March Madness Data set Analysis

By datastam_team · April 4, 2026

The line chart shows how average Adjusted Tempo has changed across college basketball seasons from 2002 to 2026. Overall, tempo has slightly decreased by 0.75 possessions per game over this 24-year span, going from 68.06 in 2002 to 67.31 in 2026.

The analysis successfully identified which conferences produce the most Final Four teams across all seasons. Two visualizations were generated: a bar chart ranking the top 15 conferences by total Final Four appearances, and a stacked bar chart showing how the top 5 conferences contributed Final Four teams season by season.

Yes — teams with above-average Effective Height do achieve better Adjusted Defensive Efficiency. Taller teams average a score of 101.83 compared to 106.40 for shorter teams, a difference of 4.56 points. Since lower scores mean better defense, this confirms that height is associated with improved defensive performance.

Yes, there is a modest positive relationship between coaching tenure and team Net Rating. The correlation coefficient of 0.1359 indicates that coaches with a longer Active Coaching Length Index do tend to produce slightly higher Net Rating teams on average, though the relationship is relatively weak.

Tournament winners stand out dramatically from all other teams across all three efficiency metrics. Based on 22 tournament winners compared to 8,658 other teams, the data tables show clear and significant differences in performance.

Tournament winners show dramatically better Adjusted Defensive Efficiency Ranks compared to non-winners. Two visualizations were generated — a histogram overlay and a box plot — both clearly illustrating how winners cluster near the very top of the defensive rankings while non-winners are spread across the full range.

The analysis produced two visualizations comparing bench contribution percentage across different levels of post-season tournament advancement — from teams that didn't make any tournament all the way up to champions. A box plot shows the distribution of bench percentages at each advancement level, while a bar chart displays the average bench contribution for each group.

A data table was generated comparing the average team Experience for Final Four teams versus non-tournament teams. The table breaks down experience averages across different team categories, giving you a clear picture of how tournament success relates to team experience levels.

Of the top 2 seeds flagged as vulnerable, 85.7% (18 out of 21) actually lost before reaching the Final Four — confirming that the vulnerability flag is a strong predictor of early exits. A bar chart breaks this down by season, showing how the rate varied year over year.

On average, teams' AdjEM ranks shift by about 3.15 positions between pre-tournament and final rankings. The median shift is just 2.0 positions, meaning most teams don't move dramatically. This analysis covers 7,962 team-seasons of data. A bar chart shows how this average shift varies by season.

Championship teams have a lower average turnover percentage (TOPct) of 17.53% compared to 18.22% for all NCAA tournament teams. This means championship-winning teams are better at protecting the ball, turning it over less frequently than the average tournament participant.

Across 25 seasons of data, the team with the highest Adjusted Efficiency Margin (AdjEM) each season went on to win the tournament 12 out of 25 times — that's a 48% conversion rate. The bar chart shows each season's top-ranked team by AdjEM, with gold bars indicating tournament winners and blue bars indicating teams that did not win it all.

Three visualizations and supporting data tables were generated to explore how 3-point attempt rates have evolved and whether they predict tournament success. From 2002 to 2026, the average FG3Rate climbed from 32.2% to 39.5% — a +7.2 percentage point increase — confirming a clear league-wide shift toward 3-point reliance. However, higher 3-point volume does NOT appear to be a recipe for championship success. In fact, March Madness champions average a slightly lower FG3Rate (33.3%) than other tournament teams (34.9%) and even non-tournament teams (35.3%). The gap is small (~1.6 pp), suggesting that how well teams shoot the 3 (efficiency) matters far more than how often they attempt it.

Two visualizations were generated to explore the relationship between conference offensive rebounding percentage (ORPct) and tournament wins. The bar chart with overlay shows the top 15 conferences ranked by average ORPct alongside their tournament win counts, while the scatter plot maps each conference's average ORPct against total tournament wins (with bubble size representing team count). Together, these charts reveal which conferences lead in offensive rebounding and whether that translates to postseason success.

Yes, Final Four teams do show slightly more balanced scoring across positions compared to other teams. The difference is modest but consistent — Final Four teams have a lower standard deviation in position scoring (3.97 vs 4.08) and a smaller scoring range (9.64 vs 10.07), indicating more even distribution across Center, Power Forward, Small Forward, Shooting Guard, and Point Guard.

The analysis compared average AdjEM (Adjusted Efficiency Margin) for conferences across two five-year periods: 2017–2021 (early) vs. 2022–2026 (late). Two visualizations were produced showing which conferences improved the most over the last decade.

Among 1,603 March Madness tournament teams, there is a weak negative correlation (-0.2530) between Steal Rate and Adjusted Defensive Efficiency. This means teams with higher steal rates tend to have better (lower) defensive efficiency ratings, though the relationship is not particularly strong. Two visualizations were generated: a scatter plot showing the distribution of all teams with a trendline, and a box plot breaking down defensive efficiency by steal rate quartile.

The analysis produced two visualizations breaking down Net Rating by tournament seed. The first bar chart shows the average Net Rating for each seed number, illustrating how higher-seeded teams (lower seed numbers) generally have stronger net ratings. The second chart highlights which seeds outperform a linear expectation model — seeds shown in green are punching above their weight relative to what you'd expect based on a straight-line projection from seed 1 to 16.