Twenty-Eight Teams Changed Their Opening Five: In the NBA, Continuity Is Now the Exception — and the Apron Is Footing the Bill

There is a folder on my laptop called "Game-1 sheets." When I started Hoopistani in 2026 I decided the team sheet was more honest than the box score: the box score tells you what happened, the sheet tells you who was given the chance. Two years later, at the FIFA U-17 World Cup in Delhi, I built a 504-player ledger of club affiliations and registration status. That exercise turned into a habit, and the habit comes back every opening night.

This season it came back with thirty sheets. Twenty-eight of them carried at least one name in the opening five that had not been on the previous season's first-night sheet. A few teams changed four names — four-fifths of the five, roughly half of what we loosely call the core. Two teams changed nothing: Denver and Orlando.

And in San Francisco, one name was written in the same place for the eighteenth time.

Five hundred and four names in, I stopped believing the fees. Thirty sheets in, I stopped believing in continuity. What is left is machinery — the cap sheet, the apron line, the trade-matching rule, and the arithmetic that decides what it actually costs to keep the same five together for two seasons.

The numbers first, because everything else leans on them. Twenty-eight of thirty teams changed at least one starter. Of last season's opening-night starters, seventy-one started again for the same team and twenty-one changed uniforms — roughly a quarter of the league's starters moved in a single offseason. Only three players have started for the same team on opening night in each of the last ten seasons: Stephen Curry, Nikola Jokić and Devin Booker.

This is where I have to stop and write down what I do not have. There is not a single offensive rating, defensive rating, net rating or pace figure in the dataset. No team's salary sheet, no contract length, no distance from the apron line. When I opened my own file, dozens of columns were empty. Claiming to fill them is the mistake every season-preview feature makes: mistaking a count for a conclusion.

So this report splits in two. One half is what the sheet proves — the breadth of churn, its direction, and its three or four clean exceptions. The other half is what I can only model, and what has to be printed as modelled.

The context is on paper, not on the floor. The 2026 collective bargaining agreement introduced a second apron, roughly $17.5 million above the tax line and rising about ten percent a year. The penalties above that line are widely reported and I am taking them from CBA summaries, not from any team's cap sheet: no aggregating salaries in a trade, no sending cash, no taking back more salary than you send, no access to the taxpayer mid-level, and a first-round pick frozen at No. 30 if a team stays over in three of five seasons. Trade matching is tiered too — below the first apron a team can take back about 175 percent of outgoing salary plus $100,000; above it, 125 percent.

Think about what 125 percent does. You cannot send one large contract and receive one large contract. You have to stack three or four small ones — and above the second apron, stacking is banned outright. Improvement therefore arrives through minimum deals and exceptions, at the back of the rotation, which means the opening five changes as a by-product of cap mechanics rather than basketball judgement.

I priced a World Cup from a bedroom once and the bedroom held up, but that model had FIFA's prize pool and the Club Benefits Programme in it. This one does not have the inputs. So the model goes in bands, not points.

First layer: a team that wants to hold its top eight or nine together for three seasons watches its salary base rise every year through built-in raises. Second layer: as the base rises, the team drifts toward the apron, and crossing it removes the aggregation tool. Third layer: without aggregation, an in-season injury at one position cannot be fixed by the easiest route. Stack those three and you get what I call the continuity premium. I will not put a single number on it, because I do not have a single cap sheet. What I can say is that the direction is one-way: the longer you hold a core, the more expensive your ability to repair it becomes. A rational front office reads that and answers the same way every time — keep the core, churn the edges.

That is where four new starters stop looking like chaos. It means a team changed its four cheapest slots — minimum contracts, two-way players, the mid-level — while its two most expensive contracts stayed exactly where they were. In the statistics it reads as churn. On the cap sheet it is one column moving.

On the floor, the cleanest line I found was that a starting five shares the court for roughly five of forty-eight minutes. The other forty-three get assembled by matchup, foul trouble, fatigue and scoreboard. The claim is technically correct and statistically demonstrable, but its real force is managerial rather than tactical. Publicly devaluing the starting lineup buys a coach the two permissions he most needs: to bench a veteran, and to start a young player for a matchup. The moment "start" becomes ceremony, the argument around it dies.

There is a second consequence. If games are decided in the last five minutes, then starts are the wrong proxy for a bench player's role. A player who started twelve of eighty-two games but closed every one of them has a role far larger than twelve — and we have no public instrument that measures it.

And there is the strangest datapoint of all: nineteen seasons, nineteen different opening fives, offered up with a shrug and the word "weird." That shrug matters. In a culture where rotation volatility is normal, churn is not news. In a culture where it is abnormal, every change is a headline.

The two exceptions deserve care, because exceptions define the rule. Denver's continuity is built around a Jokić-centred system in which the edges rotate and dissolve into the whole. Orlando's continuity is a young, cost-controlled core that has not yet met its first real contract squeeze. One team has bought its continuity; the other has not yet been billed for it. Putting them in the same column is a category error.

What gets lost in all of this is that churn is not new. The Clippers have not repeated an opening five since the mid-1970s. Toronto and Sacramento have never done it. That is a forty-year record, not the product of a recent rule change.

What is new is that churn is now cheaper than continuity. Instability used to be an illness. It is now a strategy. You treat an illness; you budget for a strategy. And the moment a front office puts churn in the budget, our language has to change — from "this team fell apart" to "this team is paying its flexibility bill."

Here I have to point at the weakest joint in the whole structure. The material claims that recent champions were among the heaviest-churn teams, which would mean continuity is not a prerequisite for a title. If true, that is the strongest wall in the argument. It is also the least supported. It references an Oklahoma City title and "New York this past season." I went back to my file. I have no primary record that verifies the second claim.

There is a timeline problem too. If Curry is in Year 18, and if an Orlando starter is in his second straight season in that role, the whole frame points forward rather than back — meaning the title being referenced belongs to a season I have no document for. That sounds small. It is not. If the champions-churn claim collapses, the piece loses its tallest storey and becomes a descriptive statistics report. The counts — 28 of 30, 71 versus 21, nineteen opening fives in nineteen seasons — are solid and repeatable. The champion story is the part that will travel furthest and be checked least, and in my trade that combination is the dangerous one.

What I could not find anywhere in the material is who plays the last five minutes. Nobody was selling me a closing-five index, so I am going to build one, publicly, so that someone can prove me wrong. The method is simple: for every team, every season, list the five players who logged the most minutes with the score within five points in the final five minutes. Then measure the overlap between two seasons. A team with a hundred percent turnover in its opening five and eighty percent continuity in its closing five is the real exception in this churn story, and right now we cannot see it.

I have no evidence that closing-five continuity correlates with winning, and I am not claiming it does. I am claiming that if we accept the last five minutes decide games, we should be measuring those five minutes — and we are not.

Then there is the part that never makes the dataset. On opening night a twelfth man's name appears where a starter's was, a phone rings in a house somewhere, a coach who trained that kid for free three years ago finds out he is in an NBA starting five. you had to be there — because behind twenty-eight teams making four changes each there are a dozen phones ringing, and a few phones that did not ring at all. The player who lost the spot had an agent who took no calls that week. My job is to remember that churn is a statistical event and a household event at the same time.

The stadium empties first; the wage bill notices eighteen months later. Here the order runs backwards — the bill arrives first and the empty seats follow. A team that keeps rewriting its depth rewrites its audience's habits too. The jersey someone bought last year is being worn in another city this year. I have no sales report and I will not invent a number. I will leave it as a question: if a league turns over forty percent of its depth every year, who exactly is the fan attaching to?

If apron mechanics are the engine, then the most useful consequence is that rule changes are a leading indicator of next season's opening fives. Tighten the apron and churn accelerates. Loosen it and a few teams will risk holding a core. That is the signal nobody is watching, because it is written on the CBA pages rather than at the trade deadline.

Three things I will watch this season. One, the list of teams sitting within ten million dollars of the second apron at the deadline — that is where the real drama is. Two, the closing fives of Denver and Orlando, not just their opening fives, because if the two match, we have our first real data point. Three, the first time a coach has to admit he benched a veteran purely for matchup reasons — that is the day we learn whether "starts don't matter" is a doctrine or an alibi.

I have thirty sheets. Twenty-eight of them contain a change. The two that do not are not the most interesting teams this season. The interesting ones are the teams that changed four and still called themselves stable — because they had no other way to say it.