Tom Glavine was voted into the Hall of Fame recently, but if you look at his career statistics and compare them to today’s pitchers, his line almost looks like it comes from a different century — the 19th, not even the 20th. For the first three-plus seasons of his career, he struck out around five batters per nine innings and walked three. Only one qualified pitcher had anything like those ratios last season, and it was Dallas Keuchel, who had an ERA over 5.00. Somehow, Glavine’s ERA was a shade under 4.00 for that period.
Advertisement
That’s not the Glavine who made the Hall of Fame. No, it’s the Glavine who showed up for the next eight seasons, when he put up a collective ERA under three and became one of the steadiest frontline pitchers in the National League.
Still, even in his prime years, Glavine struck out six per nine and walked three. In terms of statistics that stood out during his great run, there are two, and they are weird — at least statistically weird. The things that Glavine did best between 1991 and 1998 were: suppress homers and leave men on base. Why are those weird? Those are the two statistics that contain perhaps the most noise in baseball.
About 150 plate appearances into a given season, we can start to trust a pitcher’s demonstrated strikeout rate. That’s about the time their groundball rate becomes meaningful as well. Walk rate takes a little longer, maybe about halfway through the season. Left-on-base rate, which measures how often runners on base score? It never becomes meaningful within a season. Home run rate? Derek Carty found it takes almost three seasons before you can trust a pitcher’s home run rate.
So you’d be forgiven if you didn’t trust Glavine’s 1991 breakout season; your process would’ve been fine. But over those eight seasons, Glavine did teach us that players can demonstrate the ability to strand runners and suppress homers; we just have to give them enough time. That mean’s a player’s age and career rates are hugely important. By focusing on statistics that become meaningful quicker and then using the larger career sample to consider the rest, it should be possible to find the pitchers on the cusp of better days.
Home run rates
Let’s take a look at the Glavine method in action. First, take all the pitchers who threw over 70 innings last season and sort them by home runs per nine innings. Here’s the bottom half in homers per nine, sorted by strikeouts-minus-walks (the strongest in-season statistics), with their ERA and their career home rate included. A good bounce-backer should have a high ERA, a high strikeout-minus-walk rate and a homer rate that is much worse than their career rate.
| Name | IP | ERA | K-BB% | HR/9 | Career HR/9 |
|---|---|---|---|---|---|
193.1 | 2.84 | 25.4% | 1.54 | 1.34 | |
180.2 | 4.63 | 24.6% | 1.30 | 1.04 | |
107.2 | 2.59 | 23.1% | 1.59 | 1.12 | |
166.1 | 4.22 | 22.8% | 1.52 | 1.15 | |
178.2 | 3.53 | 20.7% | 1.36 | 1.33 | |
90.0 | 4.10 | 20.4% | 1.40 | 1.40 | |
92.1 | 4.19 | 20.3% | 1.95 | 1.95 | |
179.1 | 3.91 | 20.3% | 1.25 | 1.16 | |
119.1 | 4.68 | 19.9% | 1.28 | 1.28 | |
73.1 | 3.07 | 19.7% | 1.60 | 1.78 | |
120.0 | 5.48 | 19.6% | 1.88 | 1.60 | |
180.0 | 3.75 | 19.2% | 1.20 | 1.41 | |
89.2 | 4.52 | 19.0% | 1.91 | 1.91 | |
143.2 | 4.39 | 18.9% | 2.13 | 2.16 | |
105.2 | 4.17 | 18.8% | 1.28 | 1.47 |
The answer is Yu Darvish. Among qualified starters last year, Darvish was ninth in strikeouts-minus-walks, nestling in nicely among most of the Cy Young finishers in the top 10. He also had the second-worst home run rate in the top 30 by strikeouts-minus-walks. Our law of samples says he should continue to do well in strikeouts-minus-walks, have a better home run rate next year, and bounce back.
Advertisement
But there is at least one factor beyond the control of the player that sticks out when you look at Darvish’s career home run rates. In 2017, the league’s home run rates took a lurch forward as the baseball itself was different. Lower seams meant less drag meant more homers — but Darvish suffered more than most. Split his numbers into two periods around that ball change, and it leaps off the page.
| Period | ERA | Darvish HR/9 | MLB HR/9 |
|---|---|---|---|
thru 2016 | 3.29 | 0.91 | 1.01 |
after 2017 | 3.84 | 1.39 | 1.29 |
Steamer projections over at FanGraphs have Darvish giving up 1.24 homers per nine last year and returning to that high-threes ERA with tons of strikeouts. But it’s possible that there’s something about the way that he pitches — he’s very heavy on breaking balls and cutters in particular — that has led to such a homer explosion with today’s ball. At least his spin rates returned to normal by the end of the season.
Left-on-base rates
The real Glavine genius in the statistics seems so paltry when you sum it up: he had a career left-on-base rate of 73.9 percent, while the league average tracked very tightly around 71.5. Not the stuff of glowing biographies or Hall of Fame speeches, but pitchers have batters on base so often this stuff matters. Repeating the approach from above, here are the pitchers who had low left-on-base rates last season.
| Name | ERA | K-BB% | LOB% | Career LOB% |
|---|---|---|---|---|
4.63 | 24.6% | 66.8% | 74.0% | |
4.68 | 19.9% | 67.9% | 67.9% | |
5.48 | 19.6% | 68.0% | 72.6% | |
3.76 | 16.5% | 65.8% | 72.7% | |
5.08 | 16.3% | 59.4% | 69.5% | |
5.05 | 16.1% | 66.4% | 72.8% | |
5.03 | 16.1% | 67.5% | 70.1% | |
4.67 | 13.9% | 67.8% | 75.7% | |
5.46 | 13.7% | 59.9% | 63.6% | |
5.26 | 13.6% | 65.9% | 71.4% | |
4.91 | 13.3% | 67.8% | 69.1% | |
5.40 | 12.4% | 66.0% | 62.6% | |
6.27 | 12.3% | 68.5% | 71.6% | |
6.09 | 12.0% | 64.0% | 70.8% | |
6.22 | 11.3% | 65.1% | 70.4% |
Aaron Nola got the double whammy, as his homer rate was higher than his career rate, and he was also unable to strand runners this year. But to get a sense of the impact here for Nola, consider that his ERA last year was 4.63. If you did nothing but replace his left-on-base rate last season with his career rate, his ERA last season would have been 3.93. And that’s saying nothing about his inflated homer rate.
You might notice one weird thing about this list compared to the list above. There aren’t as many good pitchers on it, really. That’s because there’s a link between strikeout rate and stranding runners, one that makes some sense: Batters can’t usually advance on strikeouts, and a high strikeout pitcher could go to the well a couple of times in a row to keep a guy from scoring from third. Generally, left-on-base rates don’t strangle a season by themselves, but in Nola’s case, they sure made a difference.
Advertisement
Stuff and location metrics
One of the reasons we developed a metric for Stuff (the physical shape, spin, and velocity of a pitcher’s pitches) and Location was that they moved even faster than strikeouts and walks, some of the faster-moving stats in the pitching pantheon. So it’s really only following the method further to try and identify pitchers who threw the ball well but got poor results last year, even if we’ve left strand and homer rates behind.
The good news is, we’ll find some of the same names on this list, which compares starting pitchers (minimum 750 pitches thrown) by last year’s ERA- rank (from Baseball-Reference) against their Pitching+ rank (more on Pitching+ here).
| Player | Stuff+ | Location+ | Pitching+ | ERA+ Rnk | P+Rank | Diff Rank |
|---|---|---|---|---|---|---|
111.7 | 101.5 | 107.5 | 455 | 66 | -389 | |
94.1 | 104.0 | 101.0 | 663 | 281 | -382 | |
113.1 | 102.1 | 106.6 | 456 | 75 | -381 | |
96.6 | 104.2 | 102.8 | 562 | 187 | -375 | |
101.6 | 101.2 | 102.3 | 575 | 207 | -368 | |
102.5 | 99.2 | 104.2 | 479 | 135 | -344 | |
99.1 | 102.7 | 104.2 | 480 | 136 | -344 | |
102.8 | 98.8 | 100.2 | 661 | 326 | -335 | |
97.9 | 105.9 | 101.6 | 572 | 248 | -324 | |
100.2 | 105.3 | 103.6 | 472 | 160 | -312 |
Aaron Nola and Yu Darvish bookend some other names we’ve seen before, and their inclusion here seems to cement them as good bounceback candidates. Their stuff was good, their location was good, their strikeout rate was good, they just didn’t strand runners and maybe gave up a few more homers than you’d expect.
Carlos Carrasco figures prominently on this list, too, and he deserves more than a passing mention. The good news is that he had above-average location of a full arsenal — he threw four pitches over 10 percent of the time last year — and that his combination of shapes, speeds, and locations should have produced a slightly above-average starter instead of one with an ERA over 6.00. Tim Britton did a deep dive that pointed out that the beginning of the season was disastrous and he was better as he went along, but he also referenced this graphic from the archives of Stuff+ that showed Carrasco’s stuff has been in decline.
But the changeup still rates as a positive by Stuff+, it’s mostly the four-seamer and the slider which have gone from average or better to decidedly below average. Carrasco lost movement in both directions on the four-seam (less ride and less fade) and he also threw a harder slider with nearly three inches less drop in 2021. Could he fix those things? Stuff is stickier year to year than location, and so maybe Carrasco is a lesser bet than Nola and Darvish, but the Met is still an interesting potential bounce-back pitcher, especially with a more normal offseason in place.
You’ll notice intriguing young names like Luis Patiño and Logan Gilbert on these lists, and they’ll make their own lists this offseason, but in the true nature of a bounce-back article, they don’t have the track record we want. It’s a bit more like dreaming on a young Glavine because he had the excellent changeup and you thought he could put things in order around that pitch. But, since we started this article with such a legendary changeup, let’s end it with a young pitcher with a great changeup who has a little longer of a track record than those other two, and yet appeared on two of our three lists so far: Chris Paddack.
In 2021, Paddack had an above-average strikeout minus walk rate and yet allowed nearly 40 percent of the batters he put on base to score. If you gave him the league average in left-on-base rate last season, his ERA would have been 3.80 instead of the 5.07 he actually posted. Among changeups thrown 250 times, Paddack’s was fourth-best by Stuff+, and his curveball has inched forward to league-average by that metric as well. He’s always had good location numbers. The indicators are good … provided his arm is healthy after a stem cell injection to treat a slight tear in his elbow.
But at least the changeup is there, a changeup so good it could still launch a career, years after it first debuted. We could still be waiting for that career after 2022 is done and over, for sure. But what can you do, other than to look at the stats that have been proven to be more meaningful, look past the ones that need a career to define, and look for the pitchers who are shaping and locating their pitches well?
(Photo of Darvish: Gregory Bull / Associated Press)