The culmination of my thesis project at Pratt, Baseball is a Full Context Sport: Implementing and Contextualizing Advanced Baseball Statistics for Modern Audiences introduces a way to visually integrate advanced baseball statistics into the television broadcast of games, and, with a mobile app called Game State, give users a way to engage with the game's more sophisticated metrics in real time.
The full hypothesis reads:
Baseball statistics are the main entry point to fully engaging in, enjoying, and understanding the sport of baseball.
Yet these numbers as they currently appear in the media and in stadiums provide little real insight or context into the players they are intended to describe.
I hypothesize that it is possible (if not warranted) to change the nature of these statistics to better suit today’s audiences by expanding on the common traditional statistics most newspapers, baseball stadiums, and television broadcasts use today, and by introducing carefully contextualized advanced statistics (e.g. sabermetrics) into the mainstream.
By doing so, I hypothesize that greater understanding and accessibility of sports and the athletes therein may be achieved.
Game State, the app
Much of the book is devoted to outlining the problem of the current state of baseball statistics as they relate to the modern baseball audience; eg how they are unintuitive to outsiders, fail at explaining themselves clearly, and are fail to attract new fans.
My solution to this problem was to create a mobile app that
1) shows advanced and traditional statistics in real time, arranged and displayed in a way that they can be intuitively consumed; not solely by numbers, but by information-rich graphics and visual overlays that interact with the game occurring on the television screen;
2) accommodates for baseball fans across the familiarity spectrum, from new fans to staunch traditionalists to sabermetricians;
3) condenses a wide range of information from various sources into cohesive, manageable bits, and promotes the most valid, accurate, and predictive statistics available today;
4) encourage users to learn about new statistics they do not know alongside familiar ones.
The tool, in short, should be interactive and allow for users to customize their experience.
Disrupting the television broadcast experience
While developing the interface for the app, the design naturally lead me to exploring the possibility of overlaying the corresponding information that was being viewed on the mobile device onto the television screen, propelling the utility of this tool beyond being "just" an app and into being a potentially revolutionary broadcast television experience.
I identified the component parts of a baseball game that would comprise the app. This required a comprehensive analysis of the baseball statistics being used today, understanding their natural groupings (e.g. offensive statistics, defensive statistics, related aspects of gameplay etc.), determining which metrics would be valuable or interesting to a broad television audience, and how I might better represent these statistics visually. As a result, the app is divided thusly:
Hitter vs. Pitcher Matchup
This mode focuses on the most common interaction in baseball: one hitter versus one pitcher. Information displayed in this mode includes hitter vs. pitcher history, count-specific batting statistics, and strikezone heatmaps that the user can tailor to different pitch types (batter success vs. fastballs, pitcher success with sliders, etc.)
Total Field View
This mode highlights the play that occurs in the field and on the bases. Batter spray charts illuminate hitter tendencies, while defensive shifts show how fielders are playing the hitter. Defensive Runs Saved (DRS) information shows detailed fielder information.
Win Probability
Win Probability mode eliminates traditional and advanced player statistics altogether and focuses on one simple question: how much has each player helped (or hurt) their team win? The app screen displays each team’s Win Expectancy and at-bat WPA scenarios.
Hitter vs. Pitcher
In Hitter vs Pitcher mode, the red and blue duotone bar displays the offense (always in blue) and the defense (always in red). In this bar, users see the current batter, the current pitcher, and the on-deck batter.
Plotting Skill Level
The bar graph below each player places the user-chosen statistic on a scale from Poor to Excellent, and displays the numerical estimates for common distinct ratings (for example, for pitchers a 4.00 ERA is widely considered Average, 3.75 Above Average, etc.). This allows users to understand the relative skill level of a player in two ways: visually, by looking at the where the player's number appears on the scale; and numerically, by being able to see the player's statistic alongside the distinct Poor/Excellent ratings.
Users can choose to view different statistics from a drop-down list, and can easily access more information to learn more about them (eg what it wants to measure, the scale it uses, and the formula). For batters, the list includes AVG, OBP, SLG, OPS, OPS+, ISO, wOBA, and WAR. For pitchers, ERA, ERA+, FIP, xFIP, WHIP, K%, BB%, LOB%, and WAR.
Batter Career vs. Pitcher
Below the scales is the performance history of the current batter vs the current pitcher. Although sabermetricians have shown this information to be poor at predicting future outcomes, it is nevertheless interesting, and helps illuminate the matchup for the average fan.
Count-Dependent Statistics
Some hitters are great at hitting first-pitch strikes. Some pitchers are particularly skilled at getting batters out with a 2-strike count. This window, located below the Batter Career vs. Pitcher pane, shows the pitcher’s and batter’s count-dependent statistics as the count changes during the course of the at bat.
Strike Zone Heat Maps
Users can toggle hot and cold zones, or heat maps, for both hitter or the pitcher. Users can also isolate pitch types (fastball, curve ball, etc.) in these heat maps by tapping the “select pitch type” button beneath home plate.
Leverage Index
The thin bar at the bottom of the screen is dedicated to mapping Leverage Index, which is a way to quantify the general level of “pressure” during a game. High leverage situations, for example, are ones in which the result of the current at bat will cause a large shift in the Win Expectancy of both teams. We know that a one-run game in the ninth inning is more tense than a one-run lead in the second, but LI helps us to quantify how much more.
The demarcations in the bar are spaced in exponentially smaller increments. This is to visually balance the scale, where low-to-mid leverage situations exist from only 0 to 3, and anything over 3 is considered “very high” leverage (the highest LI being 10.9, where the home team is down by one in the bottom of the ninth with the basses loaded and two out).
Total Field View
Total Field View is dedicated to displaying 1) the placement of defenders/defensive shifts, 2) base runners, and 3) batter spray charts.
Defense and Offense on the Basepaths and in the Field
Players have a plus sign next to their name and position. Tapping on a defender will bring up his Defensive Runs Saved (DRS) information. Tapping on a batter or base runner will reveal information related to the player's current position on the field, such as the player’s Speed Score (Spd), which informs on the player’s base running/stealing ability.
Users can view individual defensive statistics, or they can choose to group defenders together to get a sense of the overall defensive ability of the team’s infield, outfield, or pitching battery, all with relevant position-specific statistics (such as rARM for outfielders with strong throwing arms, or rGDP for participating in double plays, etc.).
Spray Charts and Defensive Shifts
Users can toggle Spray Charts, which display the frequency a batter hits to certain parts of the field, and Defensive Shift, which allows users to see how much a defense has shifted out of traditional locations to defend against the hitter.
Win Probability
The idea behind the Win Probability mode is rather simple: which player has most helped their team win the game? This game-viewing experience nearly the eliminates statistics that so many find confusing.
Win Probability Chart
The top section of the screen shows the overall Win Probability Chart, which is a box of 18 columns representing the tops and bottoms and each inning, and a jagged line in the middle adjusting up and down as either team increases or decreases its chances to win the game. The median of the box represents a 50% chance of winning for either team, and the top and bottom lines represent 100%.
At-Bat Win Expectancy
The larger graph shows the Win Expectancy (WE) during the current at-bat (shown on the left) and possible WE after the at-bat (shown on the right). The lines represent WE added or subtracted for potential results from the current at-bat and the change in Win Expectancy associated with those events.
These WEs add up to form a player’s Win Probability Added (WPA), which are tracked in this app by overall season and individual game. A player’s WPA has been added to the duotone player bar at the top, allowing users to see which player has helped or hurt their team the most.
Thesis presentation
For those interested in reading the full text, you may do so here: