The Minnesota Twins are clutch. They have managed to hit .314 with runners in scoring position this season, despite hitting .280 overall. Their .314 average with RISP is the highest of any team going back to the year 2000 (I couldn’t find pre-2000 data). And, unsurprisingly, the difference between their overall batting average and their average with RISP is also the largest of any team since 2000. Is this random variation, or are the Twins on to something?
The Twins’ season is within the realm of statistical fluctuation. Since 2000, teams have hit .266 overall, and .267 with runners in scoring position.
In our dataset, we see a normal distribution of batting average with RISP, around the mean of .267. Overall, 68.5% of teams had an average one standard deviation from the mean – between .251 and .283. There were 96.2% of teams that fell within two standard deviations of the mean – between .234 and .299. And every single team in the dataset fell within three standard deviations of the mean – between .218 and .315.
Interestingly, the difference between overall BA and BA with RISP was also normally distributed. However, the Twins’ difference of 34 points fell exactly three standard deviations from the mean. But even this is within the realm of random statistical fluctuation: there are 270 samples in our dataset (30 teams, nine years). So, if the data were normally distributed, we would expect exactly one team to be outside of three standard deviations from the mean. The Minnesota Twins are that team. And, luckily for them, they are hitting better with RISP, rather than worse.
From a macro perspective, we may write this Twins season off as random variation – chance which is benefitting the Twins, but which is not of their own doing (and therefore is not repeatable). They’re just happening to get a very disproportionate amount of hits when runners are in scoring position.
But this is baseball, not stats 101: we want an explanation. And the Minnesota Twins have been good at hitting with RISP for the last several years:
(Note: this chart was made about a week ago; the Twins are now hitting .280 overall and .314 with RISP this year, and still have a 34-point difference)
At first glance, it appears that perhaps the Twins are onto something: they’ve been better with RISP every year since 2004 (a negative difference signifies that the Twins were better with RISP than overall). But perhaps not: my hypothesis is that any team has a 50-50 chance of performing better (or worse) with RISP in any given year. Thus, while the Twins have performed better with RISP for the past five years, they performed worse for each of the four years before that. Meaning, that in the last nine seasons, they have performed better five teams – which is consistent with my hypothesis.
The Twins are also known for an organizational philosophy emphasizing "small-ball" and abhorring strikeouts – perhaps this has something to do with their increased performance with RISP over the past five seasons. I ran the numbers in my database, and found that the correlation between strikeout rate the difference between BA overall and BA with RISP is .20. That’s rather low.
The numbers do show that the Twins tend to strike out less often than other teams: overall, players have struck out in 19% of their at bats, while players on the Minnesota Twins have struck out in 17.59% of their at bats. However, this strikeout percentage is not the lowest in baseball since 2004, when the Twins began their RISP prowess – the Angels and Orioles have lowest K percentages in that time. Here is how they have fared with RISP during that stretch:
As you can see, neither team has performed nearly as well as the Twins with RISP during this time.
Furthermore, this season the Twins have struck at the second-lowest rate in the majors. The only team who has a lower strikeout rate is the Seattle Mariners. And the Mariners are hitting 12 points worse with RISP than they are overall.
The Twins have been extremely good with runners in scoring position this year, but their performance is within the realm of random statistical fluctuation. Furthermore, the possible explanation that their emphasis on putting the ball in play has allowed them to hit so well with RISP seems to hold no water.
While it is possible that the Twins are able to identify players who hit well with RISP and/or play a specific type of baseball that causes players to hit well with RISP, the evidence in this study suggests that the Twins have simply been beneficiaries of random variation. And next year they, like every team, will once again have a 50% chance of having a higher batting average with RISP than overall.
Post-script:
Here are the other teams who have had the highest difference between their overall batting average and their batting average with RISP in the past nine years, and how they have fared in the following season:
2003 Royals: .274 overall, .304 RISP (+30)
2004 Royals: .259 overall, .267 overall (+8)
2006 White Sox: .280 overall, .307 RISP (+29)
2007 White Sox: .246 overall, .243 RISP (-3)
2005 Angels: .270 overall, .296 RISP (+26)
2006 Angels: .274 overall, .274 RISP (0)
2007 Tigers: .287 overall, .311 RISP (+24)
2008 Tigers: .273 overall, .267 RISP (-6)
2007 Tigers: .287 overall, .311 RISP (+24)
2008 Tigers: .273 overall, .267 RISP (-6)
2003 Athletics: .254 overall, .277 RISP (+23)
2004 Athletics: .270 overall, .260 RISP (-10)
2004 White Sox: .268 overall, .291 RISP (+23)
2005 White Sox: .262 overall, .259 RISP (-3)
2002 Astros: .262 overall, .285 RISP (+23)
2003 Astros: .263 overall, .262 RISP (-1)
As you can see, only one of these other eight managed to beat their overall BA next season.