After being drafted in the first round of the 2014 draft out of LSU, Aaron Nola quickly became a top prospect in the Phillies’ farm system. He made his major league debut in 2015 and has been a key part of the rotation since. In 2018, Nola had a breakout season. In 212.1 innings, he struck out 224 batters and had a 0.98 WHIP, which led to a 17-6 record and 2.37 ERA. With this performance, Nola finished third in the NL Cy Young voting, behind only Jacob deGrom and Max Scherzer. Expectations were very high going into the 2019 season, but they were unfortunately not met. Nola did increase his K/9 from 9.5 to 10.2, but his WHIP increased to 1.27 and he went 12-7 with a 3.87 ERA. These numbers are still good, but Nola is clearly capable of performing better. Accusing juiced balls for Nola’s decreased effectiveness would be easy, but it would be more helpful to look at his data in more depth to find more specific reasons.
Comparing 2018 and 2019
Nola’s four pitch types are a four-seam fastball, a two-seam fastball that could also be called a sinker, a changeup, and a knuckle curve. Usage, average velocity, batting average on balls in play (BABIP), and CSW% will be used to draw conclusions on if Nola may have had worse luck in 2019.
These are the values for the 2018 season. One thing that stands out is the low BABIP. Nola’s total BABIP on all pitches was .251, which suggests that he may have benefited from luck. Looking at his 2019 numbers will help analyze his performance a little more.
As expected, Nola did not keep up the low BABIP in 2019 on every pitch other than his changeup. Despite this, he did increase the average velocity on each of his pitches by about one mile per hour. It is also now more helpful to look at CSW% with two seasons to compare. He had roughly the same percentage of called/swinging strikes on the two seam fastball and changeup. The 3.3% increase in CSW% on the four seam fastball is good, but the 6.8% decrease in CSW% on the knuckle curve is concerning. The higher usage rate and lower percentage of called/swinging strikes shows that this pitch was not as effective at putting hitters away as it was in 2018. All of these statistics show that Nola had better results in 2018, but there is more that can be done to analyze his 2019 results.
2019 Predictive Modeling
To help analyze where things did not work out for Nola in 2019, some predictive modeling will be used to see how his 2019 numbers compare to what he would be expected to do based on his metrics. These expected stats will be calculated using the random forest algorithm.
Nola’s 2018 data is the training set for the model, while his 2019 data will be tested. The increase in BABIP seen in the tables above seems to indicate that Nola had worse luck in 2019. The expectation is that Nola’s expected stats from the model will be better than what actually happened. Models will be made to predict swings and hard hit balls, with the difference between his 2018 stats and both his real and expected stats being analyzed. A negative number would mean that his 2018 stats were greater than either the real or expected 2019 stats.
According to the model, Nola should have had a smaller decrease in swing percentage on the four seam fastball, a significant decrease in swing percentage on the two seam fastball, a slightly smaller decrease in swing percentage on the changeup, and a larger increase in swing percentage on the knuckle curve. More swings on the knuckle curve and fewer swings on the two seam fastball stand out the most here. Swinging at the 2-Seam fastball more than expected would likely lead to more balls put in play that would increase BABIP, while swinging at the knuckle curve less than expected would likely lead to fewer balls put in play to increase BABIP.
Hard Hit Balls
The model for hard hit balls also produced some interesting results. One observation is that the model was very accurate for fastballs. The changeup and knuckle curve gave more interesting results. According to this model, the increase in Nola’s hard hit rate on changeups should have been almost three times higher. It should be noted that the changeup is the only pitch where Nola did not see a major increase in BABIP in 2019. The knuckle curve’s results are the most interesting here. In 2019, his hard hit rate with this pitch was about 14% higher, a major increase. However, the model predicts a 12% decrease. This definitely falls in line with the idea that Nola had much better luck in 2018,
Aaron Nola is an excellent pitcher for the Phillies and will be a major part of their rotation as they try to make the postseason. To make a deep run in the playoffs, the Phillies will benefit greatly from getting the 2018 version of Nola back . The 2019 version of Nola saw some worse luck with more hits to go along with the increase in walks. It would be easy to accuse the alleged juiced balls of being responsible for the decrease in Nola’s effectiveness, but home runs are not used in the calculation of BABIP. The slight increase in velocity Nola had in 2019 could be combined with better luck on hard hit balls and fewer walks to make him an elite pitcher once again.
|Jordan Jones on New Yakkertech Tagging Interfa…|
|Crafting a Gameplan… on The Jordan Hicks Dilemma: A Qu…|
|A Quantitative Evalu… on Avoid the Dead Zone: An Extens…|
|SB on Zach Davies: What Changed?|
|Can Detroit Tigers B… on Improving Pythagorean Winning…|