The turning of the calendar, with its artificial demarcation of time and progress, makes for an excellent opportunity to pause and reflect. After all, we've all grown weary of adding "small sample size" to the ends of all our sentences. April has felt like a particularly long month for fans and general managers alike, so it seems like we should know something about our baseball team by now. What do we know about our baseball team now?
Not very much.
Statistics stabilize at different rates. That is to say: statistics never truly stabilize. Numbers provide correlation but never causation; we can track how many times the sun rises at the beginning of the day, but we can never use that data to prove that it'll rise tomorrow. Instead, increased data provides us with an increased confidence in whether the results we've seen are representative of that player's true talent. This research was pioneered by Russell Carleton of Baseball Prospectus, and has been adjusted over the years by himself and others.
Charlie Adams at Beyond the Box Score wrote on this topic a week ago, and the result was this: by the end of April, only two offensive statistics can be considered to be "reliable": the rate at which a batter swings, and the rate at which he makes contact. (Carleton included a third rate, pitchers per plate appearance, which I'll add to the chart.) Even the word "reliable" requires hesitation on the part of the author, since all it means is that the R-squared of the data is above 50%, which means that we are now slightly more confident that the data is meaningful than that it is randomness. It's not the point where we can assume it as fact, only the point where we can begin to trust it. The more plate appearances, the more data, the more trust.
We're not done, though. On top of that, we also have to accept the fact that these "stabilized" statistics don't really tell us that much about the future. We don't know what players will do based on these numbers; we can only be reasonably be sure that what they did do was really them. But whether this is part of a new ongoing trend or a correction in the previous data, we can't be certain. Analysis like this makes me want to go back to writing short stories.
Four hundred words of equivocation! Let's stop. Here are the numbers for you to absorb, for all Mariners with more than 70 plate appearances (note: these numbers are through 4/29 because I am lazy and did not update them last night):
2013 | Career | 2013 | Career | 2013 | Career | ||
Name | PA | Swing% | Swing% | Contact% | Contact% | P/PA | P/PA |
Kyle Seager | 112 | 38.4% | 46.3% | 80.0% | 82.7% | 4.24 | 3.70 |
Kendrys Morales | 110 | 45.7% | 47.9% | 80.2% | 78.6% | 3.67 | 3.87 |
Justin Smoak | 107 | 42.3% | 43.7% | 73.5% | 77.3% | 3.91 | 3.87 |
Michael Morse* | 100 | 53.5% | 51.2% | 69.4% | 74.8% | 3.87 | 3.71 |
Dustin Ackley | 93 | 40.7% | 39.5% | 87.7% | 85.7% | 4.12 | 4.11 |
Brendan Ryan | 75 | 42.9% | 45.6% | 81.1% | 81.7% | 3.72 | 3.77 |
Let's take them in order of significance.
Kyle Seager: is great. We know this. It was a popular sentiment among the national experts that Kyle's pleasant 2012, in comparison to expectations, represented an early peak rather than an elevated one; analysts like to treat their mistakes as aberrations. April has thrown even more evidence toward the latter conclusion.
As Logan noted a couple of days ago, Kyle's increased selectivity has led to an increase in his walk rate without an accompanying spike in strikeouts (numbers that we can start talking about in another 50-100 PAs). This is because he's swinging at fewer pitches both inside and outside the zone, and seeing half an extra half of a pitch with each at bat. He's also not just watching a bunch of first-pitch strikes, either: he's seen pitcher's counts in 25.8% of his plate appearances this season, as opposed to 28.8% over his career.
It's no certainty, but so far, Kyle Seager looks like he's made some real, positive adjustments to his game.
Kendrys Morales: is the same. What's most interesting about Morales' rates is that he's swinging less than ever before, but he's swinging the same amount on pitches in the strike zone. No, it's the garbage that he's laid off of, posting a dramatically lower swing rate (26.2%) than his career norms (33.2%). Everything else that's changed so far (the BABIP, the disturbing drop in ISO, the softening platoon splits) will have to wait.
Justin Smoak: is awful. Why talk about him? Let's not talk about him.
Michael Morse: broke a finger. We may as well treat Morse as having two 50 PA seasons so far, one by Bad Mike Morse, the other by Bad as People Used It in the Eighties Mike Morse. We'll throw out both of them for the time being.
Dustin Ackley: lost his biceps. Perhaps it's most surprising that despite all of his mechanical tweaks, Ackley's swing and contact rates have remained extremely faithful to his (short) career patterns. With Dustin, even more than Kendrys, it's a matter of figuring out why his power has vaporized, and why he's gone from a 39.8% ground ball rate in 2011 to a painful 56.2% this year.
Brendan Ryan: has been the Anti-Seager. Like Kyle, he's swinging less this year, but that's the end of the happy coincidences. Instead, Ryan's swinging at more pitches outside the zone than he ever has before (31.0%, compared to a career rate of 28.2%) and far less actual strikes (54.6%, compared to a career rate of 63.7%). Maybe Brendan has lost faith in his ability to put the ball in play, or maybe he's bunting more and doing it poorly (although B-R's bunting data is glitched right now, so I can't tell for sure). Maybe he's still really sad that LucasArts got dismantled.
Uncertainty isn't fun. It certainly doesn't make for popular journalism or analysis. But if we want to give up the phrase "small sample size", we have to replace it with "we can with X% confidence say", which is more syllables and is unalliterative and terrible. So for now, the only thing we can truly say with confidence is that this was an unfulfilling conclusion.