clock menu more-arrow no yes mobile

Filed under:

Looking at Chris Sale’s conflicting projections

Will the newest Red Sox starter be great in 2017? Or just very good?

Chicago White Sox v Philadelphia Phillies
Will Sale have some hiccups in Fenway Park?
Photo by Mitchell Leff/Getty Images

The offseason sucks, man. You can stare out the window and wait for spring for only so long. After next weekend, football’s done for the year, and basketball courts and hockey arenas just can’t stack up to the thrill of the diamond. Aside from the Australian Baseball League — which our intrepid foreign correspondent Anthony Rescan has chronicled for your consideration — nothing resembling America’s pastime will exist until pitchers and catchers report.

In this trying time, projections bring us something resembling solace — they tell us how our favorite players and teams should fare in the coming season, thereby giving us something to look forward to. FanGraphs released Steamer projections back in November following the conclusion of the World Series; throughout the last few months, they’ve also published ZiPS projections for 27 of the 30 teams (only the Mets, Reds, and Rockies remain). For the statistically oriented fan, these numbers cultivate excitement and anticipation for the season to come.

But what happens when these projections disagree? Seeing Steamer and ZiPS reach different conclusions about a player’s potential is like seeing parents fight — well, maybe not that bad, yet still rather jarring. That’s why, here at Beyond the Box Score, we aim to resolve these disputes in a recurring series. Over the next several weeks, we’ll examine some players with vast differences in their projections, coming to a firm conclusion about their 2017 potential. While it can’t compare to actual baseball, it’ll help to get us through these last few months before MLB makes its comeback.

First up is Chris Sale, the 27-year-old lefty who recently swapped his White Sox for Red ones. No matter where you look, Sale is projected to be an ace in 2016; it’s the scale of that acehood that differs:

Sale projections, 2017

Projection IP K% BB% HR% BABIP LOB% ERA FIP fWAR fWAR/200
Projection IP K% BB% HR% BABIP LOB% ERA FIP fWAR fWAR/200
ZiPS 199.2 27.9% 4.9% 2.3% .297 79.9%* 2.79 2.88 6.2 6.2
Steamer 213.0 25.3% 5.3% 2.7% .283 72.8% 3.42 3.37 4.7 4.4
*ZiPS doesn’t project strand rate, so I calculated this manually, using Sale’s lifetime hit-by-pitch rate of 1.4 percent.

To figure out whether ZiPS’s optimism or Steamer’s pessimism is most justified, let’s take a deep dive into the data. We’ll examine the five areas of disagreement — strikeouts, walks, home runs, batting average on balls in play, and strand rate — and come to a conclusion about which projected outcome is correct.


Like most front-line starters, Chris Sale makes his living on strikeouts. Since he joined the White Sox rotation in 2012, he's fanned 27.8 percent of the hitters he’s faced, the seventh-highest clip among qualified starters. ZiPS predicts he’ll remain in line with that level of play, while Steamer expects a bit of a dropoff. Which system should we trust?

That 27.8 percent figure doesn’t tell the whole story. Sale’s evolved quite a bit over his five years as a starter, particularly with regards to the K. In 2012 and 2013, he notched a 25.6 percent strikeout rate; he then lept ahead to a 31.3 percent strikeout rate in 2014 and 2015; and he tumbled back to a 25.7 percent strikeout rate in 2016. To put that in simpler terms, Steamer thinks the 2016 Sale is the new normal, whereas ZiPS splits the difference between the new and old.

Sale’s punchout decline last year occurred, as it so often does, because he got fewer swings-and-misses: After running a 13.8 percent whiff rate in 2014 and 2015, he dropped down to 11.3 percent in 2016, more in line with his 10.8 percent whiff rate from 2012–13. That dip, in turn, likely has something to do with his pitch selection:

Image via Brooks Baseball

Hitters have never been able to figure out the changeup; since 2012, it has a whiff rate of 18.8 percent, the highest in his arsenal. Yet for whatever reason, Sale decided to stop leaning on it last year, instead opting for more sliders (2012–16 whiff rate: 15.9 percent) and fastballs (2012–16 whiff rate: 11.1 percent).

Why would Sale forgo his most deceptive pitch? At the same time as this change, his fastball velocity plummeted two ticks — from 95.6 mph in 2015 to 93.6 mph in 2016 — while the changeup remained steady at about 86. Josh Kalk’s research suggests a larger fastball-changeup velocity gap is more advantageous, so with less clout on the heater, the cambio might have suffered as a result. (Indeed: Sale’s changeup didn’t get as many swings-and-misses in 2016, in addition to decreasing in usage.)

That velocity decline wasn’t because of age or injury. As Sale told’s Scott Merkin in May, he intentionally threw softer to focus more on locating his pitches:

"That's probably the biggest part of my change, is not throwing every single pitch as hard as I can every inning, every out. I waste a lot of pitches doing that," Sale told prior to his Saturday start against the Twins. "I've noticed being able to throw strikes down in the zone where last year my fastball was just getting crushed.

"You can throw 96, but if it's up, they are going to hit it. I'm starting to realize it's more location than it is speed and velocity.

"There are still times to overpower guys," Sale said. "But when you are going through a game, there's no reason to throw a 0-0 pitch as hard as you can just because you gave up a hit before that. Now you are 1-0 and now you are more mad than you were before, so you rear back and try to throw one harder."

This alteration hasn’t gone unnoticed. Sale’s shift in selection stupefied our sister site, South Side Sox. In July, they wrote on his move away from the changeup, remarking he was “making an extreme choice in his pitching style to favor contact over whiffs.” On the surface, his decision seems to have paid off: Sale’s BABIP in 2016 — which we’ll discuss in a moment — fell to .279, his lowest mark since becoming a starter.

That BABIP didn’t cancel out the missing Ks, though. A strikeout rate in the mid-20s is still elite — Sale did rank 13th in the majors in that regard last year — but it can’t compare to the echelon Sale used to occupy. If he keeps his velocity low, and continues to stay away from his changeup, Sale will most likely sustain last year’s strikeout rate, to align with Steamer’s prediction.


Unlike with strikeouts, the projections align pretty closely here. Both Steamer and ZiPS know Sale has impeccable control, given his 5.0 percent walk rate last season, and they appraise him accordingly. But the difference is still a significant one, and one that illustrates two different ways — one of them superior — to evaluate his profile.

In 2016, Sale threw 66.9 percent of his pitches for strikes. Of those strikes, 26.7 percent were balls in play. And as mentioned above, he fanned 25.7 percent of the batters to step in against him. Those stats are relevant because, back in 2013, RotoGraphs’ Mike Podhorzer ran a regression to find an expected walk rate equation:

xBB% = 0.7598 + (-0.7300 * Str%) + (-0.5729 * I/Str) + (-0.2341 * K%)

By this formula, Sale should have issued a free pass to 5.8 percent of opposing hitters last year. That would seem to endorse the (relatively) negative Steamer view — Sale got lucky in 2016, which means he’ll regress back to normal in 2017. Should we make that our conclusion?

Not in my opinion. Two factors complicate Sale’s case. First, there’s the matter of another critical pitcher skill, perhaps more important than strike-throwing, efficiency, or strikeouts: pitch sequencing. In 2015, following up on Podhorzer’s research, RotoGraphs’ Alex Chamberlain unveiled a new, improved expected walk rate equation. This one incorporated the same three metrics as Podhorzer’s, while adding in 3-0 count rate as a proxy for sequencing ability:

xBB% = 0.598 + (-0.494 * Str%) + (-0.595 * I/Str) + (-0.264 * K%) + (0.515 * 3-0%)

In this light, Sale’s stock shoots up. Of the 909 hitters to step in against him in 2016, just 16 worked him to a 3-0 count, which translates to a 1.8 percent clip. Plugging that into this formula, we receive a 5.0 percent expected walk rate — actually lower than his actual walk rate. With more information taken into account, our perspective on Sale’s control becomes a bit more bullish.

Then there’s the second factor to keep in mind: framing. Given that Sale threw 52.3 percent of his pitches in the strike zone, and coaxed the hitter into swinging 32.9 percent of the time when he didn’t, we’d expect him to have a strike rate of 68.0 percent — more than a full percentage point above his actual strike rate. As my colleague Evan Davis explained back in June, the White Sox abandoned framing in 2016; their catchers combined to cost them an astounding 26.5 runs via framing, per Baseball Prospectus.

That might not be the case in Boston. The Red Sox have some uncertainty behind the plate, but all their prominent catching options have some upside. Christian Vazquez is one of the best receivers in the world, with 20.7 framing runs in less than 900 career innings. And while neither Swihart nor Leon has framed well in the majors — where they’ve accumulated -6.7 and -8.3 framing runs, respectively — they’ve each performed much better in the minors, suggesting they could improve in 2017.

The figures involved are pretty similar — over 900 batters faced, 4.9 percent versus 5.3 percent is a difference of four walks — so Sale could feasibly go in either direction. Still, given his skill at avoiding 3-0 counts, and his probable 2017 battery mates, I’d predict some improved control from Sale this year. In this regard, we’ll go with what ZiPS projects.

Home Runs

Now, for the most consequential of the three true outcomes: the long ball. Sale’s never been a homer-prone pitcher, but in 2016, he — like pretty much everyone — gave them up more than ever before. His 3.0 percent home run rate last year was markedly higher than the 2.5 percent clip he posted in the four previous seasons. Although either Steamer nor ZiPS expects that pace to continue, they don’t see eye-to-eye on the scale of his regression.

Like many high-strikeout pitchers, Sale also gets a good amount of fly balls. Last season, his 41.2 percent ground ball rate ranked him 51st out of 73 qualifiers. Plus, as a southpaw, Sale faces a lot of righties (81.0 percent of opposing hitters since 2012); combined with the short porch in Fenway Park, that could lead to some more balls leaving the yard.

But Sale has two things working in his favor. Those Ks mean he doesn’t allow a ton of air balls: Even in 2016, with his diminished velocity, he gave up a fly ball or line drive to 38.9 percent of batters faced, compared to a league average of 37.9 percent. And while Boston is something of a bandbox, it can’t compare to Chicago’s Guaranteed Rate Field, which is more homer-friendly to righties and lefties.

Home runs can fluctuate a bit from year to year, since they’ll always come from a small sample — last year, Sale’s adversaries hit 27 of his 3,431 pitches out of the ballpark. This year, however, he’ll have the more expansive confines of Fenway Park for half his starts, which along with the strikeouts is reason enough for optimism. The lower home run total that ZiPS endorses checks out in my mind.

Batting average on balls in play

You know that old adage about pitchers exercising little control over the outcomes of balls put in play? Sale's career BABIP encapsulates that pretty handily:

Image via FanGraphs

In his five seasons as a starter, Sale’s gone from average, to sub-average, to sub-sub-average, to way-above-average, and back to sub-sub-average. Over that span, his BABIP as a whole sits at .293, on which ZiPS takes the over and Steamer takes the under (in this regard, the projections are flipped, perhaps because of their different methodologies). With something this volatile, where should we place our faith?

The flukiness shown in that above graph makes coming to a firm conclusion difficult. This graph, though, clears things up a little bit:

Image via FanGraphs

As noted earlier, Sale seemed to prefer a pitch-to-contact approach last year; ironically, that contact was actually pretty solid. After finishing with an exit velocity of 86.2 mph in 2015, Sale jumped up to 89.2 mph in 2016. That dropped him from the 23rd-lowest to the 139th-lowest among pitchers with 1,000 pitches. Perhaps because of his decreased fastball velocity, batters started to square up the ball against Sale in 2016, which doesn’t bode well for 2017.

Not all hard contact is created equal, of course. Since balls on the ground go for hits more often than balls in the air, a fly-baller like Sale could maintain a sub-average BABIP even with a higher exit velocity. He’d have a hard time doing that in Boston, though: Steve Staude’s research from 2013 showed Fenway Park inflated BABIP more than any other stadium in the AL, and trailed only Coors Field (obviously) among all ballparks.

The .279 BABIP Sale sported last year seems like an anomaly — his hard-hit rate shot above the major-league average, and even without many grounders, we’d still expect him to have a more pedestrian BABIP. With more balls bouncing off the Green Monster in 2017, his BABIP should inflate to the level ZiPS prognosticates.

Strand rate

The last — and possibly flukiest — area of our analysis is also the one where the projections diverge the most. What fun! In 2016, the MLB-wide strand rate was 72.9 percent, whereas Sale left 76.6 percent of his runners on base; Steamer thinks he’ll come down to the average, and ZiPS foresees him rising far above it.

In this area, it’s pretty difficult to make anything resembling a case for Steamer. Over the past five seasons, Sale’s had a mediocre strand rate just once, in 2015 (when he prevented 73.2 percent of runners from scoring). From 2012 to 2016, he put up a 77.4 percent strand rate, the 12th-highest among qualified starters. Five years of this performance would seem to establish this as Sale’s true talent level.

And we have every reason to think that’ll continue. For one, Sale has always locked down under pressure:

Sale by situation, 2012-16

BE 2548 27.5% 4.9% 2.7% .294 .228 .276 .357 .279
MoB 1536 28.1% 5.8% 2.4% .291 .221 .285 .354 .279

Most pitchers fare worse once someone has reached; in 2016, the average hurler had a .315 wOBA with the bases empty, and a .322 wOBA with runners on. That Sale has done just as well from the stretch as he has from the windup — and that he’s allowed such a low wOBA regardless of the situation — testifies to his persistent ability to strand runners.

Plus, Sale’s gotten a lot better at holding opponents in place. Each of his years in the rotation, runners have tried to steal less often:

Baserunners vs. Sale

Year SB opportunities* SB attempts %
Year SB opportunities* SB attempts %
2012 269 24 8.9%
2013 295 21 7.1%
2014 221 8 3.6%
2015 285 9 3.2%
2016 286 7 2.4%
*Per Baseball-Reference: “Plate appearances through which a runner was on first or second with the next base open.”

In 2013, FanGraphs’ Matt Klaassen showed that strand rate hardly correlated at all year-to-year — no other pitching metric vacillates quite like it. It’s possible that either projection system wins out, simply because of the random variation involved. With all this evidence, however, it’s hard to imagine Sale could stop stranding runners. The ZiPS prediction may be a bit sunnier than my liking, but it’s more likely to come to pass than Steamer’s.

In the end, that’s four victories for ZiPS and one for Steamer — although that’s not quite the whole truth. Really, it’s three areas (walks, home runs, and strand rate) where Sale has a relatively positive future, and two (strikeouts and BABIP) where he has a relatively negative one. The lefty remains one of the premier starting pitchers in the game, and regardless of how much he dominates in 2017, the Red Sox will be glad he’s donning their colors.

. . .

Ryan Romano is the co-managing editor for Beyond the Box Score. He also writes about the Orioles for Camden Depot, and about politics for The Diamondback. Follow him on Twitter if you enjoy angry tweets about Maryland sports.