SLG vs. ISO: Allen Craig's Power Outage

We've started unveiling the 2014 Viva El Birdos Community Projections here at VEB this week. On Wednesday, Rui wrote up the community's collective projections for the Cardinals' cleanup hitter, Allen Craig. I found it interesting how many VEBers predicted that Craig's power-hitting would return in 2014 after a substantial falloff in 2013, especially given the contours of that power outage. I thought it would interesting to examine Craig's power outage in 2013 using Slugging Percentage (SLG) and Isolated Power (ISO).

ISO vs. SLG

There are two stats that measure a player's power-hitting: SLG and ISO. SLG measures something different than ISO does, and each has its own distinct formula. Their differences make ISO the superior stat for measuring strictly power-hitting.

We all know that some hits are better than others. A home run is better than a triple, which is better than a double, which is better than a single. The more bases a player touches en route to his ultimate destination on a hit, the better. Batting average (BA) doesn't reflect this reality. It treats every hit the same. In BA, a single is worth the same as a homer, triple, and double. A player has a .300 batting average whether he went 180 for 600 with 50 homers or 180 for 600 with no dingers.

SLG was developed as a metric to reflect the type of a hits a player notches. It bases hit values on the number of bases a player takes on a given hit. Singles are multiplied by one, doubles by two, triples by three, and homers by four. The formula for calculating SLG is:

( [ Singles ] + [Doubles * 2 ] + [ Triples * 3] + [ Home Runs * 4 ] / At Bats

The problem with SLG as a metric is that the weight it gives each type of hit is completely arbitrary. A homer is not worth four times as much as a single or twice as much as a double. Yet that's how SLG weights them.

The other problem with SLG, at least in terms of its value as a metric measuring power-hitting, is its inclusion of singles. This means that a player highly skilled in batting for average can inflate his SLG with a high BA, even if he hits a lot of singles.

For an example, let's compare the 2013 seasons for Marlins outfielder Giancarlo Stanton and Cardinals second baseman Matt Carpenter. Carpenter posted a .481 SLG, which was one point higher than Stanton's. This even though Stanton hit 13 more homers than Carpenter in 213 fewer PAs. Making a wider comparison, 47% of Stanton's hits went for extra bases while 37% of Carpenter's hits did the same. Yet Carpenter's SLG, a rate stat, was higher than Stanton's. Using ISO, we get a much different picture of Stanton and Carpenter's 2013 batting than with SLG.

ISO excludes all singles, so it only measures a batter's extra-base hits (XBH). This means it reflects only his power-hitting and not his singles hitting. ISO is a metric that demonstrates a player's raw power. The easiest way to calculate ISO is simply:

SLG - BA

The formula strips one layer of the SLG formula. A more complicated ISO formula is:

( [ Doubles ] + [ Triples * 2 ] + [Homers * 3 ] ) / At Bats

ISO is not perfect by any means, but it's better for measuring power than SLG. A comparison of Stanton and Carpenter's 2013 ISOs shows the difference in using the metric. Carpenter's SLG was one point higher than Stanton's. But Stanton's ISO was .231 to Carpenter's .163. The 68-point difference between Stanton and Carpenter's ISO better reflects than SLG the fact that Stanton hit for a larger share of XBHs.

ISO vs. SLG: Allen Craig's Power Outage

Now let's look at Craig's power-hitting drop between 2012 and 2013 using ISO and SLG. The following graphs show Craig's SLG and ISO and the MLB average SLG and ISO by year, which gives us a baseline against which to compare his drop.

SLUGGING PERCENTAGE (SLG): ALLEN CRAIG vs. MLB AVERAGE

ISOLATED POWER (ISO): ALLEN CRAIG vs. MLB AVERAGE

It's important to note that Craig's BA went up by eight points from 2012 to 2013, yet his SLG dropped by 65 points. Craig's ISO, which focuses only on XBH, fell by an even larger amount. Craig experienced a 12.45% drop in SLG and a 33.95% drop in ISO. Because of Craig's high BA, he still had a higher-than-average SLG despite having an average ISO in 2013. Craig's ISO shows that he hit for a lot fewer XBH in 2013 than in 2012, but SLG doesn't show the drop as clearly because of Craig's high, single-infused BA.

The following chart shows the percentage share for singles (1B), doubles (2B), triples (3B), and home runs (HR) out of Craig's overall hit (H) totals for 2012 and 2013.

Craig's XBHs fell across the board as a share of his overall hits. Meanwhile, his singles total and share shot upward. In 2013, Craig turned into a singles hitter. It will be very interesting to see if the VEB community projections for Craig's 2014 prove true and his power returns.