FanPost

Predicting the Future — How Can We Know How Players Will Adjust to the Next Level?

The offseason is a long and arduous journey of self-discovery, something of a yearly walkabout during which each of us must ask ourselves: what can we do with an extra 3 hours each day? For some, the answer is surely that if you cannot be watching baseball, at least you can use that extra time to study baseball.

As sabermetricians, our job is something of a diviner's art, the modern version of the oionopolos of Mycenean Greece, who would map the trajectories of birds across the skies in attempt to find correlations between wildly disparate phenomena. What patterns can we find that will reveal outcomes of the future?

I'll be looking at a question that has always interested me greatly: what should we be looking for in order to better be able to predict how a young player will adjust to the next level? Some prospects are sensational throughout every minor league stop, but then are unable to translate such success to the major leagues (Matt Wieters, for instance, even though he's been pretty good in the MLB). And then some players don't seem to lose a beat as they make the jump across levels (Todd Frazier, for instance is pretty much the exact same player in the MLB as he was in AAA).

Two rates in particular have drawn my interest in relationship to the jump: BABIP and BB/K. For so long the sabermetric community has placed an extreme amount of importance on BB/K when evaluating minor league players. BB/K provides a bit of a general picture of more refined statistics that we have no access to, like Contact % and O-Swing %. BB/K has come to represent almost something of a "maturity" factor, in that minor league players who walk as much as they strikeout seem to be the ones who best understand hitting.

Conversely, BABIP for years was viewed as something to be mostly regressed, a statistic which exemplifies a high degree of luck, and thus the best prospects can be found by weeding out those whose lines are largely fueled by extremely high BABIP or finding players who would have had good batting lines if their BABIP weren't so low. In the last couple years, however, the sabermetric understanding of BABIP has improved greatly. It is no longer considered simply luck (for hitters, by the way, this is all for hitters), although of course luck does remain a significant aspect within the statistic. Hitters have control over BABIP in how hard they hit the ball, and what in what trajectory they hit the ball. Line drives have close to a 70% chance of being hits, whereas infield fly balls are closer to 0. But defense also plays a huge role in BABIP, as a good defense will suppress it and a bad defense will allow balls in play to fall as hits all over the place. In the minor league, the company line is that defenses are worse and thus high minor league BABIPs should be taken with a grain of salt. This is undoubtedly true, but how bad exactly are minor league defenses, and how much of a discount should we give players?

And so I wanted to see how much correlation there is in translation of offensive performance (for which I'm using wRC+ because it's park and league adjusted, which is extremely important when crossing minor league levels) to the next level and BABIP, BB/K, BB%, and as something of a control wRC+ itself.

Before we look at some data, something important to think about is that as you'll see is that because we're crossing levels, we're inherently looking at statistics that should correlate LESS to each other than they would in the same level. We wouldn't necessarily expect a player's numbers to remain the same across all of the minors and into the majors, but that doesn't mean there won't be any correlation. It'll just be smaller.

I used the data of every qualified rookie in the MLB of the past two years across Low A, High A, AA, AAA, and the Majors. I used only two years of rookies because the study is about adjusting to the next level, and most players don't spend more than two years in any one stop in the minor leagues (especially if they make the majors), so I didn't want the MLB data to outlast a similar adjustment period. The idea is that, yes, we're working in small samples, but that's all that we have for the minor leagues, so the question is how much correlation can we draw from the imperfect data.

The other inherent flaw here is that by using rookies for the study, these are players who will have done better in the minor leagues than the average minor league player, because they've made the Majors. But this was an unavoidable problem because I wanted to study the AAA to MLB jump.

Without further ado, here are some charts, and then I'll talk about them.

20130124193154_medium

20130124194334_medium

20130124193300_medium

Here's what we have:

GRAPH 1: Average Rate Change Across Levels

This graph is showing how these rate stats changed for the players in the study as they jumped levels. The first thing you notice about this is that players did not get worse jumping levels in the minor leagues, as I would have expected them to. Instead, on average, players either stayed about the same or got a little bit better in all four rate categories as they went up a level, until they reached the Major leagues, at which point there was a major drop off in production (31% in BB/K, 17% in BB%, 8% in BABIP, and 17% overall in wRC+).

The next noticeable thing here is that BABIP does not drop off nearly as much going from AAA to MLB as BB/K.

The last thing is that the jump from AA to AAA sees a fairly large increase in BB%, and hence BB/K. This is a trend that is a elucidated a bit by...

GRAPH 2: How Steady do Rates Remain When Jumping Levels?

This graph runs a linear regression on the rates as they jump levels, coming up with a correlation for each stat. The closer the correlation number is to 1 (or -1), the greater the correlation, and the closer to 0, the weaker the correlation. For something that is completely random, we would expect the correlation to be 0.00. Here are the observations:

Firstly, BB% has a strong correlation across levels, meaning that players generally maintain their BB% fairly consistently as they cross levels, even moreso than their overall offensive production, as seen by wRC+. This makes sense, as BB% has shown a very strong correlation from year to year in the same level (MLB). A recent study by Matt Klaassen on Fangraphs demonstrated a 0.765 correlation between yearly BB% within the MLB, so the 0.66 correlation from AAA to MLB makes a lot of sense.

The next interesting thing is that there is no observed correlation whatsoever between BABIP in High A and in AA, but that there is a correlation (albeit a small one) from Low A to High A, from AA to AAA, and from AAA to MLB. There are a few possible ways to read this, but one take is that there are two different difficult jumps for minor league hitters — one of these jumps is from AA to AAA, where the correlation of BB% drops. The other is from High A to AA where the BABIP correlation drops. My take is that it's possible that what this signifies is that minor league defenses get a lot better in AA (or yes, perhaps pitchers are better at pitching to weak contact), and essentially separate the wheat from the chaff in terms of hitters' BABIP abilities. And then in AAA, hitters in general have learned to take pitches out of the strike zone better, and so this jump creates more variation because of the pitching side. Overall BB% increases, but correlation decreases, as once again, the wheat is separated from the chaff. Pitchers who can't throw strikes or get hitters to chase start to lose effectiveness in AAA (Danny Hultzen, anyone?), and most hitters benefit, though I would imagine that the lower correlation also signifies that the hitters who have serious trouble chasing pitches out of the zone face their biggest challenge in this jump (Derek Norris, anyone? remember his incredible BB%s in the lower levels?).

Graph 3: Predicting the Future — Correlation Between wRC+ and Previous Level's Rates

This is the graph that faces the big question head-on, how much correlation is there between these other rate stats and the transition to the next level in terms of offensive production?

First thing's first — there is no correlation from Low A to High A for either BB/K or BABIP. There is, however, a small correlation for wRC+, but where is that predictive power coming from here, if not from BB/K and BABIP? I think it has to be coming from ISO, which is the other part of the equation that I've left out of this study, partly because it's just a whole 'nother bag of worms for another day, and partly because my intuition tells me that Isolated Power is something which shouldn't really be affected by crossing leagues... theoretically, how far someone can hit it really only has to do with them, it's how often they do it which would change across levels. Point is, I'd like to do a similar study with ISO. But back to this, if neither BABIP nor BB/K are amounting for the predictive ability of wRC+, from Low A to High A, what's left are power and speed. And the fact that minor league offensive statistics are not as important as we might like to believe in A ball.

Next, when jumping from High A to AA, BB/K is much more important in predicting the future than BABIP. This feels right to our sabermetric understanding, as the jump from High A to AA is often considered the hardest minor league jump, and again BB/K has usually been used as the most predictive part of offensive performance in the minor leagues. However, the jump to AAA is very different. The better predictor of wRC+ from AA to AAA is BABIP, by a lot. What this tells us is that the hardest jump in terms of BB/K is indeed from AA to AAA, but that BABIP does start to become predictive in AA, meaning that if a hitter can get to AA with a high BABIP, there's a decent shot they'll be able to sustain it.

These two different jumps seems to me the most important find here, and allows for questions which may help us in prospect evaluation. Can a hitter sustain a high BABIP in AA? Can a hitter sustain a good BB/K rate in AAA? If the answer is yes, these are hitters who might be poised to adjust well to the next level.

So for whom does this bode well for? Someone like Michael Choice has shown that he can sustain a high BABIP in AA, but still needs to pass that BB/K test in AAA. Josh Donaldson's highly improved BB/K in AAA in his third stint there perhaps indicated that he had made that jump. Someone like Jesus Montero went from a .381 to a .320 BABIP between High A and AA and never regained his extremely high BABIPs from the low minors. Last year his BABIP was .296, which is alright, but not nearly enough to make up for a bad BB/K and only average power. As someone who was billed as a prospect would "mash" (which is usually code for someone who doesn't walk that much but hits the ball hard), his drop in BABIP to AA should have been a red flag in terms of his adjustment abilities. Not that he won't be able to eventually make that adjustment, as this study is about short term transitions, but it's still something to be aware of.

In 2012, the A's had a lot of hitters come up from AAA and immediately translate their offensive production to the big leagues, something that hadn't happened in a long time. Going forward, what do we see? I think this study suggests that Grant Green might enjoy a good transition to the majors — Green maintained a high BABIP in AA, and then improved his BB/K in AAA, both of which presage good things for his jump.

Predicting the future for minor leaguers is difficult and inexact. But there are things to look for — and I'd imagine they've got a stronger correlation with future performance than the patterns of birds across the sky. But who knows.

X
Log In Sign Up

forgot?
Log In Sign Up

Please choose a new SB Nation username and password

As part of the new SB Nation launch, prior users will need to choose a permanent username, along with a new password.

Your username will be used to login to SB Nation going forward.

I already have a Vox Media account!

Verify Vox Media account

Please login to your Vox Media account. This account will be linked to your previously existing Eater account.

Please choose a new SB Nation username and password

As part of the new SB Nation launch, prior MT authors will need to choose a new username and password.

Your username will be used to login to SB Nation going forward.

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

Join Athletics Nation

You must be a member of Athletics Nation to participate.

We have our own Community Guidelines at Athletics Nation. You should read them.

Join Athletics Nation

You must be a member of Athletics Nation to participate.

We have our own Community Guidelines at Athletics Nation. You should read them.

Spinner.vc97ec6e

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9351_tracker