2% 6.2% 2.68 4 56.7% 43.0% 99.7%
Billy Wagner 31.6% 7.0% 2.06 276 74.2% 25.2% 99.4%
Robbie Ross 19.8% 8.1% 2.48  0 81.8% 17.5% 99.3%
Kenley Jansen 39.9% 9.0% 2.17 51 91.2% 7.9% 99.1%
Joel Hanrahan 26.2% 10.6% 3.55 100 69.6% 29.5% 99.1%
Carlos Marmol 31.9% 15.2% 3.2 117 44.9% 53.7% 98.6%
B.J. Ryan 28.5% 10.8% 3.02 115 76.5% 22.1% 98.6%
Chad Qualls 18.3% 6.6% 3.83 51 64.5% 34.0% 98.5%
Shawn Kelley 25.4% 7.8% 3.54 0 58.6% 39.9% 98.5%
Randy Choate 20.4% 10.3% 3.97 6 71.8% 26.7% 98.5%
Scott Schoeneweis 16.1% 9.8% 4.43 9 79.8% 18.5% 98.3%
Jesse Carlson 19.4% 8.0% 3.63 3 49.4% 48.9% 98.3%
Kevin Gryboski 12.3% 12.6% 4.07 2 92.4% 5.8% 98.2%
Juan Rincon 21.7% 9.9% 3.95 3 66.1% 32.0% 98.1%
Ronald Belisario 20.3% 9.3% 3.13 4 85.7% 12.4% 98.1%
Will Ohman 21.9% 10.1% 4.13 3 57.0% 41.1% 98.1%

There are a lot of closers on this list.

Shawn Kelley’s inclusion on this list gives me pause, however. I asked him recently how he avoids platoon splits and he said that he throws two sliders. The hard, biting one, and then a slurve that “changes a bit, gets a little slower and has more depth.” So that effect could be present — I talked to Luke Gregerson about his three-to-seven sliders, for example.

But the idea behind the use of a second slider is that you have one that looks more like a curve — it’s more 12-to-6 instead of sweeping, and that allows the two-pitch pitcher to have a secret third pitch that doesn’t break into the lefty’s barrel. George Kontos told me that after his TJ surgery, a focus on his mechanics made his slider break more up-down instead of sweeping, and that he found lefties easier to deal with with his new slider.

Looking at this another way: how many saves do pitchers with big, sweeping sliders get? Let’s cull the pitcher list and use only pitchers that throw a fastball and slider more than 95% of the time, and then look at the horizontal breaks on their sliders.

Among the relievers that almost exclusively throw fastballs and sliders, the top 15% by horizontal movement have averaged 30 saves since 2002. The bottom 15% by movement have averaged 49. This isn’t a strong correlation — sweeping sliders by Carlos Marmol, Bob Wickman and Chris Perez have led to more than 100 saves in the same time period. On the other hand, owners of more curveball-type sliders include: Joel Hanrahan, Rex Brothers, Brad Lidge, B.J. Ryan and Billy Wagner. Even Jonathan Broxton’s slider is less sweeping than downward.

Well. We didn’t find a great statistical backing for our intuitive belief. Maybe a small effect. Still, if you start your search for the next closer with velocity, strikeout rate, and a nod at platoon-split-busting arsenals, you’ll be in a good spot.