Power Factor: A Better Way to Measure Power
As the sabermetric revolution has enhanced our knowledge of the game, new stats have given us better ways of assessing players. When it comes to power, for example, chicks still dig the long ball, but most serious analysts have begun to see the problems with using arcane stats like RBI.
The current statistic of choice for statheads trying to measure players’ abilities to hit the ball far is Isolated Power. ISO is simply the difference between a player’s slugging percentage and his batting average—or, in other words, his extra-bases-per-at-bat. By subtracting BA from SLG, we reduce the impact of a player’s contact ability so we can get a better idea of his raw power.
But while it’s improvement over simple slugging percentage, ISO has its flaws. It, too, is largely a product of batting average, and as such can fluctuate with BABIP and contact skills—both of which are irrelevant to a statistic called “Isolated Power.”
Let’s take two hypothetical players: Tony Stark and Bruce Wayne. Stark hits .350 with a .550 SLG. Wayne hits .200 with a .400 SLG. Both weigh in at a .200 ISO, suggesting their raw power is roughly equal.
But that’s not right. If Wayne somehow managed to bring his average up 150 points, would all the extra hits be singles? Stark is clearly a better hitter, and they both produced the same amount of extra bases. But if the two players had similar contact skills, Wayne would undoubtedly be a bigger slugger.
Let’s take the real-world example of 2010 teammates Carl Crawford and B.J. Upton. According to ISO, Crawford (.188) showed a little more power than Upton (.187). But did Crawford really outslug Upton, or was his extra-base edge simply a product of having more hits?
Of Upton’s 127 hits, 60 went for extra bases; Crawford managed 62 extra-base hits out of the 184 times reached he safely. Yes, Crawford had more extra-base hits, but a much higher proportion of Upton’s hits (47%) were doubles or better than Crawford’s (34%). An Upton hit was 50% more likely to go for extra bases than a Crawford base knocked, yet ISO said their power was virtually identical.
You can glean some useful information from ISO, but clearly “Isolated Power” is a product of contact skills and luck, and therefore it does not actually isolate power.
Enter Power Factor, or “PF.” PF removes contact skills and BABIP fluctuations from the equation by expressing power as a proportion of batting average:
For the sake of neatness, I use ISO instead of SLG in the numerator; to convert my numbers to the more traditional format, simply add 1. In this incarnation, PF could also be described as extra bases per hit.
As such, the possible range of PF stretches from 0 (i.e., a batter who hits only singles) to 3 (a batter whose hits are all home runs). For some perspective on what PF scores look like, here’s a basic percentile chart:
PF allows us to compare power between hitters with varying contact skills. Returning to our earlier example, B.J. Upton had a .789 PF in 2010—29th among qualified hitters—while Crawford weighed in with a below-average .612 PF, good for 84th in baseball. For some context, Kevin Kouzmanoff’s PF was .603.
Other high-contact hitters whose power seems less formidable in this light include Adrian Gonzalez (.714 PF), Vladimir Guerrero (.653), Shin-Soo Choo (.613), Hanley Ramirez (.583), and Andrew McCutchen (.570). At the other extreme, Mark Reynolds (1.182), Carlos Pena (1.077), Aramis Ramirez (.876), Matt Kemp (.807), and Lyle Overbay (.778) rank higher than they would on a basic ISO chart.
PF also gives us insight about whether a slumping hitter is maintaining his power, or if his lack of pop is the cause of his offensive struggles. For example, Toronto Blue Jays DH Adam Lind saw his average slip from .305 in 2009 to .237 in 2010, and a suffered a corresponding 69-point drop in his ISO. But his PF changed by a much smaller proportion, going from .843 to .793, suggesting that he’d have no problem regaining his status as an intimidating slugger if he can put the bat on the ball more often.
On the other hand, PF shows us that some apparent power surges are merely the result of increased contact. Take Chicago White Sox outfielder Alex Rios. In his bounceback 2010 campaign, Rios upped his ISO by 25 points—was he actually more powerful, or was it simply the product of more balls falling for hits? That his PF barely budged—it moved from .599 in 2009 to .609 last year—suggests the latter.
Of course, it’s fun to play with numbers and these findings are interesting, but that doesn’t mean every statistic should be taken seriously. You could multiply each MLB hitter’s pitches-per-plate-appearance by his shoe size and it would mean nothing (though it sure would be fun to see the league leaders). Even if it makes intuitive sense, we have to make sure the model fits with reality—especially with a formula as admittedly crude as this one.
How does PF hold up to empirical data? As it happens, it does very, very well. Among the 153 players who had at least 400 plate appearances in both 2009 and 2010, PF had a robust season-to-season correlation of .719; among the same group of players, ISO managed a correlation of just .664.
That’s not a huge difference, but it’s a clear one. PF is more consistent year-to-year than ISO. That implies that PF is a better method of measuring power than ISO.
But wait! There’s more. Correlation between 2009 ISO and 2010 ISO was .664, but correlation between 2009 PF and 2010 ISO was .675. That’s right, ladies and gentlemen—PF is a better predictor of future ISO than is current ISO.
If the same argument is used to explain why FIP is better than ERA, can’t we assume that PF is a more accurate measure of raw power than ISO?
Maybe I’m off the mark here. It’s certainly possible that I missed something, and if I did I’d love to be proven wrong. At the very least, there’s probably a way to improve the model. But given the combination of PF’s intuitive appeal and mathematical consistency, it seems clear that this is a better way to measure pure power.
Edit: In case it wasn’t clear from my explanation of how I tweaked the formula from its “traditional format,” I am not the first to invent Power Factor. I had the idea for it before I discovered that it already existed and I’ve never seen anyone else do this kind of analysis, but I wasn’t the first to come up with the idea of dividing bases by hits.