Elite Starters and Pitching Trends Revisited

Last year I wrote a series [Part 1, Part 2, Part 3] on whether we were properly valuing elite starting pitchers. The basic premise was that we were likely to see fantasy-relevant starters (not openers) throw fewer innings again in 2019. I hypothesized that if there were fewer and fewer starting pitchers providing 200 innings of excellent quality, then elite starters were being undervalued, especially in leagues using quality starts.

That research led me to investigate which pitchers were most likely to produce elite seasons. I concluded that pitchers who were coming off an elite season were more likely to return top-tier, if not elite, value. As a result, I argued that those pitchers should be drafted more aggressively than starters going just one or two rounds later. I'm condensing the points a bit, but that was the spirit of the series.

Did those conclusions hold up? I think so, but I’ll let you judge.

Trends in Pitching Use

2019 extended some but not all of the patterns I described last year.

Certainly, the expansion of the starter and the increased reliance on bullpens continued to eat into starter innings. Innings per Start dropped from 5.4 (base 10) in 2018 to 5.2 in 2019. Likewise, the number of quality starts dropped from 1,996 down to 1,794, a 10% year-over-year decline.

Conversely, 2019 had three more complete games than 2018, and one more starter who reached 180 IP. This year, it occurred to me to check how many of those pitchers with more than 180 IP were able to maintain an ERA better than league average. It turns out that number increased from 25 pitchers in 2017, to 28 in 2018, and to 30 in 2019. That’s two straight years of growth. Those increases are modest enough that maybe they are an anomaly. Or it may be that despite reduced starter usage, managers and teams have calculated how and when to use starters more effectively. Those increases are minor enough that they aren't a clear reversal in the trend, but they are significant enough to suggest that we've reached the saturation point for relievers and the trend for starters has leveled off or even started to rebound.

One potential reason to be skeptical of that is the addition of the extra roster slot for 2020. That seems like a further opportunity for teams to lean on their bullpen to get through games, especially as more teams focus on load management. On the other hand, teams won’t be able to rely on lefty specialists to carve through perilous sections of an opposing lineup, so maybe we will see managers give more opportunities to their starting pitchers.

At the very least, 2019 was close enough to what I expected that the core concept about the scarcity of top-tier starters held true, but if anything, the situation seems more complicated than it did last season.

Were Elite Starters More Valuable?

Last year, I defined an “elite starter” as a starting pitcher who returned a value of at least $26.88 in a $260, 12-team league while using a 70/30 split. My thoughts on that have evolved some, but I’m going to keep that definition because it’s still functional. I picked the original setup because it offered middle-of-the-road settings with an aggressive hitting-to-pitching split that recognized how most leagues allocate more space and budget to hitters rather than pitchers. Here are the top 20 pitchers for Wins leagues in 2019:

And here is that data graphed against previous years:

The 11th through 20^th best starters actually improved relative to the top 10, but the stratification that I described last year is still present in this data set.

From the data above, we can see how elite starters were exceptionally more valuable than other starters just a few spots below them. For example, Gerrit Cole ($42.4) was marginally more valuable than Clayton Kershaw ($20) and Walker Buehler ($18.4) combined. That’s confounding to think about, but here's the comparison:

Some of the difference is that 212 innings of a 2.50 ERA might not look like it provides twice the value of a 3.15 ERA, but Cole’s ERA value is worth about $7 compared with the $3.50 of value from Kershaw and Buehler’s 360.2 innings of 3.15 ERA.

To illustrate the difference in the stratification between top-ranked hitters and the top-ranked pitchers, here are the top 20 hitters for that same league:

In Wins leagues, there were 30 hitters between Justin Verlander at $45.8 and Zack Greinke at $26.0.

Meanwhile, Quality Starts leagues had markedly similar results.

Surprisingly, 2019 had the closest alignment I’ve seen in pitcher value between QS leagues and Wins leagues. Based on what I saw last winter, I would have expected QS leagues to have an even greater exaggeration in elite-starter value, but in 2019 they had slightly less stratification. I don't have any reason to explain that except for my hypothesis above that teams are finding ways to use their starters more effectively.

We can say with confidence that elite starters were still exceptionally valuable compared to starters just a tier below them. Of the three elite starters in Wins leagues and the four in QS leagues, two were players from the pool of elite starters the year before. The other two were candidates I identified last season.

Do Elite Starters Return Their Draft Cost?

Here things get trickier. Early-round picks provide little opportunity for profit. Managers can either hope that picks return their value in an absolute sense OR that positional scarcity will provide greater profitability over other players at that position.

As I wrote above, the second half of my conclusion was that elite starters were most likely to come from those starters who’d been elite the year before and who had been a top-100 player in the previous season. For 2019, that meant Chris Sale, Jacob deGrom, Max Scherzer, Justin Verlander, and Corey Kluber. The core conclusion that I reached last year was that those five starters were significantly more valuable and more likely to return fair value than those starters going immediately after them. That part turned out to be mostly true.

The five starters did outperform those starters going nearby, but the real story here was that six of the top-10 starters provided $5.10 of value or less. Those top-five starters, the ones who were at least top-100 players in 2018 and then elite in 2019 provided an average value of $20.06 with a median of $20.70. By comparison, starters 6-10 provided an average value of $9.24 with a median value of $0. If you were taking one of the first ten starters off the board, you wanted one of those top-five players. It gave you much higher odds of getting your money's worth out of the player.

The average value of $20.06 for those five starters was still below my expected average of $28.20 for a post-elite starter. Maybe that’s just year-to-year variance. Maybe it is the small sample size of five players with a single year’s data. After all, 2018 did offer the highest number of elite starters (six) at any time in the last eight years. Perhaps it made sense that we would see a correction back to only three elite starters in 2019. Regardless, for elite pitchers, it was the greatest variance in year-after results since 2013, and it’s difficult to look past.

Moreover, the elite starters came from where we expected: Verlander and deGrom were on last year's list, and Gerrit Cole was one of the candidates I tagged as most likely to emerge as an elite starter. For QS leagues, Flaherty was on the list, but the measures weren't nearly as confident in his ascent.

Sale, Kluber, and Snell

Last year, I concluded that previously elite starters were more likely to provide useful (even if not elite) seasons to fantasy owners. That didn't happen for Corey Kluber, Chris Sale, and Blake Snell. Corey Kluber’s 35.2 innings had a net value of $0 while Chris Sale’s injury disrupted 25 starts provided a pitiful $5. Both of those players fell below my 20^th percentile outcome of previously elite pitchers. If it happened to only one of them, I’d feel better. However, for two of the five to suffer from that fate, it feels like a distinct loss. Obviously, we're talking about a difference of a single-player season, and maybe just one injury, but it raises questions about the stability of elite pitcher value in comparison to those hitters going near them. I'll address this more at the end.

Blake Snell was a different case, I think fantasy managers had reason to hope he could repeat, and his injury luck feels even more random than Kluber's or Sale's. However, Snell didn't have the previous track record of players likely to be elite starters. Consider Gerrit Cole’s status this season. Compare Cole’s excellent 2018 and elite 2019 to Blake Snell’s mediocre 2017 and elite 2018. Blake Snell exemplifies the type of elite starter who was less likely to repeat with another elite performance. That’s not to suggest that Snell won’t bounce back this season, but his lack of track record didn’t have clear analogs among pitchers who provided elite seasons.

In contrast to Snell, Cole’s two-year combination doesn’t guarantee that he will again offer elite value, but based on the elite starter performances of the last seven years, he and Jacob deGrom are about as safe a bet as anyone in the league.

Justin Verlander is a close third, but the prospect of age regression is awfully steep as he enters his age-37 season. In the seven years of data I’ve examined, only one pitcher was elite at the age of 36, Justin Verlander. To find another example of a pitcher generating an elite season at the age of 36 or older, I had to go back to Roger Clemens at the age of 42 in 2005. Clemen's use of PEDs makes him a problematic comparison, but at this point, I think we can say that Verlander, like Clemens, has had a Hall of Fame career. That makes him exceptional in ways that will defy categorization.

Conclusion

Despite 2019's chaos, we do have relative clarity on where elite starters come from and that those players who offered elite value are more likely than other early-round starters to return their value in the season after. After I had written the first two articles in the series, my editor Alex Roberts referred me to a piece from Ariel Cohen that drew a similar conclusion. I would certainly recommend that piece as well.

The last point I thought about after publishing the series last season was about early-round hitter value compared to early-round pitcher value. Last year Corey Kluber, as the last elite starter off the board, was going as pick number 24, and Aaron Nola was going as pick number 25. That provides us with a simple breakpoint to divide the difference between elite starters and second-tier starters. We can then compare that to the hitters going in that same range. Here's the difference:

There are different ways to cut up the data for this, but most of the ones I tried gave me similar results. Early round hitters are the safer play, but there is a more significant difference in value between the elite starters versus another starter going just one or two rounds later.

I'm not inclined to make the argument that this means everyone should be rushing out to draft deGrom, Cole, or Verlander. The evidence above could easily be interpreted as a demonstration of why the LIMA strategy is more useful than ever.

In the next article, I'll go into how I'm planning to draft starting pitchers this season. Certainly, I want to get one of those three arms, but I'm concerned about the recent "pocket aces" trend, and my guidance is not as simple as "spend your first-round pick on the best starter available."

More Fantasy Baseball Analysis