Step 1: Turn production into dollars
The unit of production is WAR (Wins Above Replacement). One win on the open market has cost roughly $8 million to $9 million per WAR in recent free agent classes; Diamond GM uses $8.5M/WAR as its baseline. That figure comes from dividing total guaranteed free agent dollars by the wins those contracts were expected to buy.
A player projected for 4.0 WAR next season is therefore worth about $34 million of on-field value in that season alone. That is the gross number — before you subtract what he actually costs.
Step 2: Project forward, not backward
A single season of WAR is noisy. Defensive metrics stabilize slowly, pitcher win totals depend on sequencing, and a 120-plate-appearance sample tells you very little. Good models weight the most recent season heaviest but blend in prior years and a peak-season anchor so that a star who had an injury-shortened year is not written off.
Then you age the projection. Position players typically peak between 26 and 28 and shed roughly 0.3 to 0.5 WAR per season after 30. Pitchers age less predictably but carry more injury risk, so their multi-year projections get discounted harder.
Step 3: Subtract the contract
Production value minus salary owed equals surplus. This is why a good player on a huge deal can be worth almost nothing in trade while a solid regular making the league minimum is a genuine asset.
Pre-arbitration players are the extreme case. A 3-WAR player making $760,000 generates something like $24 million of surplus in one season. Multiply that across five remaining control years and you understand why teams refuse to trade young, cheap regulars for two months of a rental.
For arbitration-eligible players you cannot just repeat this year's salary. Arbitration awards escalate with playing time and counting stats, so the model projects a rising curve — often 60% to 80% jumps year over year for productive players.
Step 4: Adjust for control years
Control is leverage. Six years of a young player is not six times more valuable than one year of the same player, because the far-out seasons are uncertain and the player gets more expensive. But it is meaningfully more valuable.
Diamond GM applies a compressed control multiplier — roughly 0.88x for a pure rental up to about 1.07x for six-plus years — rather than the brutal discounts some public models use. Deadline history supports this: contenders regularly pay real prospect capital for two months of an ace, because the wins arrive when they matter most.
Step 5: Layer in role and position
WAR is not perfectly comparable across roles. Relievers throw 60 innings, so even elite ones rarely clear 2.0 WAR, yet they command significant trade returns. Catchers accumulate less playing time but are scarce. Platoon bats have their value suppressed by playing time rather than by quality.
The fix is small, bounded multipliers — a modest premium for catcher, shortstop and center field, a leverage floor for relievers so a good arm never grades as a large negative, and caps that prevent a role player from ever showing up as a franchise cornerstone.
Where prospects fit
Prospect value follows a steep curve. A consensus top-five prospect can be worth $50 million or more in surplus terms; a top-100 name in the back half of the list lands closer to $10-15 million; a team's ninth- or tenth-best prospect is worth low single-digit millions.
Diamond GM builds that curve from national top-100 rank when available, falling back to organizational rank, then adjusts for age relative to level, ETA, position and current-season statistical performance. Each adjustment is intentionally dampened so they cannot compound into absurd numbers — a hot month in High-A should nudge a value, not double it.
Reading a grade correctly
Surplus value is a starting point for an argument, not a verdict. It cannot see medicals, clubhouse fit, roster crunch, service-time manipulation, or the simple fact that a team 1.5 games out of a wild card values 2026 wins far more than 2029 wins.
The right way to use a trade grade is directionally: a deal inside roughly 15% is basically even, and anything beyond that is worth explaining. If your gut and the model disagree, one of you is missing information — figure out which.