Box reader gv27 posted the following nifty tidbit in last night's What's In a Number thread. I started out writing a comment in response, got all carried away, and thought I'd better post it separately.
I thought I'd share an oddity I found while flipping through the Baseball Encyclopedia. Certainly, this doesn't happen often. I found three players who drove in 100-runs in a season while hitting LESS then 10 home runs. Check this out:
Bill Brubaker ('36 Pirates): 6-HR 102-RBI
Bob Elliott ('45 Pirates): 8-HR 108-RBI
Tommy Herr ('85 Cardinals): 8-HR 110-RBI
Tommie Herr! I remember Bill James writing about that at the time. Something to the effect of "it's not because Vince Coleman hits in front of him, it's just a fluke."
Of course, it wasn't always a fluke. Before 1920, this happened all the time. Honus Wagner drove in 100+ runs 9 times, but his season high in HRs was 10, which he did twice.
How did he do it? Well, he was Honus Wagner, for one thing. Wagner was a slugger - he led the NL in slugging pct. six times, and was 2nd three more times.
The last time Wagner drove in 100 runs was 1912, when he led the league with 102. (I know everything about 1912, as is well known.) What was different then?
Triples, for one thing. A great slugger, like Wagner, hit lots of triples. In 1912, Wagner had 20 3B and 7 HR. Granted you don't drive in quite as many runners with a triple, because you're not driving yourself in. But otherwise a triple does just as thorough a job as a HR of clearing the bases.
In 2004, NL teams averaged 4.64 runs per game, which is happily very close to the 1912 average of 4.62 runs per game. But in 2004, NL teams averaged 177 HR; in 1912, they averaged 36 HR. How did they get the same number of runs across the plate? As we shall see, other types of extra base hits aren't the answer.
In 1912, NL teams were averaging 9.1 hits and 3.2 walks per game: 12.3 runners. In 2004, those numbers were 9.0 hits and 3.4 walks: 12.4 runners. Nevertheless, the hit component was significantly different, and in such a way as to lead us to expect many many more runs from 2004 teams.
Here are the components of NL offenses in 1912 and 2004:
Note - we don't actually have caught stealing data for 1912, which is why I marked it in bold. I have plugged in the number 110 as an estimate, and I'd better explain my reasoning.
We have very good reason to believe that 1912 teams were not stealing bases at a 72% success rate, like in today's NL. Probably not even close.
NL catchers averaged 77 assists in 2004. Teams were averaging 33 caught stealing, so we can make a crude guess that catchers are making about 44 assists on plays in the field. It would actually be slightly higher, because you can have a caught stealing without a catcher getting an assist. Obviously not all of the 44 assists came on the 74 sac bunts - there'll be nubbers in front of the plate, failed sacs where the catcher throws out the lead runner, failed plays at the plate where the catcher nails another baserunner, etc. etc. All kinds of stuff, but just an average of 44 plays in the season.
NL catchers averaged 210 assists in 1912. Who were they throwing out? Mostly, it had to be people trying to steal bases. There was a lot more bunting in 1912, it's true - an average of 96 more sac bunts in 1912, and we have no idea how many people were attempting to bunt for a base hit and getting caught. Still, it's not enough to account for all the extra assists, even if a catcher was making every last play.
Because I have absolutely nothing else to go on, I'm going to assume that 100 of those catcher assists in 1912 were on plays in the field. That's simply using the same ratio between sac hits and catcher assists (without caught stealing.) It's the best way I can think of to imagine how many more fielding plays catchers had to make.
And that leaves 110 catcher assists which just may represent a reasonable guess at caught stealing. I think that's a conservative estimate, myself. It suggests that teams were stealing bases at a 64% success rate in 1912. I don't think they were that efficient. In the 1912 World Series, there 16 stolen bases and 14 caught stealing: a 53% success rate.
But what all this means - at last, we cut to the chase - is that the NL in 1912 was losing more baserunners, via double plays and caught stealing, than NL teams were in 2004. Many, many more baserunners. And as we already noted, they had slightly fewer people on base via the hit or walk to start with (12.3 per game as opposed to 12.4); plus they hit hardly any home runs. How did they drive in and score just as many runs?
Errors. Lots and lots of errors. They were playing infield with glorified oven mitts. It means that NL teams in 1912 had loads of baserunners who actually show up as outs in the batting record. Except there they were, standing on first base, dancing off the bag and thinking about getting to second.
How many more errors? NL teams averaged 96 errors in 2004. NL teams averaged 250 errors in 1912. Some of these would have been advancement errors; they didn't put anyone on base, they just advanced a runner. How many? I don't know. There were 28 errors in 1912 World Series; 18 of them put a batter on base who was supposed to be out. The rest of them either allowed a runner to either advance or stay on the bases when he should have been out (because a fielder dropped the ball.) I have no idea if that 64% figure from the 1912 Series applies to regular season errors. If it does, it suggests each team had essentially an extra 160 hits that don't show up in the batting record.
Anyway, that's another reason Hans Wagner was able to hit 7 HR with 102 RBI. And that was nothing; in 1901 he had 6 HR and 126 RBI.
That's my quick and dirty analysis; there are Roster members and regular readers who know way, way more than I do about how to look at defensive numbers (Robert Dudek! Tango Tiger! Calling out to you!) and make some sense out of them. Maybe they can help account for this... phenomena?