Business

Intelligence and
Analytics in Major
League Baseball
By: Joseph, Sree, Ron, and Jonathan
The Foundation: Moneyball to Big Data
Scale
Historical roots and the immense growth in data collection.

● Early History: The Oakland Athletics used readily available traditional statistics in new ways,
famously documented in Moneyball.

● Modern Data Collection: Today's techniques go far beyond Moneyball. Data on every pitch is
captured by a doppler radar system, sampling ball position 2,000 times a second.

● Scale: Batter swings, exit velocity, and launch angle are recorded. Cameras behind third base
record player position 30 times a second. A massive terabyte of data is captured each game,
now common in professional, college, and even some high school parks.
Utilization of Volume

Performance
Velocity
Key characteristics of data Value
Variety
Veracity

Data in
Major Baseball generates a massive amount of data,
measure at a terabyte per game

League
Baseball: Utilizing a data warehouse

In-game
Decision
Data analysis assists with
key decision-making that Example: Batter & Pitcher
tendencies & situational
should ultimately outcomes

Making
contribute to wins
Utilization of Performance
Data in Major League
Baseball:
Player Selection

• Data Analysis helps teams acquire players

through simulation of the data
• Teams can acquire player through other
teams or sign players whose contracts with
other teams have expired (free agents)
• Predictive Analysis
• Example: Simulation of situational
player performance
Applications: Improved Performance and
Organizational Buy-in
Developing players and critical management steps needed for
analytics adoptions
Improved Performance

● Doppler radar data provides detailed information on pitch delivery (e.g. spin, release, direction,
and path)
● Analysts use this data as a feedback driver to assist in changing a pitchers delivery or motion for
certain pitches.
● For example, Justin Verlander revived his career after being traded to Astros.
● When the Astros hired Jeff Luhnow, there was initial player resistance to changes like new
defensive configurations.
● Breakthrough:
1. The program demonstrated the collection and use of real-time player data.
2. Former players were equipped with software skills as minor league coaches explained the program
to up and coming players.
 Future Trends in MLB Analytics: AI & Biometrics
The Next Evolution: Big Data & AI
● Integrating massive data sets (Big Data) with Artificial Intelligence (AI) to drive analysis.

Shift from BI to Predictive Analytics

● Traditional BI: "What happened?" (Historical data)
● Predictive Analytics: "What will happen?" (Forecasting outcomes)

New Frontier: Biometric Data

● Focus on predicting and preventing player injuries.
● Initial emphasis on pitcher health and performance.

Core AI Components
● Machine Learning (ML)
● Natural Language Processing (NLP)
● Expert Systems (Using rules & heuristics)
Conclusion
Maximizing Value
The Keys to Analytic Success
● Culture: A commitment to data-driven decisions.
● People: The right skills and expertise.
● Leadership: Strong, executive buy-in.

The Goal: From Insight to Action

● Move beyond just describing the past.
● Use predictive tools to shape future outcomes.

Our Journey Today

● The Evolution of Data
● Current Applications
● Organizational Challenges
● Future Trends
Critical Thinking Questions
Question #1: Baseball executives typically call their analysis programs
“analytics.” Based on this chapter’s BI and Analytics definitions, would you say
that their programs are more Business Intelligence or more Analytics? Or,
some of both?

● Business Intelligence (BI) analyzes historical data to determine what happened or what is
happening now. It helps organizations learn from past mistakes and successes.
● Analytics involves the extensive use of data and quantitative analysis to develop predictions
of what will happen in the future, enabling organizations to anticipate developments and
make changes now to improve future outcomes.
● Application to Baseball: Baseball programs use both.
● BI components include descriptive analysis like tracking where a batter tends to hit the ball
(what has happened) to position fielders (what is happening now).
● Analytics components include predictive simulation to forecast how a player would perform
in the future (against specific pitchers/parks), and the use of biometric data to predict and
prevent injuries. Therefore, their programs are likely a combination of both Business
Intelligence and Analytics.
Question #2: Why do you think baseball teams use tailored
software applications for their data analysis, instead of Excel?
● Excel is a popular and powerful program with a good statistical package
● Excel still has its limitations at the capacity it can generate the data
○ Scale of Data
○ Advanced Analytics
○ Automation & Integration
○ Customization & Visualization
● Excel struggles with performance and stability of millions of rows or complex multidimensional data
(scale of data)
○ Excel software cannot effectively manage and analyze the complexity of all data collected
○ Examples: Player tracking, Historical performance & statcast data
● Excel has limited support and is not built for doing high-level statistical analysis
● MLB’s tailored software provides real-time analysis while excel requires a lot of manual input for
certain automation which proves to be inefficient
● Excel’s visualization capabilities are limited while the MLB system are built specifically for baseball
decision-making
Question #3: Baseball teams use “scouts” to watch young men play
at high school/college levels. Player potential is based on an analysis of
doppler and video data. Do you think there will come a day when scouts
are no longer needed by major league teams?

● Data-Driven Evaluation: Player potential is increasingly based on data and doppler analysis data,
capturing objective metrics like ball position.
● Data Scientists Role: Data scientists are essential, combing strong business acumen, a deep
understanding of analytics, and an appreciation of data.
● The Argument for Continued Scouting: Objective data (analytics) can quantify performance,
traditional scouting (human judgement) may capture elements not quantifiable by sensors. (e.g.
mental fortitude, performance under pressure, or other character traits)
Question #4: Most teams have at least a dozen data scientists and other analysts in their programs.
Analysts earn high salaries and benefits. Office space, equipment, hardware and software are costly as well.
What would you roughly think the data analysis program would cost a major league team each year?

Staffing Costs (The People)

● Team Size: ~12+ data scientists and analysts per team.
● Salaries: Average $119,000 | Senior $175,000.
● Annual Base Salary (Minimum): ~$1.43 Million
○ (12 analysts x $119k)
Technology Costs (The Infrastructure)
● Data Collection: Doppler radar & high-speed cameras (across all parks).
● Data Storage: Massive systems (data warehouses, lakes) to handle 1 Terabyte per game.
● Software & Hardware: Specialized tools (Hadoop, NoSQL) for real-time data processing.
The Bottom Line: A Multi-Million Dollar Investment
● The ~$1.5M in staffing is just the starting point.
● Total cost must also include massive investments in:
○ Hardware (cameras, radar, servers)
○ Software Licenses
○ Employee Benefits