The year 2003 feels like a very long time ago for me. I had just moved into my first “big person” apartment (which is now six addresses ago, depending on how you count things) in Chicago (two cities ago). My wife was still my girlfriend (our long-distance year), and my little brother graduated from high school. In more important news, it was the last year in which the National League compiled a winning record against the American League in interleague games. That year, the Senior Circuit got the better of 137 of the 252 games conducted between representatives of the two leagues. Since then, the American League has run off a string of 11 (and likely to be 12 in a couple of weeks) seasons in which they have emerged victorious.
From 2004-2015, American League teams won 54.7 percent of their games against National League opponents during the regular season. In the twelve World Series that have been played in that time period, the leagues have split them with six apiece (Thanks, Giants!), although 10 of the past 13 All-Star games (and in fact, sixteen-and-a-half of the last 20!) have been won by the American League. Of course, we’re playing fast and loose with some very arbitrary endpoints, but even a cursory glance at the evidence shows that there’s some imbalance in the AL/NL “rivalry” over the past decade.
Last week, former Baseball Prospectus editor-in-chief Ben Lindbergh, along with former Baseball Prospectus writer Michael Baumann, former Baseball Prospectus writer Dave Cameron, and… Jeff Sullivan gathered together on an episode of The Ringer MLB podcast to try to figure out why this was. Is it possible that the explanation goes deeper than just “sometimes the coin just comes up heads 12 times in a row?” If it were a fun fact where one team had dominated another over the course of a decade in some meaningful way, we might be able to pull apart some specific practice where those teams differed, but how to handle a whole league? Is it possible that the American League is somehow better than the National League?
Leagues encompass a heterogeneous bunch of folks and philosophies, and there’s constant movement back and forth within and across the leagues both in terms of players as well as decision-makers, and in the case of the Astros, entire franchises. Technically, only one General Manager from 2004 is currently still in his job (Brian Cashman of the Yankees), though there are a few other cases of General Managers being “promoted” though still retaining many of the responsibilities historically associated with “General Manager,” such as Billy Beane and Brian Sabean. Still, in the last decade, there’s been a change at the top for most teams. But is there something inherent in the American League that makes for a better baseball team? Has the American League—on the whole—been better about doing some crucial thing that a team needs to do to win?
The tempting answer is to simply invoke the ecological fallacy and say that it’s “because of the DH.” That’s the most visible difference between the two leagues, and it could be the case, but we need a little more evidence than that. So, let’s see if we can come up with a reason why the American League has been better than the National League.
Warning! Gory Mathematical Details Ahead!
The Sunday Night Baseball Hypothesis
We noted above that from 2004-2015, the American League has won a total of 54.7 percent of the interleague games played. One thing that we know for certain about the American League is that it is (and has been) the home of the New York Yankees and the Boston Red Sox, historically the two biggest spending teams in baseball. Here’s a graph showing the average Opening Day payroll, by league from 2004-2016.
Case solved! It’s because the American League spends more money than the National League!
Well, not really. We know that the Red Sox and the Yankees tend to outspend everyone, and when we remove them from the payroll data, we get this graph:
It looks like these two groups have spent roughly the same amount as each other over the years. And when we look at “interleague games since 2004 not involving the Yankees or Red Sox” we see that the American League teams have still won 53.6 percent of them. Even more, the 12-year streak is pretty much intact, although in 2004 and 2011, the non-Yankees/non-Red Sox teams played exactly .500 ball against the NL. It doesn’t seem to be a mega-money issue.
OK, Maybe It Is the DH?
There are two strains of argument on this one. One is that National League teams struggle specifically in games in American League parks because their teams have been built with the assumption that for 152 games out of the year, they will be playing without a DH. So, during those 10 games each year where they get one, they end up designating some random bench guy while the American League team can insert the big, lumbering, if defensively challenged guy whom they signed last offseason into their lineup.
The evidence supports this, halfway. I’ve previously estimated that moving from non-DH rules to DH rules brings an extra .11 runs per game to a National League team’s bottom line. For the American League, on the other hand, the effect of going from DH rules to non-DH rules is actually a penalty of about .46 runs per game. That cuts both ways; while NL teams might not be able to reap the full benefit of the DH rule, they should be cleaning up when it comes to games in NL parks and the AL has to play without a vital cog in their offense, right?
From 2004-2015, the American League really has had an advantage over the National League in AL parks, winning just over 60 percent of the time. (Yes, you read that right.) Now, the AL’s success could be because of the DH or it could be because the AL was just better. Swinging around to the other side of that analysis, we see that the AL also won 49.1 percent of their games in NL parks (ha! NL wins!), although given that the road team in general only wins 46 percent of games (and the AL team is always the road team in an NL park), it means that the AL is ahead of that curve as well. The DH imbalance probably contributes to the difficulties that the NL has in interleague play, but even when the AL’s great advantage is taken away, they still seem to have been better over the years.
The other strain of argument on why the DH might be affecting things is that because of the presence of the DH, American and National League teams construct their roster in different ways, and that the American League way of constructing a roster turns out to be more efficient. This falls apart when we look at the evidence around that. We know that designated hitter is actually one of the most highly compensated positions on the field… erm, on the team. Teams also seem to pay more per win for their DH than they do for other positions on the open market, so the DH is both expensive and inefficient, on the whole. American League teams aren’t required to sign an expensive DH (though the numbers suggest that they do anyway), but National League teams are excused from doing so. That means that (Yankees and Red Sox aside) AL teams begin their payroll with the same basic number of dollars to spend and then make an expensive and inefficient purchase that NL teams don’t have to make.
Maybe there’s a triple bank shot theory that says that the presence of the DH caused AL teams to realize that they needed to find a way to cover for the fact that the DH was so inefficient that they needed to innovate more. (In the same vein, maybe it was the presence of the Yankees and Red Sox that made AL teams more likely to really search hard for an answer, but do we really believe that the NL teams were on cruise control?)
I suppose that there’s an argument to be made that the designated hitter might rob some talent from the National League. Some of those guys could be playing on NL teams as defensively challenged first basemen or left fielders. There’s also the idea that the presence of a DH allows an American League team to be able to take a chance on a player who might quickly lose his defensive value (or not have any defensive value to begin with) because if the defense doesn’t pan out, there’s a place for him. Maybe some of that’s in there, but let me instead suggest that we consult Occam’s Razor on this one.
Another argument that gets raised at places like—ahem—Baseball Prospectus is that the teams who were earliest on the bandwagon of hiring people from places like—ahem—Baseball Prospectus were in the American League. (Somewhere in Pittsburgh, BP alumnus and Pirates’ Director of Baseball Informatics is probably wondering if his team has been needlessly batting the pitcher all these years and no one bothered telling them.)
Did the American League just have better functioning Front Offices?
Well, one way that the AL could show off their better brains was doing better in scouting and player development. To test that hypothesis, I looked at data from Baseball America’s Top 100 prospects and sorted them by league. We see that from 2004 to 2011, the trend line was generally downward for the NL. (It’s worth noting that before 2013, the NL had 16 teams to the AL’s 14, so if everything were equally distributed, we would expect the NL to have a slight edge in the raw count of prospects.) This suggests that the AL was doing a better job on the whole than the NL in developing prospect currency, although that trend reversed starting in 2012. It’s also worth pointing out that prospects don’t always pan out.
The point of having good prospects is to have a good pipeline of young players who can be paid “cost controlled” wages and still put up value (or traded). How to measure which league was better at this is actually a little tricky.
· We can look at how much cumulative WAR cost-controlled players put up in each league. This gives us a raw amount of value, although it can be inflated by one league or another just having more young players around. (Also, the NL having 2 more teams for many of the years in question.) That can be a feature too, because if one league has a lot of cost-controlled players around to fill spaces, even at replacement level, that can be helpful.
· We can look at that on a per-player basis. This solves for the problem above, but could be confounded by one league bringing up a lot of cup-of-coffee guys.
· We can make sure that we’re only looking at players who meet a certain cut-off for playing time (I chose 50 IP or 250 PA as proper), which will artificially inflate our per player WAR estimate, but gets around the cup of coffee problem.
Here are the graphs over the years for position players who were in their first, second, or third years. The first graph is the cumulative WAR put up by those players by league, and the second is the average WAR per player with a 250 PA minimum. We see that in terms of the absolute number of WAR put up by freshmen, sophomores, and juniors was fairly even over most of those years, although again, the NL had a two-team advantage during those years. But, looking at the per player estimates, we see that the American League was just better at developing (or acquiring?) players who could put up more value. (Note: I extended the window to the first six years of a player’s career, which would cover both their minimum wage and their arbitration-eligible years, and those graphs told the same story.)
Here are those same graphs for pitchers from 2004-2015.
Here we see that with pitchers, the AL got more out of its young pitchers on an absolute basis (again, even spotting the NL two extra teams), and in most years on a per-player basis (min 50 IP).
But that wasn’t the only way in which American League front offices made their bones. The other way in which “smarter” front offices are supposed to help a team is that they can identify bargain-priced players in the free agent bin. So, I compared the two leagues in their abilities to nail down value from players who were free agent eligible and were making less than $5 million per year. I took the per-player WAR and again we see first for position players, then for pitchers, that the AL front offices were better bargain hunters than their NL counterparts. (Because of a data error, I could only go back to 2005 on this one.)
So, we have pretty good evidence that the American League may have been beating the National League in the most boring way possible: developing better young players and signing bargain free agents.
Make the National League Great Again
This may not fully satisfy everyone out there. There may be a person or two reading this saying “Yeah, but what was the real driving force behind the AL having better players?” I think that the real “question behind the question” is whether there’s something inherent in the American League, on a structural level, that makes it better than the National League. Is the National League doomed? I doubt it. If the answer really is that American League teams, on the whole, made decisions that turned out better, then that dominance was either a function of luck, which by definition can’t be predicted, or skill, and there’s no reason to believe that a new generation of brilliant baseball ops people would avoid the NL specifically. In that sense, it’s just the coin coming up heads 12 times in a row. Unlikely, but possible.
Let me question the entire premise though. This entire time, we have used the fact that the American League has won more interleague games as “proof” of its superiority. And you might think “Well, what else should we use? It seems like a pretty obvious measure of comparison.” I’d agree that there’s evidence here that over the past decade, the American League has been head and shoulders above the National League, but I think that almost in spite of what the interleague record says.
We saw above what we knew intuitively about the one structural issue that does separate the leagues, the DH, and how it specifically plays out in those interleague games. We estimate that the NL picks up an extra .11 runs per game from being able to use the DH for a couple of days. This is because they generally don’t carry a proper DH, and are playing with their eight regular position players and a ninth guy who’s really a glorified bench player. The American League, on the other hand, is playing with their eight regular position players and a ninth who gives them a .46 runs per game boost over those eight regulars. The obvious advantage goes to the AL team in the AL park. When the teams go to a National League park, both sides are playing with eight position player regulars. Yes, the AL team will score less than they normally do on average, but AL teams generally score more runs than do National League teams (in 2015, AL teams scored 4.39 runs per game while NL teams scored 4.11).
As we saw, the NL has had a slight advantage in NL parks over the past decade, but the AL absolutely cleaned house when they could use the DH. That pattern will probably continue. I don’t think that the AL will continue their streak to infinity and beyond, but I think that we have to acknowledge that the playing field is tilted slightly toward them in interleague games. It’s possible that the NL will eventually catch up and surpass the AL in terms of drafting, developing, and finding bargain bin guys. And at that point, we can sincerely call them “better” than the AL, at least on the things that we actually want a team to be better at. And eventually, they’ll win a season series, but I could see some years in which the NL was measurably better and yet, the AL won the head-to-head matchups based on the DH problem.