Every offseason a star changes teams and the same argument starts up again: does this move change who wins it all? The argument is usually loud and almost never has a number in it. This year's version is LeBron James, who, according to the box scores we pull, spent the 2026-27 preseason in a 
Philadelphia 76ers jersey after a season with the 
Lakers.
Numbers don't get loud, so we handed them the roster. The sim plays the 2026-27 season 10,000 times, each one a slightly different year with its own lucky bounces and unlucky ones, and then we count how often each team ends up in the Finals. The 76ers made it in 714 of the 10,000 seasons.
The Bottom Line
That works out to roughly 1 trip in every 14 seasons. It is a perfectly respectable place to sit if you are not trying to win the thing. LeBron's presence is worth about 1.4 points of Finals odds to Philadelphia, and the sim does not think that is enough to move them out of the middle of the pack.
How a Computer Plays a Season 10,000 Times
First the sim gives every player a score for how much the team's point total moved while he was on the floor, after adjusting for who was playing with him and against him. Then it adds those scores up for each team, weighted by how many minutes each player is expected to play. That makes one number per team, called strength.
Next comes the part that sounds like a casino. For every game left on the schedule, the sim rolls a weighted die: the stronger team gets better odds, the home team gets a small bump. It plays out the whole year, sets the playoff bracket (including the play-in games), plays that out too, and writes down who made the Finals. Then it does it again. And again. 10,000 times.
One more thing, and it matters: the sim does not pretend to know exactly how good a team is. Each season it also nudges every team's strength up or down a little, because a rating is a guess, not a fact. That is what keeps it from acting like a genius.
The Teams That Show Up the Most
Here is how often each team reached the Finals across all 10,000 seasons. The top of the list is crowded: no team gets there more than about one season in seven, which is the sim's polite way of saying nobody is a lock.
Denver Nuggets13.6% · 1,357 of 10,000
Toronto Raptors12.9% · 1,286 of 10,000
Cleveland Cavaliers12.8% · 1,279 of 10,000
San Antonio Spurs11.8% · 1,179 of 10,000
Boston Celtics11.6% · 1,162 of 10,000
Detroit Pistons10.4% · 1,040 of 10,000Los Angeles Lakers10.0% · 1,003 of 10,000Philadelphia 76ers7.1% · 714 of 10,000Philadelphia is 12th. The Lakers, the team LeBron left, are 7th at 10.0%.
So What Does LeBron Actually Add?
This is where anyone hoping for a headline gets a small disappointment. The version of the sim we are using rates LeBron at +1.2 points per 100 possessions compared to an average player. The stricter version, which only looks at what happened to the score with him on the floor, says +0.2, with a margin of error of about 1.6 either way. In plain words: close to average. The sim has no feelings about legends. It has a spreadsheet.
It also prices in missed games. He played about 2,009 minutes last season, which the sim turns into roughly 22 minutes a game over 82 games. That is the number the sim bets on.
To see what that is worth, we ran the season again with his spot filled by the kind of player any team can sign tomorrow (a replacement-level player), and used the same random rolls so the only thing that changed was him. These what-if runs used 30,000 seasons so the gaps are not just luck.
76ers with LeBron6.8%76ers with a replacement-level player instead5.4%Lakers if LeBron had stayed11.1%Lakers without him10.3%LeBron adds about 1.4 points of Finals odds and 0.8 points of title odds to Philadelphia. His move cost the Lakers about 0.9 points of Finals odds. The move shifted the odds a little and left Los Angeles still ahead of Philadelphia.
Take Away One Player: The Biggest Drops
We did the same trick for the best teams: pick a player, swap him for a replacement-level one, replay the season, and see how far the team's Finals odds fall. Below is the player each top team leans on hardest, plus the three biggest names in Philadelphia.
Nikola JokicDenver Nuggets
Victor WembanyamaSan Antonio Spurs
Kawhi LeonardToronto Raptors
Donovan MitchellCleveland Cavaliers
Derrick WhiteBoston Celtics
Tyrese MaxeyPhiladelphia 76ers
Cade CunninghamDetroit Pistons
VJ EdgecombePhiladelphia 76ers
LeBron JamesPhiladelphia 76ersGaps of a point or less are mostly dice noise. The big ones are not.
Where the Sim Is Strong
We ran this exact sim on last season, 2025-26, to see how it would have done. Its best trick is learning quickly. Each week of real games, it updates its idea of every team.
The test: how many of the 16 teams that really made the playoffs were in the sim's top 16?
Before the season even started, the plain "same teams as last year" guess beat the sim by one. Six weeks in, the sim had run away with it.
That is the whole personality of the thing. It starts out modest, keeps its eyes on the scoreboard and gets sharper every week. It is a good reader of what is happening and an average guesser of what is about to.
Where the Sim Is Weak
It cannot see injuries, trades that happen mid-year, or how old anyone is. It does not know whether players fit together or whether somebody quit on the coach in November. It only knows what the numbers from last season, and the games so far, say.
Here is what that looks like in practice. The sim's guess for Indiana and Memphis before last season was a modest plus. They both finished far worse than that. The files cannot say why, which is exactly the problem.
Indiana Pacers
Memphis Grizzlies
New York Knicks, the 2026 championsThe title is the hardest call. The Knicks won it, and the sim gave them 5.8% before the season and 6.5% with a week to go. The sim's favorite that week was Oklahoma City Thunder at 42%. A model can be right about the shape of a season and still be wrong about one team that gets hot for two months.
Two more honest limits. We have only tested this on one season, so we are not claiming it is better than the other versions we tried. And the 2026-27 rosters come from only two to five preseason games per team, so they will change.
Explained
Simulation: a computer playing out a season with weighted dice instead of real players. Do it 10,000 times and you can count how often each thing happens.
Rating (per 100 possessions): how much a team's scoring margin changes with a player on the floor, compared with an average player. Plus is good.
Replacement level: the kind of player a team can sign off the street any day. A star is measured against him.
Margin of error: how wrong a rating could be. A player rated +1.2 with a margin of error of 1.6 could really be anywhere from about -0.4 to +2.8.
Method & Sources
Data: ESPN box scores and play-by-play from the last two NBA seasons, through the same pipeline that powers STAATY's NBA pages. Player ratings come from 78,000 stretches of play with the same ten players on the floor, rebuilt from substitutions. The 2026-27 rosters are the teams players appeared for in the preseason; the sim's clock is set to the day before opening night.
The sim ran 10,000 seasons for the headline numbers and 30,000 for the what-if comparisons. This version uses a box-score-based expectation for each player. It scored best of the versions we tried on 2025-26, but the differences between versions were small. The current schedule file has 80 games per team, not 82, so a couple of games per team are missing.
The sim has been told about the move. It says it will believe it when it sees it in the standings.