How Our Numbers Work

Why we show our math

Most scouting tools hand you a number and ask you to trust it. We do the opposite: every score in NYSFitSports can be traced to published inputs and weights, and the weights on this page are rendered from the same code that computes the rankings — they cannot drift from what the product actually does. If a player grades out well, you can see exactly why; if you disagree, you can see exactly where.

Portal IQ — the talent score

Portal IQ answers one question: how good is this player, independent of any team's system? It is a composite of six pillars, each built from the player's season statistics, with fixed, published weights — nobody tunes them per team; that's the point of a national ranking.

• Production — 24%. Box-score volume: scoring, all-around impact, and the extra possessions a player creates.
• Efficiency — 20%. Scoring efficiency, credited more at higher usage — a low-volume specialist doesn't get the same credit as a first option shooting the same percentage.
• Ball Value — 12%. Playmaking minus turnovers: what a player's ball-handling is actually worth.
• Defense — 16%. Steals, blocks, defensive rebounding, and defensive rating, combined.
• Context — 8%. Opportunity and environment: role and minutes. The schedule-strength part of this pillar is inactive this season — see the note below.
• Pedigree — 20%. Strength of the program the player produced at. Honest caveat: part of this signal is level inertia — high-major players tend to land at high-major programs — which is why this pillar is deliberately capped, not dominant. Currently inactive: the program-strength field is not populated for this season's pool, so this pillar is the same neutral value for every player and is not affecting anyone's rank.

The whole composite is then scaled down for small samples — a player with limited minutes hasn't proven their numbers, and the score says so rather than pretending 40 great minutes equals a full season.

Statistical safeguards under the hood: rate stats are shrunk toward position norms until a player's attempt volume earns them full credit (a 45% three-point shooter on 20 attempts is not treated like one on 200), and every stat feeds exactly one pillar, so nothing is double-counted.

Fit Score — your system

Fit answers the second question: how good is this player for you? You set the weights — shooting, playmaking, rebounding, rim protection, and the rest — pick your offensive emphases and defensive scheme, and every player is scored against that profile. On top of the base score, a set of visible multipliers adjusts the result, each one shown as a line item on the player's card:

• Offensive style fit — how the player's offensive role matches your emphases.
• Defensive scheme fit — how their defensive profile matches your scheme.
• Schedule strength — a discount when production came against weaker competition. Not currently applied: the schedule-strength field is not populated for this season's pool, so no player is being discounted for competition faced. Player cards say so individually.
• Sample reliability — a discount for limited minutes.
• Must-have gate — a penalty when a player falls short on something you weighted heavily.

Fit percentiles and grades are computed against the available pool — the players you could actually recruit — and the grade curve is deliberately honest:

• A+ / A / A− — top ~14% of the pool
• B+ / B / B− — next ~26%
• C+ / C / C− — the broad middle (~29%)
• D+ / D / D− — below the middle (~16%)
• F — bottom ~15%

How much to read into a single letter

We backtested the talent grade against what actually happened to 4,967 real transfers across five seasons — specifically, whether each one went on to earn a rotation role (20+ minutes a game, 10+ games) at the school they transferred to. The grade separates strongly at the range level:

• A range — 72% went on to earn a rotation role
• B range — 65%
• C range — 54%
• D range — 42%
• F — 35%

A 37-point spread from top to bottom. A-over-B-over-C held in all five seasons; in the most recent one, D and F finished level at 39% rather than separating.

The limits matter as much as the result: the plus and minus steps inside a range are not distinguishable. A B+ and a C+ came back under two points apart — inside the margin of error at this sample size. Treat the letter as the signal and the sign as texture: sort your board on A-versus-C, not B+-versus-B.

One further limit, stated because you would otherwise reasonably assume otherwise: those figures measure the talent grade. The fitgrade uses the same curve but answers a different question, and we have not backtested it the same way — doing that honestly needs outcome data tied to a specific coach's system, which does not exist yet. We would rather say so than imply a number we haven't earned.

Two rankings, on purpose

You'll see two national numbers side by side: our computed Portal IQ rank and the licensed sRating rank from our data partner. They answer the same question with different methods, and we show both rather than pretending ours is the only truth. The Role Jump badge deliberately uses the licensed rank — that gap is only meaningful against an independent benchmark. It flags players whose fit for your system is far above where the market ranks them. We validated it against what actually happened to 1,111 spring-2025 transfers: the gap predicts how much bigger a role a player takes on at his next school (median roughly +8 minutes per game), and it does not predict that he will improve as a player. We publish it as an opportunity signal for that reason.

What we don't do

• No character, injury, or off-court inference — statistics only.
• AI-written scouting summaries are grounded in the same statistics you see on screen, framed as statistical observations, and labeled as AI-generated. They never speculate beyond the numbers.
• No pay-for-placement. No program can buy a better grade.

When numbers change

Data refreshes when our licensed feed updates. After each refresh, the dashboard's "What's new" panel flags players who entered the portal and meaningful movers — a move counts as meaningful when a player's national standing shifts by several percentile points, a top-200 player moves 25+ spots, or their talent grade letter changes. Small day-to-day wiggle is deliberately ignored.