By Marc Normandin
There isn't much season to analyze at this point in April, but, since we've all been so desperately waiting for Baseball That Counts, we all try to make what we've seen count more than it does. Sure, it's cool that you can't throw a baseball near Chris Davis without him sending it into orbit right now, but if it happened in June, it wouldn't register nearly as much -- that would be a hot stretch in the midst of a season that had already established itself, whereas, right now, that hot stretch is his season.
When is it OK to start to believe what the numbers are telling us, though? There has to be a point where we can trust what we're seeing, where results are more than just what happened and are representative of what will happen, as well. Thanks to research by the sabermetrically inclined, we can pinpoint these moments where numbers actually matter.
Russell Carleton, before he worked for Baseball Prospectus, before he worked for the Cleveland Indians (and before he worked for BP once more), wrote for a blog on a now-dead sports network called MVN. That site, Statistically Speaking, produced quite a few stathead writers during its short life. One of the items that Carleton produced, and it's one that every baseball fan should remember this time of year, was his guide to measuring the reliability of statistics. Carleton's work, in essence, described the moment you could start to trust a player's statistics, accepting change -- for good or bad -- as legitimate.
The thing is, different stats require different playing time thresholds. For instance, you can tell a hitter is going to swing often or not often quickly, with swing percentage stabilizing in 50 plate appearances. Even this early in the season, there are a handful of players close enough that their swing percentage will have stabilized before the week is out. However, if you want to know whether a player has truly reached a new level of plate discipline, well, you're going to have to be patient, too: It takes around 200 plate appearances for walk rates to stabilize.
So, if you're a Nationals fan who is hoping that the team's hitting coaches worked with trade acquisition Denard Span over the offseason, turning him into a walk machine, you're going to have to wait another 169 plate appearances, or another 40 games or so, to find out if that's true. You can probably guess, though, given his career walk rate of less than 10 percent, that he's not going to continue to draw a free pass one-fourth of the time to the plate.
Let's be honest, though. No one is noticing walk rates right now, not this early. What people are paying attention to are the power numbers. Home runs, lofty early-season slugging percentages -- the numbers that really pop off the mid-April stat sheet. Justin Upton already has six homers, or one every 5.5 plate appearances, while new Mariner Michael Morse is hitting one out every 6.7. Then there is Coco Crisp, who has four homers already after averaging just 12 per 162 games for his career entering the season. Is there anything we can draw from these numbers yet? In short: nope.
If you thought you had to wait awhile to know if walk rates were legitimate, know that it takes 300 plate appearances for home run rates (and home runs per fly ball) to stabilize, 500 plate appearances before slugging percentage can be trusted, and 550 times up before you can trust Isolated Power (slugging percentage minus batting average, used to gauge just how powerful a player's hits are). This is why it's difficult to get excited about an early season power outburst from someone like Crisp, or even the aforementioned Davis, who has, on occasion, murdered a baseball or 30 in a given year. These things happen, and they happen in other times of the year besides April, but it's much harder to notice the two-week spike in power in someone who has spent three months slugging .475 or .400 or whatever then it is in someone whose OPS can jump hundreds of points with a productive April afternoon.
It's not just hitters who need caveats aplenty for their performances, as pitchers have their own rules to follow. Ground ball rates and strikeouts per plate appearance take about 150 plate appearances before you can trust any change in them -- the league leader in batters faced as of this writing is at 60, so they still have a handful of starts to go before we can even begin to pay attention. Fly-ball rate takes a little longer, at 200 batters faced, but strikeout-to-walk ratio and just straight-up walk rates take 500 and 550 batters faced, respectively. That's about half-a-season for someone like Justin Verlander, and even more for pitchers that throw far fewer innings.
Yes, Clayton Kershaw is pretty awesome, but he's about 500 batters faced short of us thinking he can strike out 16 times as many batters as he'll walk in 2013. Matt Harvey is an exciting young hurler and potential ace in the Mets' rotation, but he's just a little bit shy of us believing he's going to punch out over 12 batters per nine innings on the year. Given time, these things will sort themselves out, and we'll see who has made progress, fallen behind, or stayed exactly the same. Just not now, nor anytime particularly soon.
That's not to say we can't identify anything that has changed with pitchers or hitters. But, this time of year, scouting rules the day. Maybe Denard Span has changed his approach in a way that will allow him more walks. Maybe Coco Crisp isn't going to lead the A's in homers, but he's bulked up a bit and barrels up on the ball better, and will set a career-high for the long ball because of it. Maybe the league is so scared of Joey Votto that he actually will walk about 30 percent of the time. Well, no, he won't, but we can dream.
If you want to find legitimate changes before statistics will allow, scouting is going to be the way to do it. That won't tell you everything, either, as the stats have to bear out the story that scouting is telling, but it can at least help point you in the right direction, or let you know where to look. And hey, in a few months' time, you'll know for sure if Chris Davis is amazing, pretty good, or something in between. You just have to be patient in the meantime.
* * *
FanGraphs' sabermetric library has the full breakdown of these meaningful sample size thresholds, as well as archived links to Carleton's original studies. Carleton only studied statistics that could be measured over the course of a single season, so you won't find something like batting average on balls in play or ERA, as those take years and years to be trusted. Remember that, too, the next time you see someone in April hitting .400, or succeeding with runners in scoring position more than usual.
* * *
Marc Normandin writes and edits for Over the Monster, a Boston Red Sox blog, as well as SB Nation's baseball hub. He's one of many behind the e-book " The Hall of Nearly Great," and has written for Baseball Prospectus, ESPN, and others. You can follow him on Twitter at @ Marc_Normandin.