Professor Lugo

In his first start as a Kansas City Royal, Seth Lugo delivered six scoreless innings. In his second outing, he pitched 6.2 innings and allowed just one run.

Seth Lugo has excelled in his debut season with the Royals, solidifying the rotation and, alongside Cole Ragans, providing the team with a pair of legitimate aces. His success stems from a diverse pitch arsenal that includes a four-seamer, sinker, cutter, slider, sweeper, changeup, curve, and slurve. This varied mix keeps hitters off balance and has led to impressive results.

In his debut season as a Kansas City Royals starter, Seth Lugo has unequivocally justified his shift back to the rotation two years ago. This culminated in his selection for his maiden MLB All-Star Game appearance announced last Sunday.

The decision to include Lugo on the American League roster for the Midsummer Classic might be one of the simplest ones this year. Despite conceding two earned runs over 6.0 innings in a 3-1 loss to the Colorado Rockies on Saturday, Lugo maintained a stellar ERA of 2.21, the lowest among all MLB starting pitchers. Saturday's start marked his 15th quality outing in 19 appearances this season, leading all MLB pitchers in this category as well.

During spring training in February, then-Padres manager Bob Melvin was asked what stood out the most about pitcher Seth Lugo. "His curveball," Melvin replied succinctly, according to MLB.com .

In our analysis, we employed linear regression to explore the factors influencing Seth Lugo's pitch release speed by examining a range of pitch characteristics and release metrics from 2024. By applying this statistical method to our training data and assessing its performance on test data, we gained valuable insights into how elements such as spin rate, pitch type, pitch movement, and release position contribute to the velocity of Lugo's pitches. When examining the model, we analyze its effectiveness.

Performance Metrics

• Mean Squared Error (MSE): 1.63
• R^2 (Coefficient of Determination): 0.95

These metrics indicate strong model performance, with MSE signifying that predictions deviate by about 1.63 units from actual release speeds on average. The R^2 score of 0.95 shows that 95% of the variance in Lugo's release speed is explained by the model, indicating a robust fit.

Coefficient Analysis

Release Spin Rate (1.401770): A significant positive influence on pitch speed, highlighting the importance of higher spin rates for movement and deception.

Pitch Types:

• Four-Seam Fastball (FF) (1.880287): Indicates a substantial positive impact on release speed, aligning with the expectation of faster delivery for fastballs.
• Sinker (SI) (1.410912): Shows a significant positive impact on release speed, consistent with sinkers being typically faster pitches.
• Cutter (FC) (0.942424): Also positively influences release speed, though less than four-seam fastballs and sinkers.
• Changeup (CH) (0.058388): Has a minor positive impact on release speed, which could reflect specific usage contexts.
• Slider (SL) (-0.022913): Shows a slight negative impact on release speed, which is expected for slower breaking pitches.
• Splitter (ST) (-0.776234): Indicates a notable negative impact on release speed, consistent with it being a slower pitch type.
• Curveball (CU) (-2.313247): Demonstrates a substantial negative impact on release speed, reflecting its slower nature despite its effectiveness in inducing movement and deception.

PFX (Pitch Movement):

• PFX_Z (0.430417): Vertical movement positively influences release speed, indicating that pitches with more downward break tend to be thrown faster.
• PFX_X (-1.277114): Horizontal movement negatively impacts release speed, suggesting that pitches with significant horizontal break are generally slower.

Release Position and Extension:

• Release Extension (0.180089): Positive impact on release speed, indicating that releasing the ball closer to the plate increases speed.
• Release Position X (0.108258): The horizontal release position has a minor positive impact on release speed.
• Release Position Z (-0.084767): The vertical release position has a slight negative impact on release speed.

Curveball Insights:

Seth Lugo's curveball, with its markedly negative coefficient (-2.313247), is the slowest pitch in his arsenal. Its high spin rate (1.401770) creates significant drop and lateral movement, making it difficult for batters to track. This stark contrast to his faster pitches enhances its effectiveness, making it a potent weapon for inducing swings and misses or weak contact. The strategic use of varied pitch speeds and movements is crucial in Lugo's pitching approach.

Implications:

• Spin Rate: Lugo's curveball spin rate (1.401770) enhances its drop and movement, contributing to its effectiveness in deceiving hitters.
• Pitch Arsenal Diversity: Despite negative speed coefficients for some pitches, Lugo’s diverse pitch types keep hitters off balance, enhancing his overall effectiveness.
• Performance Factors: Metrics show that while Lugo’s fastballs contribute to high velocity, his off-speed pitches like the curveball rely on movement and spin for deception. This strategic variety is fundamental to his success, highlighting the importance of nuanced pitching strategies in MLB.

Lugo’s Metrics

Seth Lugo's success is highlighted by his unique pitching repertoire. His elite spin rate, particularly on his curveball, contributes to sharper movement, making it harder for hitters to connect. Despite being one of the slowest pitches, its high spin rate and movement make it formidable. His diverse pitch types are crucial for keeping hitters off balance.

Unlike traditional pitching norms focused on speed, Lugo relies on spin and movement, particularly in pitches like sliders and splitters. This strategic adaptation has proven successful, showcasing his ability to deceive batters and his value as a versatile pitcher in MLB.

What’s He Trying to Pitch?

Understanding the likely sequence of Seth Lugo's next ten pitches involves delving into the intricate patterns and decision-making processes that define his pitching strategy.

By analyzing key metrics such as release speed, spin rate, and contextual variables like the batter he faces and the inning, we aim to predict with accuracy the types of pitches he is most likely to employ. Leveraging a Random Forest classifier, which excels in capturing complex relationships within data, we seek to uncover insights into Lugo's pitch selection tendencies.

Random Forest Classifier

A Random Forest combines multiple decision trees to improve predictive performance and robustness.

Random Forest Equation:

• y(x) is the predicted class for input
• N is the number of decision trees.
• Ti(x) is the prediction from the i-th decision tree.

Majority Voting:

• For classification, y(x) is the class predicted by the majority of trees.
• For regression, y(x) is the average of the predictions from all trees.

Training Process of Random Forest

Bootstrap Sampling:

• Random subsets of the training data are sampled with replacement to train each tree Ti.

Random Feature Selection:

• At each node of Ti, a random subset of features is considered for splitting.

These processes help ensure that the model is both robust and generalizable, reducing the risk of overfitting while maintaining high predictive performance.

Implementation Context for Seth Lugo

Here's how the Random Forest Classifier is implemented to predict the pitch types:

1. Feature Selection: Identifying relevant features such as spin rate, pitch type, and movement
2. Pipeline Definition: Defining the sequence of data preprocessing, model training, and prediction.
3. Cross-Validation: Evaluating model performance and tuning parameters using cross-validation techniques.

Random Forest Model Performance and Next Pitch Predictions:

Random Forest Mean Accuracy: 0.7416 This mean accuracy indicates that, on average, the model correctly predicts the pitch type approximately 74% of the time. This level of accuracy is considered quite good for a classification task involving multiple classes, especially in a complex and variable domain like pitch prediction in baseball.

Standard Deviation of Accuracy: 0.0063 The standard deviation of the accuracy score is relatively low, indicating that the model's performance is consistent across different subsets of the data. A low standard deviation suggests that the model is stable and reliable, with minimal fluctuation in its predictive capability.

Next Ten Predicted Pitch Types

Using the Random Forest model, we have predicted the types of Seth Lugo's next ten pitches: ['SI' 'SL' 'FF' 'CU' 'CU' 'CU' 'CH' 'SI' 'CU' 'FF']

Variety and Unpredictability

The predicted sequence of Seth Lugo’s next ten pitches showcases a strategic blend of pitch types: sinkers (SI), sliders (SL), four-seam fastballs (FF), curveballs (CU), and changeups (CH). This variety is a testament to Lugo’s diverse pitching arsenal and his ability to keep batters guessing. In contrast to pitchers who may rely heavily on one or two types of pitches, Lugo’s approach ensures that batters cannot easily predict what is coming next. This unpredictability is crucial in baseball, where the difference between a hit and a strikeout can hinge on the batter’s ability to anticipate the pitch.

Curveballs: A Key Element

The model’s prediction of three curveballs (CU) out of the next ten pitches underscores the effectiveness and reliance on this pitch type in Lugo’s strategy. Lugo’s curveball is known for its exceptional spin rate and significant movement, which makes it particularly difficult for batters to track and connect with. The high spin rate generates a pronounced drop and lateral movement, creating a deceptive trajectory that often leaves batters swinging at air or making weak contact.

Curveballs are inherently slower than fastballs, which adds another layer of complexity for the batter. After seeing a fastball or sinker, a well-thrown curveball can be almost impossible to hit because the batter’s timing is thrown off. This makes Lugo’s curveball a formidable weapon in his pitching arsenal.

Mixing Sinkers and Fastballs

The inclusion of sinkers (SI) and four-seam fastballs (FF) in the predicted sequence highlights Lugo’s strategy of mixing high-velocity pitches with those that have significant movement. The sinker, with its downward and horizontal movement, is typically used to induce ground balls, while the four-seam fastball relies on its straight trajectory and speed to overpower batters. By alternating between these pitches, Lugo keeps the batters’ timing and swing mechanics constantly in flux, making it challenging for them to settle into a rhythm.

Differentiating from Standard MLB Pitching Sequences

Lugo’s approach differs from standard MLB pitching sequences, where many pitchers might rely more heavily on fastballs or another primary pitch. The diversity in Lugo’s pitch selection makes him stand out. While a typical MLB pitcher might use a fastball as their go-to pitch, Lugo’s willingness to frequently employ his curveball and other off-speed pitches adds a layer of complexity to his game. This multi-faceted strategy not only increases his effectiveness but also makes him a more versatile and unpredictable pitcher.

Combatting the Curveball

For batters facing Lugo, particularly his curveball, a few strategies could be employed. One approach is to focus on identifying the release point and spin early. Recognizing the unique spin and trajectory of a curveball can help batters adjust their timing. Another tactic is to stay back in the batter’s box and wait for the ball to break, allowing for a better read on the pitch’s movement. Additionally, batters might benefit from a more compact swing to adjust to the reduced speed and sharp movement of the curveball.

Lugo’s Standout Sequence

The diversity and strategic sequencing of Lugo’s pitches are what make him stand out. His ability to effectively mix high-velocity pitches like the sinker and fastball with the deceptive curveball keeps batters perpetually off balance. This sequence not only highlights Lugo’s skill in pitch selection but also his understanding of how to exploit batter weaknesses. By consistently changing the speed and movement of his pitches, Lugo ensures that batters can never settle into a predictable pattern, thereby enhancing his effectiveness on the mound.

This strategic variety is a key reason why Seth Lugo is a standout pitcher and a strong candidate for the midsummer classic.