Average Limitations

When averages misinform and mislead —precision, causality, and predictability provide a repeatable path to better decision-making. The history of Baseball offers many examples for how statistics can be improved. From Henry Chadwick’s pioneering work in the 1800s to modern data-driven strategies, baseball has shown how continually refining metrics can significantly evolve game understanding and strategy.

Here are three examples of early baseball stats:

1. Runs Batted In (RBI): Introduced in the early 20th century, RBIs provided a way to measure a player’s contribution to scoring runs and their value as a clutch hitter. Prior to this, the focus was mainly on batting averages, which didn’t fully capture a player’s ability to drive in runs when it mattered most. The RBI metric shifted attention to a player’s situational effectiveness and impact on the team's ability to win games.
2. Fielding Percentage (FPCT): Early baseball statistics often overlooked defensive capabilities, focusing mainly on batting and pitching. The introduction of fielding percentage in the late 19th century allowed teams to better evaluate defensive performance. By calculating the ratio of putouts and assists to total chances, fielding percentage offered insight into a player's reliability in the field. This metric pushed teams to pay closer attention to defense as a critical component of winning strategies.
3. Stolen Base Tracking: In the late 1800s, teams began to recognize the strategic importance of base-stealing and started tracking stolen bases as a separate statistic. Before this, baserunning prowess was not quantified, leaving teams without a clear measure of a player’s speed and ability to disrupt pitchers and defenses. By counting stolen bases, managers could identify and leverage players who brought added value through their speed and aggressive base-running, leading to more dynamic and strategic gameplay.

These early "averages" improved player evaluation and management decision-making, but there was still plenty of room for improvement.

This article on Historic Baseball highlights how the Moneyball approach, adopted famously by the Oakland A's in the early 2000's and rooted in sabermetrics, brought another set of advancements in precision, causal understanding, and predictability to baseball statistics.

Here’s a breakdown of the Moneyball-related improvements and how they reflect a general path for making progress in any area where measurement, learning, and decision-making is important.

Precision:

The Moneyball approach showcased how using advanced metrics greatly increased the precision of player evaluations. While traditional statistics like batting average and ERA provided a broad overview, newer metrics like On-Base Percentage (OBP) and Slugging Percentage (SLG) offered clearer, more detailed insights into player value. This precision allowed teams to identify undervalued players whose specific contributions—such as consistently getting on base—were crucial to winning games. By focusing on precise, impactful skills rather than generalized assumptions, teams were able to make smarter, more effective roster decisions and game-time decisions.

Causal Understanding:

Moneyball also marked a significant improvement in understanding causality within the game. Traditional scouting often relied on surface-level metrics that correlated with winning but didn't necessarily reveal the root causes of success. By shifting the focus to metrics that showed clear causal links to scoring and preventing runs, teams could better identify which player actions directly contributed to victories. This led to strategies that prioritized actions with proven, tangible impacts, reinforcing evidence-based decision-making over intuition.

Predictability:

The use of detailed sabermetrics greatly enhanced predictability in player performance and game outcomes. Unlike subjective scouting methods, which were prone to bias and inconsistency, data-driven models could forecast results more reliably. This meant that teams could construct rosters and implement strategies with greater confidence, maintaining competitive performance even with limited budgets. The ability to predict outcomes with greater accuracy created a sustainable advantage that set the Moneyball approach apart from conventional methods.

General Path of Improvement:

These advancements in precision, causal understanding, and predictability illustrate a common path for progress in any research or learning field:

1. Enhanced Precision: Shifting from broad averages to detailed, relevant metrics improves the accuracy of assessments and strategies.
2. Understanding Causality: Moving from correlation-based analysis to causally linked metrics reveals what truly drives outcomes.
3. Better Predictability: Precise, causal understanding drives better forecasting, reducing uncertainty, and promotes more consistent decision-making.

This trajectory of progress shows that initial reliance on simple averages can provide a starting point, but are limiting compared to more refined measures. By moving toward precise, causally relevant insights and employing them for better predictability, any discipline—from sports and business to software development —can keep improving.