Every stat we use to evaluate tournament teams — what it measures, why it matters, what the data says, and how much weight it carries in our model.
Points allowed per 100 possessions, adjusted for opponent quality. The single most predictive metric for tournament survival. Defense travels — offense gets disrupted by pressure, shot-clock manipulation, and hostile environments.
Points scored per 100 possessions, adjusted for opponent quality. Elite offenses can survive poor shooting nights and find points in multiple ways. The pairing with AdjDE is what creates tournament-proof profiles.
When evaluating auction value, the most important question isn't "how good is their offense?" or "how good is their defense?" — it's whether both are elite simultaneously. Defense alone with a weak offense reaches E8 at just 7.5%. Offense alone with weak defense: 9.7%. Neither compensates for the other. Both in the top 20 together: 52.3%.
Percentage of available offensive rebounds secured. Second-chance points are the most reliable way to manufacture extra possessions in grinding, low-efficiency tournament environments when half-court sets are well-scouted.
Effective field goal percentage — field goal percentage adjusted for the extra value of three-pointers. Captures shooting quality more accurately than raw FG% by accounting for the 3pt premium. Teams that shoot efficiently from high-value spots win close games.
Percentage of offensive possessions that end in a turnover. Ball security is a floor metric — teams that give it away create opponent momentum and destroy their own half-court offense at the worst possible moments in tournament play.
Percentage of available defensive rebounds secured. Ends opponent possessions definitively — every missed defensive rebound is an invitation for the upset. Especially critical in close games where one second-chance bucket changes everything.
Percentage of offensive shot attempts (and points) generated at the rim — dunks and close two-pointers. Teams with high rim access shoot higher percentages, draw more fouls, and generate the most efficient shots in basketball regardless of game environment.
Percentage of opponent shot attempts that come at the rim. Defenses that force opponents to shoot farther away degrade their efficiency — limiting dunk/close-2 access is the most reliable way to hold teams below their expected scoring output.
% of field goal attempts taken from three. High 3PA rates introduce more game-to-game variance — great when they're falling, catastrophic when they're not. In the tournament, one cold 3-pt shooting night is all it takes.
% of opponent field goal attempts from three. Defenses that force opponents into long twos or drive-to-contact situations limit variance — the opponent can't get hot from three if you don't let them shoot threes.
Getting to the line more than your opponent, while also shooting better from the stripe. Free throws don't feel exciting until the last two minutes of a tournament game — then they decide everything. This edge score rewards both foul-drawing and FT execution.
Effective field goal percentage allowed to opponents. Complements AdjDE by isolating shooting quality suppression independent of pace. A team can have a great AdjDE by forcing turnovers while still allowing easy shots when they do get attempts.
Average years of experience in the rotation. Intuitively appealing — older teams should handle pressure better. But the data says otherwise: once efficiency metrics are accounted for, experience adds almost nothing predictive.
| Metric | Wt | Direction | E8 signal strength | Flag threshold |
|---|---|---|---|---|
| AdjDE — Adj Defensive Efficiency | 32% | Lower = better | Top-15 → 28.2% E8 rate | Rank 201+ = red flag (0-for-81 historically) |
| AdjOE — Adj Offensive Efficiency | 22% | Higher = better | Top-15 → 32.4% E8 rate | Rank 150+ with weak D = elimination profile |
| OR% — Offensive Rebound % | 11% | Higher = better | Top-30 → 20.9% E8 rate | Rank 250+ = low O-reb flag |
| eFG% — Off Effective FG% | 10% | Higher = better | Top-15 → 2.1× E8 likelihood | Rank 250+ = offensive floor concern |
| TO% — Turnover Rate | 8% | Lower = better | Bottom-30 → 2.6% E8 rate | Rank 200+ = high TO flag |
| RimOff% — Offensive Rim Share NEW | 2.5% | Higher = better | Top-30 → 18.4% E8 rate | Rank 300+ with low ORB = weak rim flag |
| RimDef% — Defensive Rim Allowed NEW | 2.5% | Lower = better | Top-30 both rim metrics → 21.7% | Rank 250+ = permissive rim defense |
| DR% — Defensive Rebound % | 2.5% | Higher = better | Moderate — floor effect | Below 68% = concern in close games |
| Net FTR — FT Rate Edge Score | 2% | Positive = better | Strong in late-game scenarios | Negative edge = disadvantage in close games |
| Def eFG% — Opponent Effective FG% | 2% | Lower = better | Complements AdjDE | Rank 200+ = soft perimeter defense |
| 3PAr — 3-Point Attempt Rate | 1.5% | Moderate = better | High 3PAr → lower E8 rate | Top-15 nationally = 3-pt volatile flag |
| Opp 3PAr — Opponent 3PA Rate | 1.5% | Lower = better | Limits opponent variance | Rank 300+ = opponent 3pt exposure |
| Experience | 0.5% | Higher = (supposedly) better | No clear gradient historically | Rank 280+ = young roster flag |