The LaMelo Bet
I built a championship model to grade this trade. I wanted it to tell me we got better. It wouldn't.
I had a meeting early that morning, so naturally I was running late. I was tearing through the house grabbing everything I needed, trying to get out the door on time, and I didn't glance at my phone once. Of course I didn't.
The first text came in while I was already in the car. One of my league contacts. I figured I'd read it whenever I got where I was going, no rush. Then my old coworker buzzed through. Then Scott (my brother-in-law). Then my cousin. Then a couple of friends. Then names I hadn't seen light up my phone in months, people I hadn't talked to since who knows when.
That's when it stopped being normal. When everybody surfaces inside ten minutes, something enormous happened. I told Siri to start reading the texts out loud, both hands on the wheel, and somewhere in the middle of them my heart dropped.
Wolves get LaMelo Ball.
Wait, what... LaMelo Ball? Really? I didn't even know he was available.
And here's the honest thing. With a name like that, the first feeling isn't "we did it." It's "what did we give up." Yes, he's talent, a twenty-four-year-old All-Star, a downhill creator with handle for days, on paper the exact kind of guard the model and the coach had spent all of June asking for. But he's also a name that comes with a question stapled to it, the injuries, the defense, the fit, and for me that question landed before any celebration could. By the time I parked, Scott had already texted the exact thing I was thinking. "This doesn't pass the smell test… our franchise just mortgaged its future for a Ball brother."
It's a good thing that the meeting I was going to was therapy...
That question didn't let go of me all day. So I did the thing I built this whole series to do. I sat down and ran the trade through the model.
It wouldn't tell me we got better.
That isn't the model breaking. That's the finding. Let me show you why.
The price
Start with what we gave up, because that part doesn't need a model. The Wolves sent out Julius Randle and Naz Reid, a 2033 first with no protection on it, the right to swap firsts in 2028, 2029, and 2030, three second-rounders, and pick 28. Back came LaMelo, Josh Green, the 33rd pick, and Mouhamadou Gueye, who the Wolves are already moving to waive, plus the books to re-sign Ayo. The deal hard-caps us at the second apron, which means in the same motion it kills the big trade exception and the full mid-level. The hole behind Gobert, the one Reid used to plug, can't be patched with money this summer. Read the first piece. There is no money.
The players roughly wash on the floor. The price is the row of future picks, a 2033 first with no protection plus three swaps, and the second-apron hard cap that kills the trade exception and the full mid-level. The hole behind Gobert cannot be patched with money this summer.
So the price is a stack of future picks and a roster we can barely touch again. For that, there's exactly one acceptable answer.
Yes, we got better.
Hold onto that sentence, because the rest of this piece is the model refusing to hand it to me.
It's not a one-year shot
Here's the part the "they mortgaged everything" take skips, and it's the part that talked me off the ledge after a week of staring at the model.
We didn't spend the players. We spent the picks.
LaMelo is twenty-four. Ant is twenty-four. Jaden is twenty-five. All three are under contract through 2028-29. That is not a window cracked open for one spring and slammed shut. That is three full seasons of a twenty-four-year-old All-Star, a twenty-four-year-old top-five-ceiling superstar, and one of the best young perimeter defenders in the league, all locked in, all still on the good side of their primes. The on-court future, the part that plays inside this window, is intact and it is young.
What actually left is the draft capital that comes after the window. The 2033 first. The swaps in 2028, 2029, 2030. The start-over button.
Green spans each player's contract, all three through the 2028-29 season. The red markers are the draft capital surrendered, every one of them dated past the window it paid for. We spent the picks, not the players.
That cuts both ways and I'm not going to pretend it doesn't. A three-year window with a core this young is a real thing to bet on, and anyone telling you the Wolves punted the future is not looking at the calendar. But it's a window with no parachute. If the fit doesn't hit inside these three years, there's no pick stash waiting on the other side to retool with. We didn't bet next season. We bet the next three, and we already pushed the chips that come after them into the pot.
What LaMelo actually is
Before any numbers, just picture the player. LaMelo is a gift to score with and a problem to get a stop with. One trip he slings a pass nobody in the building saw coming and the place goes up. The next trip a guard runs him into a screen and strolls into a layup. Elite creator, leaky defender. Hold that picture, because the numbers say the exact same thing, and they tell you how much.
The number I lean on goes by RAPM, and the idea is simpler than the name. Watch the scoreboard while a guy is on the floor, do it across every lineup he plays all season, then mathematically strip out how good his teammates and his opponents were. What's left is his own fingerprint on the game, written as points per hundred possessions. Here's the yardstick that makes it mean something: average is a zero, the real stars live up around plus six to plus nine (that's the Jokic, Giannis, Wembanyama shelf), and a plus two is a good starter.
LaMelo, measured clean off his Charlotte years, comes out around plus two. A good starter. But the two halves hiding under that number are the actual story. On offense he grades about a plus four, which is genuine star-level creation. On defense he's about a minus two, a real hole that good guards go hunting for. What makes him special and what gets him picked on are both true, and they very nearly cancel.
Now the part that matters more than it sounds. Every one of those numbers got piled up on bad Charlotte teams, in games that didn't mean much, against defenses that never once had to scheme him up for a seven-game series. You can't just copy that onto a contender and call it a day. So the model does what a good scout does in his head: it marks the offense down for the stiffer competition and the smaller role he'll play next to Ant, leaves the defense where it is, and asks what's actually left when he's the third guy on a real team instead of the whole show on a bad one.
What's left is a good player. A plus one, give or take. Useful, real, and a long way from the kind of number a championship actually turns on.
Then there's the part I didn't expect to matter this much. Naz Reid.
Here's a thing a box score can't do: it can't see a smart big quietly making the other team's offense worse. Points, rebounds, blocks, those it counts. A center who's just always in the right spot, taking the shot away before it gets taken, is mostly invisible to it. So you really have two ways to count a player. The scoreboard way, the RAPM number from a second ago, catches that hidden value. The box-score way doesn't. For most guys the two land in about the same place. For Reid, they're on different planets.
Count Reid by the box score and losing him is a minor ding, about a plus one. Count him by what actually happened on the floor and it's a plus three and change, one of the single biggest hits in the whole trade. That gap, the part of Reid only one of the two counts can even see, is the argument we're about to have.
Randle, for all the noise around him, grades like a replacement-level guy either way. The production we lost in Randle was small. What we paid for him was the picks and the cap.
Net impact in points per 100 possessions, the same yardstick. Each player has two dots: what the box score counts (gray) and what the scoreboard counts (blue). For LaMelo and Randle the two nearly touch. For Reid they are on different planets, and the planet the box cannot see is the bigger one.
We didn't trade our best player. We traded the guy whose value the box score couldn't see.
The number that wouldn't come
Here's where a clean answer falls apart. Those two ways of counting, the scoreboard one and the box-score one, both pass every test I can throw at them. I checked them on old seasons they had never seen and they grade out equally well, so I genuinely cannot tell you which one is right. And on this exact trade, they disagree, by almost precisely the slice of value a box score can never see: LaMelo's defense and Reid's defense.
Then there was Gueye, the throw-in I'd spent days agonizing over, because almost all of his value is a tiny, noisy sample of defense, and depending on how kindly I read it the whole trade tipped from a hair above zero to clearly below it. While I was writing this, CBS reported the Wolves are just waiving him. Which settles the agonizing for me. The one reading that nudged this deal up to a wash needed Gueye on the floor providing that contested value, and now he won't be. Take him out, which is exactly what a waiver does, and the most complete read doesn't get to sit at zero anymore.
Put it all together and the change in our title odds runs from a hair above zero on the read that can't see defense to clearly negative on the read that can. The friendliest honest version of this trade is basically even. There's no honest version where it's an upgrade.
The trade's effect on team net rating, same points-per-100 scale, under every defensible read. The three dots per row are three honest values for Gueye, the near-minimum throw-in. A single number pulled from a cloud that straddles zero would be false precision, so the piece declines to print one.
A single number inside a band that straddles zero is false precision with a decimal point on it. So I'm not going to give you one. I know that's the number everyone wants. The honest output of a good model is sometimes a refusal, and this is one of those times. The data cannot say this trade made us better, and the most complete read leans slightly worse.
Where the bet is won or lost
The number won't come. Fine. The reasons are sharper than the number ever would have been. I ran the trade through five separate questions, ranked not by which way they point but by how much they actually move the outcome, and the shape of the bet shows up.
Ranked by how much each lever moves the outcome, not by which way it points. Bar length is that rank, not a measured percentage. The one that points up is Edwards off the ball, and it rides on a choice. The biggest of all is the one no one can call.
The biggest swing is defense, and it's the one I genuinely can't call. In the regular season LaMelo is a bottom-third defender against the pick-and-roll, and offenses have gone at him a little harder every year. Gobert hides a lot of it. When a drive reaches the rim, his contest knocks fourteen or fifteen points off the conversion, the best backstop in the league. But the playoff version, where a smart offense puts a shooter at the five, drags Gobert away from the rim, and runs everything at LaMelo with the paint wide open, has almost no data behind it. His playoff defensive sample is basically nothing. The lever that decides playoff series is the one the data flatly will not settle. We already know this scheme exists, by the way. The Spurs ran the wing version of it at Edwards for six games.
The biggest swing I can call leans the wrong way: he is hard to count on. Build his games-played out of his own six-year record and the center sits near forty-nine, against the sixty-three I committed to as a prediction, and the shape is two humps. Either the ankle holds and he plays seventy-something, or it doesn't and he plays in the forties. His 22-game and 36-game seasons aren't ancient history. They're recent. A star across fifty uncertain games is still worth real money. He's just the largest knowable risk on the board, and he points down.
The best news is real, and it comes with a condition. The loudest knock on this trade was that Edwards needs the ball, that he can't share a backcourt with another creator. The film and the numbers retire that fear. Ant is an elite catch-and-shoot threat, top-four among thirty-one high-usage stars in off-ball efficiency, and he's genuinely better off the ball than on it. He can play next to LaMelo. Whether he will, whether he actually gives up the on-ball reps so LaMelo can run the thing, is a choice, not a measurement. The entire upside rides on a choice.
Depth leans negative. Behind Gobert now is Joan Beringer, nineteen years old. With Reid and Randle gone, the frontcourt leans on a teenager taking a real step, and the only help left is a minimum or two. Cheap tools patch a hole. They don't fix it.
The collision everyone argued about matters least. Every comparable two-creator backcourt, the ones that worked and the ones that flopped, ran a perfectly fine offense with both guys on the floor. The thing the discourse was loudest about turned out to move the outcome the least, and it quietly handed the decisive weight back to the defense.
What Connelly actually bought
Run the season twenty thousand times and the middle of the pile is a wash. But the tails aren't the symmetric thing we want them to be. If the fit clicks, if Ant cedes the reps and the ankle holds and the defense survives, we get a real bump in deep-playoff odds, somewhere between three and a half and nearly eight points of added probability that we reach the conference finals. If it fails, the collision, the hunted defense, the missed games, the most complete read says we come out worse than the team we traded away.
Each bar is a slice of 20,000 simulated seasons, sorted by how good the team turns out (net rating). The middle is a wash. The left tail comes out worse than before, the green right tail reaches the conference finals. A bet on the shape of a curve, not a guarantee.
He bought the right tail of a distribution whose middle is a coin flip and whose left tail is a genuine downgrade. That's the trade. Not a disaster. Not an upgrade. A bet on the shape of a curve.
And we've seen a version of this movie. The last time we bet big on a fit-dependent star, the Gobert deal in 2022, year one was a mess before it ever became a strength. I spent a whole piece on what that misuse actually cost and where the blame really sits. That history isn't proof of anything here. It's a warning, from our own recent past, about exactly this kind of swing.
What this leaves us with
I'm not going to bury the prediction inside the hedge, so here it is, logged cold for the season to grade. The model itself sees a mid-to-high-forties-win team, not the contender the headline wants. I committed to LaMelo playing around sixty-three games, and I'll tell you plainly that the model leans lower than that and I bet the optimistic side anyway. And the one that settles the argument the data couldn't: the Edwards-and-LaMelo two-man net rating. Near plus 8 and the fit clicked and this was worth it. Near minus 2 and it didn't and it wasn't. Reality closes the loop the model left open. Grade me on it.
Green is the logged prediction, point and band. On the win total, the gray block is what the model itself projects, lower than the bet. The bottom row is the one that settles the argument: near plus 8 and the fit clicked, near minus 2 and it didn't.
Here's where I land, and you can hold me to this too.
The ceiling is real. This was not a stupid trade. It was an expensive, high-variance one, and the model's refusal to crown it isn't a verdict of failure, it's an admission that this particular bet cannot be graded before it's played. The likeliest single outcome is that it lands somewhere near even.
But hear the thing I keep coming back to. This isn't one season. It's three. Three years of a twenty-four-year-old All-Star next to a twenty-four-year-old superstar, three cracks at Ant ceding the reps and the ankle holding and the kid behind Gobert taking the leap. A one-year dice roll, you grade by the spring. A three-year window, you give a chance to become what it's capable of. We bought the harder thing to judge, and we bought more than one roll of it.
So watch the things that actually decide it. Watch the first time a playoff offense pulls Gobert to the perimeter and runs at LaMelo. Watch whether Ant gives up the ball. Watch the ankle.
The model says probably not. But if the looks open up and the ankle holds and the kid leaps, that's the season that proves it wrong. And anyone who's read me this far knows I'm an optimist. I lost hope on this team once already, in a walking boot and a Game 2 that was over by halftime, and I hated every minute of it.
That's the bet. And I'll be rooting for it like hell.
One note on timing. I'm writing this in late June, with the trade just official and the offseason still going. The grade on the trade itself, the swap, the math, the bet, is settled and isn't moving. The roster around it is. The Wolves still have cheap money and open spots to fill, and the right minimum or mid-level add could take some of the sting out of the depth problem I just spent a section on. So read that depth worry as a snapshot of an unfinished roster, not a final word on it. Everything upstream of it holds either way.
Corrections
No corrections logged.