Matthew's Portfolio

Leading the way for the Mets in terms of sheer pitch volume was rookie Kodai Senga. Towards the left of the graph, primarily starting pitchers can be found, as they tend to go deeper into games compared to relief pitchers. On the right side of the graph, position players such as Luis Guillorme and Danny Mendick can be found, with pitch counts that do not even register on the graph.

The most frequently used fastballs among pitchers were Four Seamers. Among breaking balls, Sliders were the most commonly employed. Off-speed pitches, specifically changeups, were utilized the most by pitchers during games.

Scatter plot shows avg. launch angle vs. avg. launch speed for each pitcher. Two outliers circled: Luis Guillorme & Danny Mendick. Data clusters mostly between 10-25 degrees and 85-90 MPH. Trend suggests higher angles correlate with higher speeds; more pitcher data needed for confirmation. Lower launch speeds create weak contact, aiding defense in run prevention. Players with lower angles tend to have lower speeds; investing in them could be advantageous.

Among all starting pitchers, identified by their notably taller bars on the graph, Kodai Senga stands out as the sole pitcher who utilized Forkballs at a noteworthy frequency. Additionally, he led in total pitch count for the year. Senga employed a diverse repertoire, showcasing seven distinct pitch types in 2023. This multifaceted approach likely contributes to his effective management on the mound. The data suggests that encouraging other starters to increase their usage of secondary, tertiary, and potentially quaternary pitches could yield similar benefits. Such a strategy could also lighten the workload for relievers, providing a tactical advantage across the pitching staff.

Balls put into play are categorized into fly balls, ground balls, line drives, or popups based on their launch angles: ground balls (GB) are less than 10 degrees, line drives (LD) range from 10 to 25 degrees, fly balls (FB) span 25 to 50 degrees, and popups (PO) exceed 50 degrees. Most pitchers predominantly see ground balls as the majority of their balls in play, while popups typically constitute the smallest portion. Fly balls and line drives exhibit considerable variation among pitchers. Lower launch angles correlate with reduced exit velocities, thereby producing more ground balls. Pitchers with high ground ball rates tend to minimize hard-hit balls. However, effective team defense remains crucial for maximizing the benefits of a ground ball pitching strategy.

Among the 37 pitchers analyzed, 28 were right-handed and 9 were left-handed. Increasing the number of left-handed starters and relievers would enhance roster flexibility for strategic matchups. Many starting pitchers displayed similar pitch counts against left-handed and right-handed hitters, exemplified by Tylor Megill's nearly even distribution of 1085 pitches to righties and 1150 to lefties. In contrast, relief pitchers often exhibited specialization, evident in visual disparities on their respective bar graphs. For instance, Adam Ottavino, a right-handed pitcher, noticeably favored pitching against right-handed hitters, leveraging his inherent advantage in those matchups.

After performing PCA analysis, the first 2 components, which together explained more than 99.5% of the variance, were chosen.

Based on an SSE elbow plot, it was determined that using 5 clusters provided the most descriptive segmentation.

There is a potential issue where insufficient or less significant variables might have been chosen instead of more impactful ones.

Of the 37 pitchers analyzed: 32.4% were in cluster 0, 24.3% were in cluster 1, 5.4% were in cluster 2, 27% were in cluster 3, and 10.8% were in cluster 4

There was more variation in the cluster’s average spin rates (2039-2638) than the cluster’s average release speed (87.8-89.7) (while ignoring the cluster of position players).

This indicates that release speed may influence the distinction of pitchers less than spin rate, i.e., the team should more intensely monitor pitchers’ spin rates, not necessarily their release speeds when evaluating.

Out of the 863 pitchers analyzed, the clustering results revealed varying distribution across six distinct clusters. The largest proportion was found in cluster 0, comprising 34.6% of the pitchers, indicating a dominant group within the dataset. Cluster 1 represented 14.3% of the pitchers, followed by cluster 3 at 14.4%. Clusters 2 and 4 each accounted for 5.1% and 2.1% respectively, reflecting smaller subsets within the population. Cluster 5 encompassed the highest proportion after cluster 0, consisting of 29% of the total pitchers analyzed.

Clusters 0 and 3 exhibited the lowest average launch angles at 11.5 and 11.9 degrees respectively. Cluster 5 also showed relatively consistent launch statistics among its players. In contrast, clusters 2 and 4 displayed greater variability in their launch angles. Given that lower launch angles tend to result in groundouts, which are advantageous for pitchers, prioritizing free-agent signees from clusters 0 and 3 is recommended. It is also prudent to avoid pitchers grouped with position players, as they may not perform as effectively in pitching roles. This strategy aims to optimize team performance by leveraging pitchers with consistent and effective pitching metrics.

Clusters 4 and 2 are distinctive due to their notably lower release speeds and spin rates, similar to the NYM model, primarily comprising position players rather than pitchers.

Spin rates vary more significantly across clusters than release speeds.

Notably, cluster 1 exhibits the highest spin rates and shows the least change in run expectancy, suggesting more consistent performance.

Thus, prioritizing the evaluation of pitchers based on spin rates could offer valuable insights into their effectiveness and reliability on the field, aiding in strategic pitching decisions.