Winning percentage in multi-homer games
Posted by Andy on February 4, 2010
In 2009, there were a total of 276 instances of a player hitting at least two homers in the same game. This counts some games twice when more than one player did it in the same game. In 210 of those games, the player's team won while in 66 of them, his team lost (including this September game in which Chris Young homered three times.)
Anyway, that means that in 2009, teams won 76% of the time when one of its players hit at least 2 homers. That's pretty good. Let's take a look at historical values:
I'm surprised that the win percentage hasn't changed all that much over time. Since 1954, it's been between 70% and 80% most year except for some isolated spikes, particularly in 1966, 1972, and 1979. I would have expected that in the Steroids Era, the win percentage would have gone down a little bit since more overall scoring means additional homers gives a team a little less of an advantage than it used to.
Perhaps if we could plot this data all the way back to 1900, we'd see that win percentage was consistently higher when homers were quite a bit less common.
For what it's worth, I also calculated the ratio of instances of 2+ HR games each season to the total number of games that year. It's not quite right to call it a straight percentage of games with 2+ HR performances since some games feature two or more such performances by individual players. However, the ratio was right around 10% in the late 1950s, dropped to 8% by the late 1960s, then leveled around 7% by the late 1970s and through the 1980s (save the 1987 spike to almost 13%.) In 1993 it jumped to 9.5% and in 1994 it was 12.5%. Eventually it peaked in 1999 just shy of 15% and has slowly dropped since then, down to 11.4% in 2009. Still, given how much more common it is today than in, say, the 1970s, I would have expected the overall win percentage to be a little lower today, but it is not.
February 4th, 2010 at 11:45 am
Any way you can check the number of runs that resulted from those HR's? There's likely more runners on base in a higher run scoring context. Two HR's in 1968 might mean 2 or 3 runs while two HR's in 1998 might mean 4 or 5.
February 4th, 2010 at 11:48 am
Good point David. I'm not sure if there's an easy way to check this but I can probably do it manually for a few of the years.
February 4th, 2010 at 12:05 pm
I did some manual calcs.
In 2009, the 2+ homers comprised 971 RBI over 560 HR, or 1.73 RBI/HR.
In 1989, it was 440 RBI over 261 HR, or 1.69 RBI/HR
In 1969, it was 605 RBI over 348 RH, or 1.74 RBI/HR
That's just a small amount of data but suggests that it doesn't change all that much.
February 4th, 2010 at 2:35 pm
To Andy: How was it 9.5% in 1993 and also 12.5%?
February 4th, 2010 at 2:38 pm
That was a typo, thanks.
February 4th, 2010 at 5:06 pm
You are most very welcome.
February 4th, 2010 at 5:23 pm
That very low point for 1966 really stands out, and I can't think of a reason why 1966 should be any different from 1965 or 1967, so (unless there's a mistake in the data) I guess it just shows how much truly random variation there can be in this statistic.
I note that there have been 15 instances of a player hitting 4 home runs in a game, and the player's team won 13 of those games.
February 4th, 2010 at 11:50 pm
Doc Gooden once surrendered three homers on an opening day to KARL "TUFFY" RHODES, a guy who was a mediocre star in japan for most of his playing days, and Doc still managed to win the game. I'm not sure of the year, but '94 sounds right.
I think RHODES had maybe <20 career HRs. Although I was too young to realize it, it was one of many nails in Doc's baseball coffin.
Despite a grand canyon talent deficit, they probably have the same net worth (dollars) today.
February 5th, 2010 at 3:02 pm
I think that, rather than expecting a decrease in win percentage during the "steroid era," the more telling thing is that the percentage stabilized in the late 1990s. With more teams having multi-homer games it should tend to stabilize the statistic.
I don't think you should have expected to see a decrease in the percentages since more teams had the ability to hit more than one home run per game.
February 5th, 2010 at 3:24 pm
Anthony (or should I say apreziosi?) I had the same thought--the variation is less because in theory there are more multihomer games, so the variation drops purely as a function of the sample size getting larger. My data posted above in comment #3 confuses me though, about how there were so many multi-HR games in 1969. I will need to double-check that.
February 5th, 2010 at 3:25 pm
Gerry, why do you question my data so much? Heh.
Good ol' Gerry is always questioning me 🙂
February 5th, 2010 at 5:04 pm
Andy, why do you post such weird data?
Seriously, though, don't take it personally, as it isn't intended that way. Indeed, I usually don't even notice who it was that posted the data I question. And I think you have to admit that the two things I've questioned lately do look very strange.
February 5th, 2010 at 5:19 pm
I admit nothing!!!!!
Heh, kidding of course. I didn't take it personally Gerry 🙂 I just thought it was funny that you questioned data twice. Since graphs like the one in this post involve some manual work on my end, however, questioning is a smart thing to do.
February 6th, 2010 at 12:31 am
I'm amazed that over 10% of games feature at least one player hitting 2+ home runs. It seems like a very special accomplishment, but by those numbers, you should expect to see one every single full day of baseball. Never would have guessed it.
February 7th, 2010 at 9:23 am
I have to agree with Zachary. That's astounding. In fact, it's over 11%, and with 15 games on a full day of ball, you could almost expect to see 2 such games every day (1.7)!!!
February 12th, 2010 at 8:05 pm
Bob Horner once hit four home runs in a Braves loss.