The Crystal Ball
Championship profile benchmarks, not a projection
The Crystal Ball evaluates every Division I team against 15 benchmarks derived from historical championship and deep-tournament-run profiles. It is not a projection of what will happen — it is a snapshot of how closely a team's current profile matches the dimensions that have historically separated true contenders from the field. Teams are scored by how many benchmarks they clear and grouped into four tiers.
What It Measures
Championship profile alignment: how many of the indicators that historically characterize national contenders does this team currently satisfy? The 15 checks span offensive and defensive efficiency quality, possession-level performance across the Four Factors, win percentage, schedule resume strength, shooting quality, and external validation signals like poll recognition. Each check is pass or fail based on thresholds calibrated from historical championship-caliber team data.
Why It Matters
A single efficiency number can be built through a weak schedule. A gaudy win percentage can be manufactured through non-conference scheduling. The Crystal Ball forces a multi-dimensional test — a team must pass checks across efficiency, shot-making, ball control, resume, and validation simultaneously to earn a high score. Teams that check many boxes independently are exhibiting the same patterns as historical contenders, regardless of name recognition or media narrative.
How to Interpret
Each team's score is shown as X/15 — how many of the 15 benchmarks they currently clear. Click any team row to see the full breakdown: which checks passed, which failed, and the supporting values. Tiers are assigned based on total passed checks. Championship Tier (12–15 passed): profiles that closely match historical national contenders. Contender (9–11): strong profiles with one or two meaningful gaps. Threat (6–8): real but incomplete — could make noise in a favorable bracket. Pretender (0–5): current profile does not align with contender history. The checks are not weighted — each benchmark contributes equally to the score.
Technical Notes
- The 15 benchmarks span six categories: efficiency profile (2 checks), rank-based quality (3 checks), Four Factor margins (4 checks), composite scores (2 checks), resume signals (2 checks), and shooting quality (2 checks).
- Thresholds are calibrated against historical championship and deep-tournament-run team data and are season-specific — the exact cutoffs shift slightly each year as the national baseline shifts.
- Season context (national rankings, trapezoid boundaries, max AdjEM baseline) is pre-computed once per request and held constant whether filtering to all teams or tournament-only teams. This prevents rank-based checks from shifting when the visible team set changes.
- The score (0–100) is derived directly from the passed check count: passed ÷ 15 × 100. It is an alignment score, not a win probability.
- Checks that reference national rank use a stable season-wide ranking that does not shift when filtering by conference or tournament status.
- •The Crystal Ball measures current profile alignment, not future performance. A team that passes 14 checks today can still lose in the first round.
- •Each check is binary — pass or fail. A team that narrowly misses a threshold is treated the same as one that misses badly. The score does not reflect how close a team is to each benchmark.
- •Early-season ratings are noisier, which makes early-season Crystal Ball scores less meaningful. The visualization is most useful after 15+ games.
- •Injuries, lineup changes, and transfers are not directly modeled — the checks respond only to what the underlying ratings reflect.
- •The checklist is designed for identifying historical contender patterns, not for predicting upsets, Cinderella runs, or late-season trajectories.
- •Some checks depend on external validation data (AP Poll rankings), which may lag or be unavailable early in the season.
Illustrative: a team with AdjEM +31, elite adjusted offense and defense, eFG margin +7, Rebounding Edge positive, FFI above 80, WAB over 5, and strong shooting percentages might pass 13 of 15 checks — earning a Championship Tier designation. A team with similar AdjEM but average shooting percentages and a thin resume might pass only 9, landing in Contender. Same adjusted margin, very different contender profile.