Methodology
CourtsideIQ does not rank every player like they have the same job. Each player gets one primary role, then the model grades the stats that matter most for that role across impact, scoring, possessions, finishing, defense, and contract value.
Method map
Metric lens
Score ladder
Every player is evaluated inside one primary archetype so creators, spacers, stoppers, bigs, and developmental players are not judged by the same checklist.
Stats are compared to role-specific targets. Higher-is-better stats rise normally; lower-is-better defensive stats are inverted.
REI is the main role score. VSE, PIV, PFE, and Smart CVR explain scoring, possession value, finishing, and salary context.
Players with missing inputs are marked as estimates. Contract leaderboards exclude minimum and two-way salary noise when value would be distorted.
Core Metrics
The main CourtsideIQ score. REI compares a player against the expectations of his assigned role, not the whole league at once.
Scores how valuable a player's scoring responsibility is once efficiency and usage are considered.
Scoring output + TS% efficiency + usage context + role adjustment
100 is useful scoring value. Higher means efficient scoring under meaningful responsibility.
Captures how much a player helps a possession through scoring, passing, and turnover control.
((PTS × TS%) + (AST × 1.5) - (TOV × 1.2)) / (FGA + TOV)
Displayed on a 100-style impact scale for tiering.
Measures how well a player turns possessions into clean endings through scoring efficiency, assists, and ball security.
[(PTS + AST × 1.5) - TOV] / (FGA + AST + TOV) × 100
This is about finishing plays, not future potential.
Measures whether a player is outperforming, matching, or underperforming his salary slot.
100 = fair contract value
Awards such as All-NBA and DPOY can add context, but the score is smoothed so it does not imply impossible salary multiples.
Defensive stats are handled carefully because lower numbers often mean better performance.
Role Weighting
A role score should reward the thing that role is actually paid to do. Below are the current REI inputs and weights for every CourtsideIQ role.
| Role | Input 1 | Input 2 | Input 3 | Input 4 |
|---|
Tiers
110+ Elite
100-109 Excellent
90-99 Strong
80-89 Good
70-79 Adequate
60-69 Questionable
50-59 Poor
Under 50 Wrong Role
130+ Historic Scorer
120-129 Elite Scorer
110-119 Star Scorer
100-109 Efficient Scorer
90-99 Average Scorer
80-89 Limited Scoring Role
70-79 Low Scoring Impact
Under 70 Negative Scoring Impact
140+ Generational Impact
125-139 Elite Impact
115-124 Star Impact
105-114 Positive Impact
95-104 Neutral Impact
85-94 Limited Impact
70-84 Negative Impact
Under 70 Possession Liability
150+ Historic Finisher
135-149 Elite Finisher
120-134 Great Finisher
110-119 Good Finisher
100-109 Average Finisher
90-99 Decent Finisher
75-89 Limited Finisher
Under 75 Inefficient Finisher
150+ Legendary Value
135-149 Elite Value
120-134 Great Value
110-119 Good Value
100-109 Fair Value
90-99 Slightly Overpaid
75-89 Overpaid
Under 75 Contract Liability
Credibility Rules
A player appears under one role so leaderboards do not double-count the same profile across multiple categories.
Smart CVR leaderboards filter out contract cases that would create misleading 50x or 100x value reads.
DRtg and DFG% are inverted because a lower number is better. That prevents defensive scores from rewarding the wrong thing.
If a role formula is missing inputs, the score is treated as an estimate instead of being presented as equally complete.
The model is built around regular-season player profiles so preseason and playoff samples do not distort season-long rankings.
Rankings show the player's neighborhood, role leaderboard, similar profiles, and tier labels so the number is not isolated.
Validation
A CourtsideIQ score should never stand alone. The site now treats every number like a claim: it needs inputs, a peer group, a tier label, and a visible reason why the player landed there.
Player profiles show the actual stats feeding the role score, including the role target and weight. If the stat is missing, the score is labeled as an estimate.
Role pages and player pages compare players inside the right job first. A Perimeter Stopper is judged against other stoppers, not against high-usage creators.
Every metric gets a plain-English tier so the user knows whether a score is elite, average, limited, fair value, or a warning sign.
Leaderboards focus on regular-season players with enough games and minutes to make the ranking meaningful. Low-sample players can still be viewed, but the site flags context.
Stats like DRtg and DFG% are inverted because lower is better. This keeps defensive grades from accidentally rewarding the wrong profile.
Stories and methodology examples show the basketball reason behind outlier scores, including when a famous player ranks lower in a narrow metric than his reputation suggests.
Explainer Examples
Why REI likes him: switchable defensive activity, shot suppression, rebounding support, and role fit. The score needs to show he is not just a box-score scorer.
Why VSE likes him: elite scoring volume with elite efficiency. This is the clearest example of high responsibility without an efficiency tax.
Why PFE can be lower than reputation: PFE measures finishing plays, not total offensive greatness. Luka can be an elite engine while ranking closer to average as a pure finisher.
Why CVR likes him: superstar production at a rookie-scale salary. Smart CVR should explain contract advantage without implying impossible salary multiples.