Greek God of What?

Kevin Youkilis had a huge year in 2008, shedding the mantle of "Greek God of Walks" for something a little more, well, powerful. He had always been a very good hitter, especially in the realm of plate discipline, but last year, he became arguably the best hitter in a loaded Boston line-up, leading a team that included David Ortiz, Manny Ramirez, J.D. Drew, Jason Bay, and Dustin Pedroia in such favourites as wOBP, wRAA, REW, Fangraphs' Batting Runs, SLG, and OPS, many by wide margins. While Youk's .390 OPB was again outstanding, it was right on line with his career averages (.390, .381, and .400 in the previous three years). The .312 BA was a good step up, but not really outside what you can expect to see from a .290 hitter every once in a while. It was the .569 SLG, a full .116 points higher than his career best set the year before, that vaulted him to a slightly less dismissible position on Olympus. And behind that spike in SLG, of course, was a .092 point spike in isolated power (ISO).

The good news for Youk is that ISO is generally accepted as the more stable component of SLG. There's certainly some truth to that. ISO regresses much less heavily than BA, and the correlation is much stronger for ISO from one year to the next (the correlation coefficient-closer to 1 means stronger correlation-for ISO from one year to the next over the past 6 paired seasons is .74, compared to .42 for BA). However, ISO, on average, also changes more each year than BA. The average player's BA only changed 8.2% (.023 points) from one year to the next, while his ISO changed 20.0% (.034 points)on average. I know that sounds like it doesn't make sense, and I won't go into all the details of why, but basically, it is because ISO is spread out much more between players even though it is on a smaller scale. So while a great hitter in ISO might only be at about .250, he can be over .100 points ahead of other players just as easily as a good hitter in BA can be .050 points ahead of hitters the same distance behind him. A good hitter in ISO can drop .050 points and still be good in ISO much more easily than a hitter can drop .050 points in BA and still be good. Compared to the spread between talent in each stat, ISO is more stable. But players will also see larger swings in their ISO.

That's a lot to take in, and I don't blame you if you just skipped it. Instead of poring through it again, take a look at the following graphs. The x-axis is a player's BA/ISO in one year, and the y-axis is his BA/ISO in the next. The line shown is the line of best fit, and the slope gives you an idea of how much each stat tends to regress toward the mean (greater slope means less regression):




While BA is clearly more scattered and more prone to regression, it has much less area to move around in. Batting averages are packed into a tighter range than ISO. This added range for ISO to jump around in means that Youk, or any hitter, is actually more likely to take points off his SLG from ISO than from BA. That doesn't mean ISO is not a good indicator of power. It's an excellent indicator of power, and it separates good hitters from poor much more significantly than BA. We do, however, have to take its stability with a grain of salt, especially after large jumps like we see in Youkilis' 2008.

In the past 5 paired seasons for which we can examine both the jump in ISO in 1 year and how any large spikes in ISO regressed the following year, there have been 13 hitters who saw 1-year jumps of at least .090 points. All but three of them dropped significantly from that peak the following year. The following graph lists all 13 players along with how big a spike they saw in ISO and how far their ISO dropped the following year (note the significant "busts" among the list).


Spikes in ISO greater than .090

Yr. of Spike	Player	Spike	Regression
2003	Edmonds	.092	-.001
2003	Varitek	.113	-.054
2004	Beltre	.110	-.137
2004	Se. Casey	.093	-.099
2004	Konerko	.093	-.007
2004	Ar. Ramirez	.092	.006
2005	C. Tracy	.122	-.075
2005	Ensberg	.137	-.046
2005	Br. Roberts	.099	-.077
2006	Beltran	.172	-.071
2006	Durham	.106	-.120
2006	Teahen	.097	-.101
2007	A. Rodriguez	.099	-.075

So one-year spikes, at least in recent history, do not seem to be very sustainable. A more gradual build-up may be, but chances are very high that Youkilis sees his ISO fall back below .200 in 2009. He could still maintain a solid improvement over his previous high of .165, but a lot of the big numbers he saw in 2008 that derived from that huge ISO are likely not going to be there again.

Likewise, you should always be careful at projecting the continuation of power surges simply because they are propped up by the more stable ISO rather than a high BA. If the ISO is well out of line with a player's career, then it is likely to see some significant regression just like you would expect from, say, a BABIP that is well out of line with career norms. Steady progression in ISO, like Youk has had in the past few years, is a good sign, but when you are dealing with a one-year spike this big, scale it back to the career path you were seeing before. For Youk, that means continued improvement from his 2007 levels, but nothing like the 2008 production. So Youk, don't be too quick to drop that title as "Greek God of Walks". After all, it might not be there forever.