NL vs. AL: Which league is better?

No brainer, right?

Eric Karabell said yesterday (7/7/08 ) on his podcast that the AL was obviously the better league — I think he used the word “clobbered.”

And why not?  This year’s interleague record was 149 for the AL, to 102 for the NL.  That gives the Senior League a 0.396 winning percentage.  Friends, that’s close to Washington territory. (*)

*By the way, why did Jim Bowden think that the 2003 Reds would be any better in 2008 than they were back then? Dmitri Young, Kearns, Lopez, Aaron Boone — playing the Nationals is like Old Home Week.  Also, I’m stealing this footnote technique from Joe Posnanski.

Add to that the fact that the AL has won 8 of the last 13 World Series, and 10 straight All-Star games.  Case closed, right?

Not so fast.  When you look at the underlying numbers, it’s much less clear.

League	Runs	HR	OBP	SLG	OPS
American League	409	85	0.332	0.410	0.742
National League	401	90	0.329	0.410	0.739

Gee, that looks a little closer, doesn’t it?  Those numbers are per team, to compensate for the fact that there are 16 NL teams and only 14 AL teams.

It turns out that the difference is barely statistically significant.  In runs, it’s about 1.1 standard deviations; the NL advantage in HR is about 1.5 standard deviations.  The AL numbers are better, but only slightly better — not the kind of difference you would expect would generate the lopsided interleague results.  Furthermore, we would expect the AL to be better because of the DH.  Shaving 1/9 off the AL run numbers would put them lower than the NL.

Well, lets look at the players, then.  From Fangraphs.com, the 10 best players this year in WPA are Berkman, Burrell, Holliday, Manny, Bay, Pujols, Mauer, Chipper, Uggla, and Carlos Lee.  That’s 8 NL players.  The top 10 (actually 12, because of ties) in Runs Created from baseball-reference.com:

Berkman-HOU	95
Jones-ATL	88
Kinsler-TEX	86
Utley-PHI	83
Pujols-STL	80
Ramirez-FLA	79
Burrell-PHI	76
Holliday-COL	74
Sizemore-CLE	74
Bay-PIT	72
Bradley-TEX	72
Hamilton-TEX	72

That’s 8 National Leaguers of the 12.  Certainly looks like the NL can pull its weight here.

Maybe it’s the pitching?  Here’s Adjusted ERA+:

Duchscherer-OAK	194
Volquez-CIN	190
Lee-CLE	172
Lincecum-SFG	171
Danks-CHW	170
Haren-ARI	160
Sheets-MIL	156
Marcum-TOR	153
Zambrano-CHC	151
Billingsley-LAD	141

Only 6 of 10 here for the NL. But still! Around half of the best pitchers are in the NL.  And Santana, Sabathia, and Harden aren’t on that list (yet).

Jayson Stark had a chat about this topic after the first interleague weekend.  He took the position that the NL was at least closing the gap, if not even.  Of course, after that the NL really got clobbered in interleague play.

So, how can we possibly explain why the AL becomes unbeatable in interleague play?

Here’s the possibilities I’ve come up with.  Maybe you can add more.

1. The AL pitching and hitting are both better than the NL by precisely the same amount.  Thus, they cancel each other out in AL games, but show up big time in interleague.  But then how do you explain the preponderance of NL players atop the leader boards?
2. The NL has the stars, but the AL teams are better balanced.  The balanced teams win.  Okay, but is there a reason why we think a balanced team is better than a stars-and-scrubs team?  Also, apparently the stars and scrubs just happen to balance out to give PRECISELY the same overall stats as the even-keeled AL.  Sounds fishy.
3. AL teams make better use of the DH, because they actually have a good hitter on the payroll to play the position.  The NL teams are using utility infielders.  Two words: Jose Vidro.  Besides, that should only matter in half the games where the AL is at home.
4. The AL has more big-name aging stars.  The kids in the NL are simply awed, and fold.  C’mon, it’s the major leagues.  They get over it pretty fast.
5. The leagues are actually pretty equal; it’s just been a string of bad luck for the NL.  149-102?  That’s pretty unlikely — the binomial calculation puts it at 0.15% chance if the true odds are 50/50.
6. AL Three-True-Outcomes ball beats NL small-ball. Except that the NL has the lead in HR, BB, and K.
7. Wait, I mean that AL small-ball beats NL Three-True-Outcomes ball.  Um, now you are reaching.

It doesn’t make any sense to me.  The NL looks roughly equal to the AL.  If I didn’t know the head to head outcomes, I’d give the NL a slight advantage.  What gives?

The Frustrating LOB

Last night, the Reds lost 1-0 to the Pirates (!) leaving 12 on base. Tonight, they lost 4-3, leaving 11 on base. That, my friends, is frustrating. Last year, it seemed like the Reds were forever leaving men on. This year, George Grande and Chris Welsh are certainly upset about the missed opportunities. I decided to look it up, and see how bad it really is.

A well-posed question is half answered, so here’s the questions:

1) Which teams are best and worst in terms of scoring their baserunners? In particular, are the Reds as egregious about leaving men on as they seem to be?

2) How much difference is there between teams that are really good at scoring their baserunners, and teams that are bad? Is “baserunner scoring efficiency” a real trait, or is it just random?

With the help of ESPN’s team batting statistics, I got some data for 2007. They don’t list LOB, so I calculated it as follows:

LOB = Hits + Walks + HBP – CS – GDP – R

I think that’s correct. Once a guy gets on, he gets out or scores. Fielders choice doesn’t matter — on the play, there is no hit, and the number of runners does not change. Yes, I’ve left off all those guys called out on interference or failing to touch a bag or whatever. I think being picked off counts as a CS. I probably am failing to account for runners who are thrown out trying to take an extra base.  Those situations should be very small corrections to this analysis, and should not have any effect on the conclusions.

We can’t just look at this number in isolation, because good teams will have more baserunners, and thus will always have more LOB. So, we want to make a comparison based on the LOB opportunities.

The relevant comparison here is Baserunners, which is just LOB + R. This is a kind of “isolated baserunners” that ignores those runners who make outs on the basepaths. After all, I want to know the scoring efficiency of the men on base, as a result of their not being stranded. Maybe that’s a good way to think of it — an inverse strand rate for team batting.

Anyway, here’s the data from 2007. I normalized it to a per-game value.

If you are below the line, you are efficient with your baserunners, leaving fewer on base than most teams. The answer to question 1B) is immediately evident — the Reds are right about average. The Yankees not only get a lot of baserunners, but they leave comparatively few of them on base as well. Grrr. The Nats (aka the “Former Reds”) are the opposite — not many baserunners to begin with, and not too much success driving them in.

Why are some teams better than others? Maybe “well-balanced” teams are more efficient, because they more often have another good player coming up to drive in runners, while “single-superstar” teams are worse — that guy gets on and gets stranded. Or, maybe small-ball teams are more efficient, while Earl-Weaver-2-walks-and-a-homer team are less efficient? I don’t think there’s enough evidence to claim either one of those for sure.

How about the previous year? Here’s the same thing for 2006.

Man, look at those Yankees. And look how much Tampa Bay improved! Once again, the Reds are just about average.

So, is there any correlation? Here’s the Runs per Baserunner 2006 vs. 2007:

Teams to the top left have improved their efficiency, teams to the bottom right have degraded their efficiency. Now, this is efficiency only, so the overall number of baserunners is removed here. That’s why Tampa’s improvement looks small — their efficiency improved a little, their number of baserunners improved a lot.

Very poor correlation here, so you pretty much would have to say that the variation in R/BR is not a trait, but instead is random. Yeah, except that the Yankees, Tigers, and Rangers would say otherwise. I should probably go back and look at a few more years to see if some teams really can maintain above-average efficiency.

So, how much does it matter? That is, if your team could boost its efficiency back to the league average, how many more games would it win? I used the regression line for the 2007 data, and assumed that each team regressed to that mean. That is, I didn’t change their number of baserunners, just adjusted their runs scored as if their scoring efficiency were at the mean. Below is a table showing how many extra runs each team would have scored. Negative numbers mean that the team was more efficient than average already, so regressing to the mean would result in fewer runs scored.

Assuming 10 extra runs is an extra win, as a rule of thumb, Oakland and the Nats pick up 6 more wins, while NY loses 8, the Tigers 7, and the Rangers 6. Again, the jury is still out whether variations in efficiency are just luck (so we can say those teams were (un)lucky) or whether there is a good reason for the variations. Consistent with what we saw above, the Reds were just about average.

So, the questions at the top are just about answered. The Reds are about average for LOB. The total range in efficiency seems to be about 160 runs — pretty significant — with Washington and Oakland the worst, and the Tigers, Yankees, and Rangers the best. At first glance, it looks like there is no year-to-year correlation in efficiency, though the Yankees ongoing success is just enough to make you wonder.

(I couldn’t get this table any farther up, so it’s at the bottom. Also, this is my first sabermetric analysis — let me know what you think. If you’ve got other interpretations of these data, please comment!)

Team	Extra Runs
Arizona	12
Atlanta	-11
Baltimore	9
Boston	-1
Chicago Cubs	19
Chicago Sox	10
Cincinnati	-5
Cleveland	4
Colorado	-12
Detroit	-70
Florida	4
Houston	38
Kansas City	8
LA Angels	-45
LA Dodgers	44
Milwaukee	-39
Minnesota	25
NY Mets	-2
NY Yankees	-82
Oakland	59
Philadelphia	-26
Pittsburgh	18
San Diego	17
San Francisco	45
Seattle	-28
St. Louis	37
Tampa Bay	-2
Texas	-63
Toronto	0
Washington	64