Mike Francesa: Is the Alderson outfielder a player where on-base percentage comes first or power comes first, or both?
Sandy Alderson: Well, it would be nice to have a combination. If I could only have one, I’d give a slight nod to on-base percentage, but power is very important. Critically important.
– WFAN Interview, November 15, 2013
New York Mets General Manager Sandy Alderson set the world – or Queens, at least – on fire with these comments. The Mets have been in decline offensively over Alderson’s three seasons at the helm, scoring 718 runs in 2011, 650 in 2012, and only 619 in 2013. With Alderson at the helm, and time to install “his system”, the casual view is that the system is a failure. For three seasons the Mets’ approach at the plate has centered around getting on base and the results have been poor.
But it isn’t the system that has failed. By and large, it’s been the talent.
In 2011, Mets’ non-pitchers hit .272/.344/.405, with Angel Pagan’s .322 OBP representing the lowest of the everyday players. In 2012, .254/.322/.396, and Scott Hairston’s .299. 2013? .244/.314/.379, with John Buck and Juan Lagares getting on base at .285 and .281 clips, respectively. Strikeout rates climbed, from 16.1% in 2011 to 21.2% in 2013, as walk rates dropped from 9.3% to 8.5%. So it’s fair to say that it hasn’t been a result of the approach as much as a decline in talent.
Going beyond that though: is there actually merit to Sandy’s preferences?
Short answer: Yes.
Were we scientists, we’d try to formulate an experiment. We could track slugging percentage and on base percentage against runs scored and see how they correlate. And thanks to sites like Fangraphs, we can!
What I’ve done is exactly what I’ve described above. The graphs below compare the runs scored totals for teams over the past twenty years* to on base percentage, slugging percentage, and isolated power (slugging percentage – batting average). Keep in mind that, because the scales are slightly different, you should pay attention to the slope values for the trendlines and not their physical slopes.
*I’ve omitted the 1994 strike season as the lower run totals throw off the graphs
The first thing we see is all three are positively correlated – that is, a higher value for OBP, SLG, or ISO typically corresponds to a higher run total. Looking at the slopes of the trendlines, we see that the correlation is indeed most positive with on base percentage. The ratio of the slope of on base percentage to slugging percentage – roughly 1.81 – suggests that a one-point increase in OBP is nearly twice as valuable as an equivalent increase in slugging percentage (this 1.8 ratio has been confirmed previously).
The fact that the ISO slope is slightly higher than the SLG slope makes sense as well. ISO, by definition, is a measure of extra bases (where slugging percentage is a measure of total bases per at bat, ISO is a measure of extra bases per at bat). This makes clear sense – a double is more valuable than a single, and a home run moreso than a double.
Where the plan draws criticism (often unfairly), is the association of focusing on On Base Percentage and a perceived emphasis on walk rate. Later in the interview, Sandy dispels this perception:
It’s not about drawing walks, it’s about getting into hitter’s counts.
The “On Base Percentage” approach hinges on a single, simple pretense: don’t get out.
That’s it. Don’t get out. Drawing a walk is better than getting out. But that doesn’t make a walk better than a single, and of course Sandy Alderson wouldn’t prefer walks to hits. But players who walk more often are typically more disciplined at the plate, and as such tend to be better hitters overall. Ideal offensive players, as such, are those who get on base and hit for power. Think Mike Trout, Andrew McCutchen, and Miguel Cabrera.
Because of the fact that OBP is more valuable than Slugging, using OPS as a measure of offensive value has an inherent weakness. A better measure, arguably, would be weighted on base average (wOBA). wOBA measures a player’s offensive value by weighting all non-intentional (“intentional” in this case being intentional walks and sacrifice bunts) outcomes by their run expectancy, and then averaging that total. From the Fangraphs page, the important thing to note is that walks are the least-valuable positive outcome, because a walk can advance runners a maximum of one base. A runner on second can score on a single, but not on a walk or hit by pitch. Also, bases aren’t necessarily linearly-weighted – a double isn’t twice as valuable as a single, and a home run isn’t twice as valuable as a double – because you aren’t necessarily twice as likely to score a runner on base on a double as a single.
Don’t mistake that though; a walk is almost infinitely preferable to an out. Your team isn’t worse off, and it allows the next runner the same opportunity. And that’s how rallies are started. By not getting out.
Home runs are awesome. But they’re that much more awesome with a couple of ducks on the pond. And that’s what it’s all about – having more people reach base, so guys like David Wright and <insert your preferred first baseman> have people to knock in.