Pricing the LaMelo Trade, Part 3: The Bill

Charlotte's press release says Minnesota granted "swap rights in 2028, 2029, and 2030," and the phrasing does what press releases do: it makes three different financial instruments sound like triplets. They are not. Read the actual fine print, the pile of conditional language Minnesota's earlier trades left stapled to its own draft picks, and the three swaps come apart in your hands.

The 2028 swap is the clean one. Two picks, Charlotte takes the better, no strings.

The 2029 swap is an option so deep out of the money it barely exists, which is finance-speak for a bet that almost never comes due. Minnesota's own 2029 first already belongs to Utah's pick pool unless it lands in the top five, leftover fine print from the Gobert trade, so Charlotte's swap right only wakes up in the narrow futures where the Wolves have gotten bad enough, fast enough, to keep their own pick. Rarely alive. Enormous when it is.

The 2030 swap is subordinated debt, which means Charlotte stands at the back of the line. Before they touch anything, Minnesota's pick passes through a San Antonio and Dallas stack that can strip its upside, and Charlotte swaps against whatever crawls out.

Those were the characterizations we wrote down before pricing anything. Now the machine has priced them, fifty thousand futures at a time, and two of the three survive contact. The third needed an amendment that tells you something about Charlotte.

How you price an option on a basketball team

One paragraph of theory, no Greek letters. A swap right is worthless in most futures: if Charlotte's own pick is better, the right expires quietly and costs Minnesota nothing. Its entire value lives in the minority of futures where Minnesota's pick is the better one, which is why an average taken over all futures looks small while the thing itself is dangerous. The average payout of a fire policy is tiny; that tells you nothing about the year your house burns. And the three swaps are not independent bets, because they all ride the same underlying asset: if Edwards leaves in 2029, the 2029 swap, the 2030 swap, and the 2033 pick all get more valuable together. So the only honest total is computed path by path, summing what the package pays in each future and then looking at the distribution of that sum. Never add the averages of things that happen together.

Here is what each instrument pays. The currency is a pick's first four seasons of value above a replacement-level player (the stat sheet calls it 4-year VORP; this piece will mostly just say wins), the same currency we use to grade drafts:

The 2028 swap exercises in 41.6 percent of futures. Its typical payoff is nothing, the median is zero, and its 90th-percentile payoff is 2.2 wins, with a mean of 0.7. Clean, modest, real.

The 2029 swap exercises in 12.3 percent of futures, and here is the fine-print stat of the piece: its 90th-percentile payoff is still zero. You can sit at the 90th percentile of futures and this option has paid Charlotte nothing. But its mean is 0.7, the same as the 2028 swap's, and both facts are true at once because when the 2029 swap finally wakes up, it pays about 5.3 wins on average, more than the 2033 pick's whole expected value. This is not a draft asset. It is a catastrophe insurance policy that Charlotte bought against exactly one event: the Wolves collapsing so hard by 2029 that a top-five pick stays home. It pays in the worlds Part 1 warned about, and almost nowhere else.

The 2030 swap is where the skeleton needed its first amendment. We characterized it as subordinated debt, upside pre-stripped by the San Antonio and Dallas stack, and structurally that is still true. But it prices as the most valuable swap of the three, not the least: exercised in 66.3 percent of futures, mean payoff 1.3 wins, 90th percentile 4.6. The reason is not Minnesota's side of the trade at all. It is Charlotte's. By 2030 the simulated Hornets are cresting, a median 49-win team picking in the twenties, and a swap right held by a good team is cheap to exercise against almost anybody. Charlotte did not buy 2030 upside from Minnesota. They bought the right to throw away their own late pick in the exact season they expect to stop needing it.

Two smaller entries close the ledger, printed so that every asset in the package appears in the total exactly once. The first already happened, on draft night, when Minnesota handed over the 28th pick and took back the 33rd. Both slots were known the moment the deal was signed, so there is nothing to simulate; at the slot curve's expected values the exchange runs 0.82 wins out against 0.57 in, about a quarter of a win of four-year value.

The second is the three second-rounders, 2029, 2032, and 2033, which get the flat treatment the fine print deserves: each carries the slot curve's average value from picks 31 through 45, about 0.37 wins apiece, 1.1 wins for the trio. If all three landed at the top of that window the trio would be worth about two wins; at the bottom, about a quarter of one. Not worth simulating, worth counting. Fold both entries in and the all-in package comes to about 6.2 wins of four-year value in the mean. The option analysis below stays on the picks and the swaps, which is where all the uncertainty lives.

Where the cost lives

Every one of the fifty thousand futures answers two questions: does Edwards stay, and does Charlotte get good? Sort them into the four possible buckets and ask each bucket what the full package pays Charlotte. Baseline, Edwards stays and the Hornets stay ordinary, is 24.7 percent of futures and pays 4.0 wins on average. Edwards leaves while the Hornets stay ordinary: 33.3 percent, paying 5.3. The Hornets ascend to a 45-win-or-better 2033 while Edwards stays: 18.0 percent, paying 4.2. And the bucket where both things happen, Edwards gone and Charlotte good, is 24.0 percent of futures and pays 5.6.

The skeleton expected that last bucket to be rare and brutal, a thin tail carrying an outsized share of the bill. That is the second amendment, and it is worse news than the expectation: the joint tail is not rare. A quarter of simulated futures put a departed Edwards and an ascendant Charlotte in the same timeline, partly because the same 2029 summer that decides Edwards' address also sits mid-crest on Charlotte's arc. The bill is not a lightning strike Minnesota is hoping to dodge. Every bucket pays at least four wins on average; the bad buckets just pay more. What Minnesota is actually hoping is that its own future comes from the cheap end of each bucket's range, and to be fair, that range is wide: the total package pays less than nothing in about one future in ten (a late pick's first four years can be worth less than a replacement player) and more than 13 wins in another one in ten.

Add it up the honest way, path by path: the headline package, the 2033 first plus all three swaps, pays Charlotte a mean of 4.9 wins of four-year value, median 3.1, with an 80 percent interval from minus 0.8 to 13.0. For scale, the unprotected 2033 first alone accounts for 2.2 of that mean, and its own interval runs from minus 1.4 to 8.1. The fine print about whether a No. 1 overall Minnesota pick in 2030 would still be swappable, a genuinely unresolved contract question, moves the total by about 0.2 wins, so we ran it both ways and it changes nothing that matters. Priced in win shares instead of VORP as a robustness check, the picture holds.

We ran the machine on 2013

Numbers this far into the future have to earn trust somewhere, so we sent the machine back in time. Same pipeline, same models, fed only what was knowable in the summer of 2013, and pointed at the most infamous pick package in league history: the firsts Brooklyn sent Boston for Kevin Garnett and Paul Pierce. The picks that became the third selection in 2016 and the first selection in 2017, the ones that became Jaylen Brown and Jayson Tatum, plus a 17th and an 8th along the way. The question is not whether the machine predicts Brown and Tatum. Nothing predicts Brown and Tatum. The question is whether reality landed inside the intervals the machine would have printed at signing, or whether the truth of 2014 through 2018 lives in a tail the machine did not know it had.

Here is what the machine, knowing only 2013, said about the four deliveries, and what actually showed up. The 2014 pick: the model's 90% interval ran from 4th to 28th with a median of 18; reality delivered 17th, dead center. The 2016 pick: interval 2nd to 29th; reality delivered 3rd, inside the interval but out at its edge, a spot the model gave about a 9 percent chance of reaching. The 2018 pick: interval 2nd to 29th; reality delivered 8th, comfortably inside. And then 2017, the swap year, the Tatum year: the interval ran 2nd to 29th, and reality delivered 1st. Outside the interval. The machine gave that outcome about 3 chances in 100, and it happened.

We are not going to hide the miss, because the miss is the most instructive number in the piece. Getting to 1st overall required Brooklyn to collapse all the way to the league's worst record and then hold the lottery, faster and harder than an honestly built as-of-2013 model considered 90%-plausible. Under our strict scoring rule, three of four inside means the primary replay cell reads red, and it stands red. But read the direction: when this machine errs on a pick package sold by a team betting on itself, it errs by understating how bad the collapse can get. Applied to our own bill, that error runs one way. It makes the numbers in this piece more likely too small than too large. One honesty note on the gate itself: the pre-declared standard wants this replay plus a 2019 Paul George replay and a negative control (a case where the machine should fail, run to prove the test can catch failure), and only this one has been built and run, so the formal gate cell reads partial, with the other two named as open work rather than waived.

The bill, both perspectives

Four-year win value is the currency draft analysts share, but this publication keeps its books in a second one: title equity, the change in championship probability that an improvement actually buys. The house rule for that currency is strict. Only changes get printed, never absolute title odds, and each change is measured by re-running the same simulated seasons with and without the improvement and recording the difference in percentage points. The conversion is not a constant, which is the entire point of using it. A win added to a 49-win team moves its championship odds; the same win added to a 25-win team moves almost nothing. So the machine priced every instrument's payout on the strength of the team receiving it, path by path, in the seasons the pick actually plays.

That produces two numbers, co-equal on purpose, and they are the two numbers this series has been building toward. What the package delivers to Charlotte, priced on Charlotte's simulated futures: a mean of 8.3 percentage points of cumulative title equity across the four-season windows the picks cover, call it two points of championship probability per delivered season, with a median of 3.3 and an 80 percent interval from minus 0.7 to 23.6. What the same package costs Minnesota, priced on Minnesota's own futures: a mean of 4.5 points, median 2.0, interval from minus 0.7 to 13.0.

Read those against each other, because the gap between them is the sharpest fact in the piece. Same picks, same fifty thousand futures, same wins changing hands, and Charlotte collects nearly twice the title equity Minnesota surrenders. That is not an arithmetic error. It is the tenure bet, priced. The package pays out biggest in exactly the futures where Minnesota has collapsed, and a collapsed team's wins carry no championship weight to lose, while Charlotte's crest puts it on the steep part of the title curve right as the payments arrive. Minnesota is not paying this debt out of its title account. In the futures where the bill runs highest, that account was already emptied by the collapse that made the bill run high.

The world-state table makes it concrete. Equity delivered to Charlotte runs 4.1 points in the baseline bucket, 5.9 when Edwards leaves, 10.6 when Charlotte ascends without him leaving, and 14.1 in the joint tail where both happen. Minnesota's forgone equity across the same four buckets: 4.1, 4.7, 4.3, 4.8. Nearly flat. The debt is enormous to the collector and strangely survivable to the debtor, and both facts are true for the same reason. If that asymmetry reads like a defense of the trade, one caution before it settles: title equity is the currency of contenders, and Minnesota's line is flat because so many simulated Minnesotas stopped being contenders. A bill that became affordable because the house burned down is not a bill anyone should celebrate.

For completeness, fold the seconds and the draft-night exchange back in and the all-in package delivers 10.2 points of cumulative title equity to Charlotte against 6.0 forgone by Minnesota, and running the unresolved 2030 fine print the other way moves the Charlotte number by about 0.4. Nothing above changes.

Every number above rests on assumptions, so we grabbed the load-bearing ones and shook them: fifteen stress tests declared before any of them ran, each one a full re-run of the simulation and the pricing, none of them tuned after seeing its result. The finding that leads: nothing breaks. Fourteen of the arms re-price the bill in its own currency, and across all fourteen the total stays between 4.3 and 5.2 wins against a base of 4.8. The fifteenth re-prices everything in win shares as a currency check, and the shape holds there too. No arm flips a conclusion, changes which swap matters most, or turns the debt into a bargain.

What moves the needle most is exactly what Part 1 said should: the tenure model itself. Cut every star's departure hazard in half and the bill drops about half a win; inflate it by half and the bill rises about half a win. The third-biggest lever is the one piece of live news still open when this published: if LaMelo signs his extension, the bill drops by about 0.4 wins, because a pinned-down LaMelo removes a whole family of Minnesota-collapse futures. The cheapest version of this trade is the one where the front office finishes the job. And the arm we built out of our own worst fear, the contract-labeling sensitivity from Part 1, where every long glued contract in the history gets split at its midpoint as if a hidden re-sign lurked in each one: the walk-year effect softens from 2.94 to 2.56 in the model's standardized units and the bill moves four hundredths of a win. The cliff is not a labeling artifact.

The verdict, next to the basketball

That is the bill. What it bought was measured by a different machine under a different gate: a clean-room evaluation of the on-court deal, locked in late June before any of this pricing ran, and required by its own pre-registration to decline any number its evidence could not identify. Its verdict, in its own words: "On the most complete way of measuring (defense-aware RAPM), adding LaMelo Ball changed how good the Timberwolves are by an amount that ranges from indistinguishable-from-zero to slightly negative once the near-minimum throw-in is valued reliably. On the simpler (defense-blind) read, a small improvement. The data cannot say this trade made them better, and the most complete read leans slightly negative; Minnesota spent a mountain of future and a second-apron hard cap to get there." That evaluation refused to print a title percentage because its honest band crossed zero, and this piece honors the refusal rather than converting its verdict into our currency. The two halves of the trade statement simply sit next to each other now, priced by machines that never saw each other's work: an on-court change the data cannot distinguish from zero, and the mountain of future, itemized.

So here is the trade, stated in full and without a press release in the room. Minnesota bought the best passer the franchise has ever put next to Anthony Edwards, and it agreed to owe Charlotte a debt that pays out of Minnesota's own worst futures: a mean of about five wins of draft value, a quarter of it riding on a single unprotected pick seven years away, the rest spread across three options that all get expensive in the same timelines. The debt is survivable in the futures where the basketball works and heavy in the futures where it does not, which is the defining property of every bet worth arguing about.

And the collection date is already circled. Two max contracts in one building, two walk years, one summer: 2029. Part 1 priced one of those clocks. The other one ticks at the same rate, and as of today, nobody has stopped either of them.

---

Methodology notes: swap payoffs computed per path on the FINAL 50,000-future simulation of Part 2, valued with the draft-slot value model (posterior curve draws with realized-outcome residuals, common random numbers within each path so payoff intervals carry real outcome risk); both top1-carries branches and both currencies reported in the public export; the 2029 swap priced on Charlotte-own-pick semantics, confirmed by the finalized trade language. Totals are per-path sums, never sums of marginal means. Title equity comes from a pre-built marginal-equity curve (championship probability re-simulated with and without a strength bump at four reference tiers, common random numbers pre and post, deltas only, absolute title odds gated); the per-path conversion reads the receiving team's simulated strength in each of a pick's four delivery seasons and sets its per-season contribution at one quarter of its four-year value. The 2026 draft-night exchange is priced at slot expected value by design (both slots were known at signing); the seconds carry the flat slot-31-45 curve value with the printed band. The replay and tornado beats' prose was written before their runs completed; the equity beat filled last, from total_asset_cost.json, 2026-07-13.