The Era of Encroaching Dinger Reliance
Back in 2014, the average baseball team scored 4.07 runs per game. That was down only a tenth of a run from the year before, which was down only a little more than a tenth of a run from the year before that, but a definite trend was emerging. An average of 4.07 was the league’s lowest mark since 1981, and there were gathering concerns that offense was being suffocated. No one really knows how low is too low, but Rob Manfred considered various ideas that might re-inject some hitting. This is why conspiracy theories persist to this day.
Rather conveniently, see, offense bounced back in 2015. It surged again in 2016, and it’s surged only more over the past five months. The average team now is up to 4.68 runs per game, which feels more familiar. The surge has been powered by a well-publicized and well-examined home-run spike, but at the end of the day, offense is offense, right?
It is, and it’s good that hitters again have a chance. The balance of power had felt like it was shifted too far. But in a certain sense, you could argue that this offensive surge is artificial. For a variety of reasons, home runs are up, and they’ve gone up right when they needed to. But offenses now are so very home-run reliant. Everything to follow is probably obvious, but I might as well explicitly lay it out. Home runs are taking over the game.
For some of you, this’ll be review. Or, if not review, it’ll cover trends you already assumed. Let’s go back to 1950. Here is a simple plot, of the rate of home runs per plate appearance, league-wide:
Yeah, there’s been a home-run spike. You know that, and your neighbors know that, and your neighbors’ children know that, and your neighbors’ children’s pets know that. But keep in mind this isn’t a plot showing home-run rate on contact. This is home-run rate, overall, and even despite the league-wide increase in strikeouts, we’re at a relative maximum. The highest rate in the sample belongs to 2017. The second-highest rate in the sample belongs to 2016. There’s less contact than ever, but there are more home runs than ever. The difference is made up in the following plot.
This shows the yearly rates of time-on-base events, excluding home runs. So it’s a combination of singles rate, doubles rate, triples rate, walk rate, and hit-by-pitch rate. Another pattern emerges:
The league isn’t presently anywhere ultra-low, but there’s still an obvious dip, relative to the previous few decades. This is a consequence of the strikeout spike. Those strikeouts have to eat away at something, and they’ve eaten away at non-homers. As a consequence of both of the trends imaged above, you can see that runs are being distributed differently.
This shows average runs per game scored on homers, and average runs per game scored on non-homers. You see that the lines are going in different directions.
With 1.99 runs per game scored on homers, that line is at its highest point. But then, there have been just 2.69 runs per game scored on non-homers, and that’s the eighth-lowest point in the sample of 68 seasons. You can observe that the red line has kind of flattened out. Although league-wide offense is climbing back up, it’s entirely because of the longball. As another way to visualize that, we can combine these numbers into one.
Enter the Guillen Number, a measure initially created by Joe Sheehan several years ago. The Guillen Number is simple. The denominator is all runs scored. Just plain overall runs. The numerator is runs scored directly on homers. How reliant is the league on home runs these days? The trend is unmistakable.
Again, we’re looking at 68 seasons. The third-highest Guillen Number came in 2015, when the league finished at 37.1%. The second-highest Guillen Number came in 2016, when the league finished at 40.1%. The single highest Guillen Number for now is happening in 2017, with the league sitting at 42.6%. In other words, about 43% of all runs are being scored on homers. That’s three out of every seven. While it might not feel like the league is yet all that close to 50%, consider that it was at 33% as recently as 2014. Since then, the Guillen Number has increased by about 10 percentage points. Home runs are taking over team offense. I think most of you already knew that, but perhaps you hadn’t considered it like this.
Because I’m here, I might as well show you what this year’s team-by-team breakdown looks like. Try not to stare too long at the Giants.
The Blue Jays have scored 51% of their runs so far on homers. They’re the only team at something greater than half, but the Rangers are awfully close. More meaningfully, there have been 1,682 individual team-seasons since 1950. Five of the 10 highest Guillen Numbers are getting put up right now, in 2017. There are an additional three teams from 2016. The highest mark so far belongs to the 2010 Blue Jays, who finished at 53.1%. No one right now is threatening to depose them from the throne, but the league is moving forward as a collective. Again, this year’s league Guillen Number is 42.6%. Three years ago, there was just one team that topped that.
This isn’t inherently a good thing or a bad thing. It’s a matter of taste, and some people like movies for the dialogue, while other people like movies for explosions. Neither group of people is superior to the other, and taste feels like a silly thing to criticize. Just, understand what is presently going on. Offense is back, relative to a few years ago, but that’s happened entirely because of the home run. In the run-production pie chart, the home-run slice is getting wider and wider. Runs aren’t back because of anything else. Manufacturing a run isn’t suddenly back in style. Teams now manufacture their runs with big mighty hacks, and trying anything else can feel borderline futile. Home runs aren’t everything, but they’re closer than they’ve ever been. What that means to you is personal, but it’s something for each of us to grasp.
Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.
“…Rob Manfred considered various ideas that might re-inject some hitting.”
Nicely done.
Todays game has Earl Weavers mantra written all over it.
What’s wrong with three-run homers?
I will renew my ongoing plea to move the fences out. Fewer home runs, more triples! Please?
Much more democratic. Home runs are fascist, just like strike outs. Uber mensch etc. Let’s do 500 foot outfields so defensive shifts have 4 outfielders!
This would be awesome. You want more strategy? Try deciding how many outfielders you should have!
Worth noting multiple teams moved their fences in during that period when offense slumped in a few years ago. The Mets, Marlins, and Padres’ parks were all built during the 2000s and adjusted when power numbers dropped significantly. I’d be really interested to see how those parks would be playing today with their original dimensions– if the change would be significant, that suggests minor changes to park dimensions could bring major results in returning to a more normal HR environment.
It’s a bit amazing. After 20 years of trying to Chicxulub RBI’s out of existence, we saber-toothed quants are left with the one hit that looks remarkably like the RBI of old.
Hey – what’s with the downvotes? Saber achieved its goal of stripping context. No longer does anyone in the game think that singles/doubles/triples that produce RBI’s are worth anywhere near a HR that produces RBI’s. Maybe we didn’t expect the smallball extinction – but then nobody expects the smallball extinction.
What is Kluber’s Guillen Number Against? Seems like it would be well north of 50%. 48 Runs Allowed and 17 Homers Allowed, but that’s of course not all the information needed.
You could go through his play by play data to see how many runners were on when the homers happened.
It’s at exactly 50% for the season, but your interest in the matter may be enhanced by the fact that it’s at 75% since july 15th
Thanks!
Similarly for Sale, Scherzer, and other power type pitchers that don’t necessarily manage contact to an extreme degree but strike enough guys out and refuse to walk anyone that they are still the best pitchers in the game.
Huh: http://www.fangraphs.com/blogs/2016-al-starting-pitcher-contact-management/ Sale and Kluber both grade out very well in terms of contact management. Especially Kluber. Granted, last year, but still.
I get that there are many factors that go into this. I just find it interesting to see Colorado where they are relative to the other great hitting ball parks (TEX, TOR, MIL).
Not sure there really are many factors. COL is a pitchers park (or small-ball) – at altitude. OF is only a bit smaller than KC and both those are quite a bit bigger than 3rd place – and 1/3 acre larger than Fenway (smallest OF). And the OF subsections are consistently large too:
LF – KC > COL > PIT > non-outlier parks
CF – DET > HOU > COL > non-outlier parks
RF – KC > BOS > ARI > non-outlier parks (and COL is 4)
That’s a lot of territory to cover and no place to stick a statue-that-bats – so hits and xbh are gonna be high everywhere.
Add altitude and it just means the contact goes further and OF has to position deeper which opens a bigger gap to the infield as well.
Right, but what I was getting at is the considerable separation between COL and the other three teams. It would suggest there is something more than just fair/foul territory. No? Fenway has the “smallest” dimensions right? But they are still second to last on this graph. See what I’m getting at? Rockies are 21st in HRs, 28th in wRC+, but 4th in Runs and 5th in wOBA. Just wouldn’t peg them as a non HR oriented offense that’s all.
Additional note: Rockies team rank in HR: 10 (2016), 7 (2015), 2 (2014)
Sure there’s more than just OF dimension. Home team still has to hit HR’s since that is half the PA’s. And yes – BOS has smallest dimensions overall – but a big green thang designed to hurt HR’s on that side and a big RF. My point was that those OF dimensions show why Coors will, year after year after year, have higher non-HR (singles/doubles/triples) and altitude will raise babip and turn some xbh into HR. And Coors will always account for 50% of Rockies numbers. So the Guillen number (dependence on HR) for Rockies will always look much lower than Coors reputation as ‘hitters park’.
eg this year Coors still has the highest babip, 3rd highest singles, 3rd highest doubles, highest triples, and 10th highest HR. It’s still Coors – with lots of runs scored in lots of ways. It’s never going to be HR dependent. And the Rockies will always be unique because Coors is.
I think the homers are stil good but if the non homer runs fall below 50% it might get ugly.
I’m not one of those singles are better than HR guys but with all the talk about the HR craze it is only like 0.3 homers per team and game more than 5 years ago. Over a season that is huge but for the individual viewer that is just half a homer more per game and thus not that much more action.
We still only have like 2.5 homers per game of both teams so statistically it is like one homer per 90 minutes- not exactly a HR derby.that is not a ton of action if everything else declines (albeit that is not thay extreme either, most outs are still in play and not Ks).
I’m not too worried about the home run spike if it’s just due to players trying to elevate more. Pitchers will adjust and baseball will find a balance.
On the other hand, if the ball is indeed changed, that’s a problem.
It’s pretty well documented that the ball has changed. Manfred’s only response is that the ball is “within specified limits,” which is a pretty telling answer in my opinion. Just ensuring that more balls are consistently at the “juicier” end of the specified limits will result in more home runs without technically “juicing” the ball.
Funny thing is we are only a few years out from talking about ‘the royals model’ that included, among other things, having a lineup that doesn’t strike out much and relies on putting tons of balls in play.
That we are now worried about homers taking over the game should show you how quickly things can turn. A combo of adjustments and normalization is likely to make hot take alarmism (if it were at all introspective) feel silly in a couple years.
Or, alternatively, how hot our takes are. Even the ones that use graphs and data.
We need to lower the pitcher’s mound and unjuice the juiced ball. Then everything would work out fine.
I’m personally fine with it as is currently, seems like years ago everyone was always wishing there were more dingers. Now that there are more…maybe the sentiment has changed.
But I agree if it continues to increase that could be problematic.
That first graph reminded me that 1987 happened, and Wade Boggs hit 24 HR. *checks player page* Yup, slapped 24. Same player hit 2 HR playing a full season in Yankee Stadium.
Only 1 team in the top 10 Guillen Team numbers is in a playoff position, and that’s as a WC. GMs have to see this is no way to build a winning team, don’t they?
Interesting. This could be part of an on-going plot to turn MLB into a real-life version of the old Nintendo game “RBI Baseball”. If the Tigers sign a shortstop named Tramml then we know it to be true.
I have a soft spot for Baseball Simulator 1.000. Sometimes I feel like batters are using the “Hyper Hit” Ultra-move.
I’m going to quibble with the statement that “HRs are seizing control of all run production” (in the subhead on the FG home page), or at least the implied timeline.
Most of your thesis is spot on: HRs per game, runs from HRs per game, and the share of all runs that come by HR, are all now at record highs.
But some folks are saying that the recent HR surge is “driving out” other forms of scoring. That’s clearly not so. The HR surge started in 2015, but non-HR runs have been utterly flat since 2012, holding between 2.66 and 2.72 R/G (currently 2.68 by my count).
The big drop in non-HR runs actually came from 2007 (3.16) to 2012 (2.72). In that span, HR runs were fairly stable, starting at 1.64 and ending at 1.60, but total scoring fell from 4.80 to 4.32. Non-HR runs leaked away without any obvious link to HRs.
That 2012 rate of 4.32 R/G is pretty typical of the expansion era, whose average is 4.38. But the share of runs via HR was 37%, matching the high to that point. I saw no complaints about that share.
Since then, there were 2 years of modest scoring decline — reaching 4.07, all due to HRs falling, and causing much concern — and then the uptick since 2015 (all due to HRs rising).
I’m one of the many who find the current 42.6% of all runs via HR too high. But the tricky questions are what one sees as ideal, in terms of (a) total scoring and (b) the share of scoring by HR. I suspect it would be extremely hard to manipulate those independently.
Tangent: I find it fascinating that runs per HR has little correlation with OBP.
I split the years since 1960 in two groups by OBP.
— The low-OBP set averaged .318 OBP and 1.58 R/HR.
— The high-OBP set averaged .332 OBP and 1.59 R/HR.
Perhaps, since shifting has taken away many non-HR hits, and cut short the rallies needed to score without a HR, this is the consequence of shifting.
I’m surprised more players don’t just hit the ball into the gaping holes in the defence left by shifting. Some players could bunt for doubles, if they actually could bunt.
Homers are fun so that’s not the worst thing. But I think most would agree that balls in play are also fun, especially ones that are extra base hits. You never have as much extended live action suspense as you do when baserunners are running. I think it’s good for baseball if they can do something to get the ball moving around inside the park as well as out of it.