- Time of possession, seconds per touch, on-ball %. Through 2023-24, the NBA's number. For 2024-25 and 2025-26, the values the NBA served before the change, which we saved. From 2026-27, our estimate: 1.42 seconds per touch + 0.88 per dribble − 2.77 per game, times a fixed factor for each player set from 2025-26 (new players have none). Seconds per touch from 2026-27 is scaled by 0.92 to stay comparable with earlier seasons.
- Compare rankings, not raw numbers. Touches (−8%), front-court touches, drives (+38% to +50%), drive passes (+60% to +89%), drive shots (+25%), paint touches (+13% to +30%), shots defended (−9%), pull-up shots (−5%), open and wide open shots. The NBA counts these differently from 2024-25. Game logs change there; season totals and career logs change at 2026-27.
- Not comparable across the change. Post touches, elbow touches, secondary assists, tight and very tight shots. The NBA reclassified these from 2024-25 and individual players moved around. Game logs change there; season totals and career logs change at 2026-27.
Value Board
? What do these metrics mean?
- Salary (% Cap)
- The player's salary expressed as a share of that season's salary cap. Cap-relative figures are comparable across eras; the nominal dollar amount of that year is shown underneath.
- Surplus Cap Value
- The gap between what the player's play was worth (per the active valuation model) and what they were actually paid, as a share of cap. Positive = outplayed the contract.
- Surplus Wins
- That same surplus expressed in wins: wins provided above what the contract's salary should buy at that season's market rate per win.
- Production Value (% Cap)
- The full value the player provided — surplus plus salary — as a share of cap. Roughly: what their production would cost at market rate.
- LEBRON model
- BBall Index's LEBRON impact metric and its wins-above-replacement translation, sourced directly.
- DARKO model
- Kostya Medvedovsky's DARKO DPM rating, converted to wins via the Partnow/Noh valuation approach (impact above replacement × minutes share) along with a sliding scale multiplier based on player DPM, priced using estimated per-season $/win.
Position, role, salary, and impact-model value — one row per player-season with a logged contract. Dollar figures shown as % of that season's cap (era-neutral), with nominal underneath.
X-axis is Salary (% of cap). Dotted lines mark median salary and the break-even (or median) value line, splitting the chart into value quadrants.
The shared Season filter doesn't apply here — Trends always spans every season. Click legend entries to isolate a line.
Dashed white line = mean · solid tick (in each group's color) = median, with thinner ticks above/below marking P25/P75.
To Date = seasons of the deal already played, valued from actual production · Remaining = unplayed seasons, projected from the seed · Full Deal = both. Cap-points add each season's % of that year's cap (a 3-year deal at 10% = 30).
Click a row for the season-by-season breakdown.
Sums each team's active contracts over their remaining (unplayed) seasons only. This is not a payroll: cap holds, dead money, two-way and Exhibit 10 deals are not included.
About Value Board
Value Board is a tool for assesing player value relative to their contracts. It allows you to toggle between two impact metrics to form the 'base' of the evaluation, LEBRON and DARKO. All monetary values(% of the salary cap, estimated 'cost' per win) are normalized according to the salary cap for the corresponding seasons
Data & Credits
- LEBRON impact metric — Basketball Index (BBI) / fanspo. https://fanspo.com/bbi-role-explorer/lebron
- Offensive Archetype & Defensive Role classifications — Basketball Index.
- DARKO DPM — created by Kostya Medvedovsky (darko.app).
- DARKO valuation methodology — adapted from Seth Partnow's The Midrange Theory (Ch. 7) and an implementation by Steph Noh. https://github.com/StephenNoh/nbasalarymodel/blob/main/README.md
Methodology
Why cap%? Because the implications of any contract depend entirely on the payroll. A 20 million a year contract in 2016-16 was worth nearly 29% of the cap, superstar max territory. That same 20 million is under 13% of the cap in 2025-26. Using cap % allows us to bypass this problem, since instead of comparing raw dollars, we are essentially comparing how much of the available payroll a contract is taking. This makes cross era contract comparisons significantly more valuable.
Surplus Wins The estimated wins added by a player relative to the contract. To calculate this, you need to estimate the 'cost' of a win each season based on the cap environment. Then, it's essentially fancy division, with the players estimated wins added via the corresponding impact metric subtracted by the win cost of their current salary(ie WAR - (salary/wincost))
Surplus Cap Value same as the above, with the wins converted to dollars, then cap % using that same win cost estimate.
Production Value (% Cap) This is essentially WAR converted into cap form. (WAR /win-cost)
Two valuation models, one framework. LEBRON's wins-above-replacement estimate (WAR) is sourced directly from Basketball Index — it was not calculated. DARKO ships a rating (DPM) but no wins translation, so I used this proxy: (minutes / 1475) × (DPM + 3), along with a sliding scale based multiplier (+10% bonus for MVP-caliber, -10% penalty for replacement-level players) following the general approach used by Steph Noh in the explanation linked above. As this is primarily a backwards facing tool, I use the win cost value associated with the corresponding season.
Contracts tab. Every active contract on file (any season after the last completed one) is valued season by season. A rookie deal and the extension that follows it with the same team are merged into one functional contract; an older deal that a newer one replaced (waived, bought out, re-signed elsewhere) is dropped. Seasons already played use that season's actual production value; unplayed seasons hold a seed season's production as a share of the cap and apply it to each future cap, priced at the newest season's $/win (scaled with the cap). The seed is the most recent season, the better of the last two (for players who missed big chunks of a season), or their average. Salaries come from the contract-history snapshot (paid salary for past seasons, Spotrac for future ones). Player options (and mutual/early-termination options) are never counted, since the player holds that choice; team-option years count by default and can be excluded. Totals are in cap-points: each season's surplus as % of that season's cap, summed. Two-way, Exhibit 10 and superseded reports are left out, and the Teams view is not a payroll.
Limitations & Disclaimers
- The DARKO valuation is my own construction not an official DARKO output — make sure you don't treat it as such.
- The Projection tab is based on transposition not prediction. It takes the selected player season, calculates their estimated value add for that season, then adjusts that value based on the contract average/selected season. Ie it's not 'Jalen Williams will be this valuable in 2027'. It's, 'If Jalen Williams plays like he did in 2025, he will produce this much value relative to inputted contract structure'.
- Distribution groups with fewer than 8 player-seasons are excluded. This is only relevant regarding some Offensive/Defensive Role Designations from Bball Index.
DARKO and DARKO based value calculations are going to be significantly stronger in capturing impact over a multi season span than a single season span. Which is true for impact metrics generally(bigger samples just help a lot here), but is especially true for DARKO given that the model is designed as a predictive, forward facing estimate.