Spotting Breakouts: Spring Training Batted Ball Data

With fantasy owner always searching for an edge, I may have found one buried deep in spring training stats. While looking for Yonder Alonso’s MLB.com player page, I noticed he had some batted ball data in the form of GO/AO (groundouts/air outs). In Alonso’s case, his GO/AO value had always been greater than 1.1 until last season when it dropped to 0.87. In the regular season, Yonder’s groundball rate plummeted from 44% to 34%. Yonder admitted to making a swing adjustment and that change should be detectable in spring training. By comparing spring batted ball data, fantasy owners can get an idea of those hitters who may be ready for the flyball revolution.

Note: While flyball and line drive rates are available for comparison, I will only use groundball rate (GB%) because it stabilizes quicker, has less stringer bias (tough call between some line drives and flyballs), and is only one set of benchmarks to memorize.

I’m going to have two foci for this article. Part 1 is all math and disclaimers. It’s the process I used to go from GO/AO values to ground ball rates to launch angles. Part 2 contains the results from Part 1 as a simple procedure for finding launch angle breakouts.

Part 1: Math and Disclaimers

When I saw the GO/AO values I had no idea their equivalent groundball numbers. I had to build my knowledge of GO/AO values from scratch. I am used to just seeing the batted ball rates which include both hits and outs. This conversion was easy to find using the Gameday database.

I took all hitters from 2015 to 2017 (used these seasons because launch angle data is available) with at least 200 PA and compared their GO/AO value to their GB%. I ended up with a logarithmic curve because GO/AO is a ratio and a change in either value will not change the output linearly. For a final equation, I ended up with:

• GB% = .1971* ln(GO/AO)+0.4396
• r-squared: 0.85

While not a perfect correlation, it does provide a needed conversion from GO/AO to GB%.

Now that we have the GB%, it can be converted into launch angle (LA). Using the same sample of data, I found the following equation to move from GB% to LA:

• LA = -50.9*GB%+36.5
• r-squared: .67

Not a perfect correlation but I am just trying to create a simple procedure to find some players to target. The table with all three rates is in Part 2 for the math adverse crowd.

The other issue is a lack of spring training with most regulars maxing out at 70 PA. With the small sample of PA, I compared the 2017 GO/AO numbers from spring training to those in the regular season with a minimum 50 PA. The results were predictably horrible as the should be.

• Regular Season GB% = 0.0558*ln(GO/AO)+.433
• r-squared = .14
• Standard Deviation of predicted versus actual: .068

Again, I’m not looking for perfection from a few spring training plate appearances. I’m just trying to find a small reason to value one similar hitter over another.

Now, the individual GO and AO (and HR) values are available in the spring training leaderboards. I used these values to get a more accurate estimate (standard deviation was 2% points smaller), the corrected value requires an owner to cut-and-paste into Excel and manipulate the values. I wanted the process to stay simple and it will for now.

Time to move onto the results.

Part 2: Utilizing spring training batted ball data

As calculated in Part 1, here is the comparison matrix for GB%, GO/AO, and launch angle (LA):

Using this table, Alonso had a 1.33 GO/AO in 2016 and it dropped to 0.87 last year. Eyeballing the numbers, his GB% declined from around 50% (actually 44%) in 2016 to around 42% (actually 34%) in 2017. The declines were almost identical.

Now, here are the spring training to major league values:

A ton more regression since the estimates are working as a projection.

Examining Alonso, he was projected for around a 45% GB% in 2016 but it just dropped to just 43%. Regression is heavy with this formula. The change doesn’t seem like much but considering it was just 66 spring training plate appearances, it possibly shows a change in approach.

When spring training gets going, owners can track a player’s 2018 GO/AO values on either a player’s home page or on the leaderboard at MLB.com.

Additionally, owners should have our 2017 GB% leaderboard handy. Start at the top and look for hitters who may improve if they put the ball in the air more. Some names which stick out from the top 20 qualified hitters are Eric Hosmer, Christian Yelich, Tommy Pham, Josh Bell, and Robinson Cano.

I think owners should be leery of putting too much stock in any spring training stat. Instead, owners should be looking for is any small edge, especially as the player talent curve starts to level off. Finding that Yonder Alonso or Ryan Zimmerman gem could be huge. Also, picking up several guys who are airing it out will increase the odds that one or two will take a major step forward. This spring I will be testing out how useful is the data, how much stock to put into the numbers, and its future usefulness.