Methods

The model compares three scenarios for one team at a time. They are what the team does today, the league's paid streaming hub with no free games, and the hub with some games kept free over the air. It starts from facts about the team's market and schedule and from a set of assumptions you can change. For each scenario it estimates how many households are reached, how many of those are headed by someone under 35, a range for rights revenue, and a friction score for one example household. It also works out whether each county can watch free, has to pay or has no way to watch. The map, the memo and the case study all run the same model, so their numbers always match.

Every sourced number has a letter grade next to it. D is documented, which means I read it from a dated source. I is inferred, which means it is calculated from documented numbers by a rule this page states. S is speculative, which means it is my judgment and comes with a range. Numbers the model calculates do not get a grade, because they depend on the assumptions. The model version at the top of each page says which version produced them.

A team's market is the list of counties where its local games are sold today. For each county, the household count, the householder's age and broadband subscriptions come from the Census Bureau's American Community Survey 2020 to 2024 five-year estimates (tables B11001, B25007 and B28002). The data files keep the margin of error for every household estimate.

DACS 2020–2024 5-year, B11001

Each territory has a grade for how well its county list is sourced. A means a dated primary source names the area and the county list follows directly from it. B means a dated primary source names the area but not the counties, so I inferred the list by checking zip codes on the seller's own page or by using the television market. C means I inferred the list and could not check it against a source, so it should be treated as a working definition.

I count a county as covered when its center of population, from the 2020 Census, falls inside the FCC service contour of the team's station or of a station that repeats it. Covered means a household there could pick up the games with an antenna. It does not mean the household owns one.

Utah Jazz. KJZZ 14 and 43 translators cover 23 of 60 counties, 1,102,139 households, which is 63.9% of the market.

Minnesota Timberwolves. KARE 11 and 7 translators cover 31 of 398 counties, 1,674,407 households, which is 32.7% of the market, for the 15 games KARE 11 carries.

Boston Celtics. No station carries the games over the air. They are on NBC Sports Boston on Peacock, which is paid, so no county is covered and no local game is free today.

Phoenix Suns. Arizona's Family and 16 other stations cover 10 of 15 counties, 2,758,723 households, which is 96.6% of the market.

New York Knicks. No station carries the games over the air. They are on MSG+ on DAZN, which is paid, so no county is covered and no local game is free today.

DFCC TV Query and FCC TV service contour files

This method has a few weaknesses. One point stands in for a whole county, so a county whose center of population is just outside a contour counts as uncovered even if some of its towns get the signal. The FCC does not record which station a translator repeats, so I took that from RabbitEars and checked it against each FCC license. A license also does not prove that a translator is on the air. And the contour files do not state the field strength they were drawn at.

In each scenario, every county gets one of three kinds of access, and this is what the map draws.

Nobody has published where the hub would be sold. I assumed it would be sold in the same territory as the team's stream today, which means no county is unavailable under the hub. In the hybrid, a covered county is free only on the nights a game is simulcast, and the map shows that by filling in part of the circle.

Reach has a free part and a paid part, and the two are shown separately as well as added together. The free part is every household in a covered county, for the games that air over the air. It does not depend on any assumption. The paid part starts with the households that could pay, which are the ones with a broadband subscription and no free way to watch in a county where the stream is sold. The households that would pay are that group multiplied by the sign-up rate, which is an assumption. Households reached is the free part plus the households that would pay.

Households with no way to watch are in a county with no over-the-air signal and have no broadband subscription at home. That count comes straight from the FCC contours and the Census, so none of the assumptions change it. In the hybrid, every figure is an average across the 82 games, because the free part only applies on simulcast nights.

When a team's games are on a regional sports network today, as the Celtics' and the Knicks' are, the model treats that network as a paid stream. No local game is free, and the households that would pay are counted at the same sign-up rate. I do not count cable subscribers who get the network in a bundle separately, because no public source separates them from streaming subscribers. MSG's annual report only gives a combined figure for its two networks DSphere Entertainment 10-K, Nov 2025.

Under-35 reach applies the same rules to households headed by someone under 35 and then multiplies the result by a number that the fan survey will set. The survey is a convenience sample posted to a fan community, so it is not a random sample. Once it has run, this page will say how many people answered and when. Until then, the multiplier stays at 1.0× and keeps its S grade.

The survey only changes the multiplier if at least 150 adults answer, including at least 30 on each side of 35. In that case the multiplier is the share of under-35s who would pay for the hub divided by the share of everyone 35 and over who would, rounded to the nearest step on the slider and kept between 0.6× and 1.2×. Someone counts as willing to pay if they chose $15 a month or more, which is the lowest answer on the form that covers the hub's price. The hub's price is $124.99 a season, which works out to $10.42 a month over a year.

The sign-up rate is the share of households with no free way to watch that would pay for the stream. No hub exists yet, so this is a judgment and is graded S. The slider starts at 6% and runs from 2% to 40%. I based that range on three documented numbers.

Revenue is always shown as a range. The size of the hub's rights pool depends on how many teams join. The reporting gives two points, $850M a year for 25 teams or fewer and $1.2B for about 29, and I drew a straight line between them DSBJ via MarketScale, Aug 17, 2026. Those figures are what the league would reportedly seek and where an offer could land. Neither is a signed deal.

Sports Business Journal reports that the money would be split by market value but does not say how that is measured. So I calculated each team's share two ways, once by market size and once as an equal split, and the range runs between them. Market size is the number of households in each team's metro area, from the Census. That undercounts a team whose market covers a whole state, and two teams in one metro area split its households.

I assumed teams join in groups. The 13 former Main Street Sports Group teams (the old Bally regional networks) go first, then the 4 NBC regional network teams, the 5 independents and the 5 teams with no reported position, and the 2 reported holdouts last DSBJ via MarketScale, Aug 17, 2026. The reporting points to the first 22. Nothing has been reported about the order of the other 5, so I added them alphabetically. At the default of 25 teams, that puts the Nets, the Bulls and the Nuggets in the hub and leaves the Rockets and the Wizards out. A team's own figures always assume that team is in the hub.

DACS 2020–2024 5-year, B11001, metro areas ICountyline method

In the hybrid, the team gets its hub revenue minus the pay cut for the simulcast. It also gets a share of over-the-air ad revenue for the games it keeps on a station, once someone enters that figure. If no games are kept free, the hybrid is the same as the hub and there is no pay cut. Revenue today is the team's stream subscriptions multiplied by the net price, plus over-the-air ad revenue. The net price is the stream's yearly price minus platform costs. I could not find a public ad revenue figure for any team, so that number stays blank until someone enters one, and today's revenue shows as Blank until then.

Friction measures how much it takes for one household to watch every game. The score adds 10 points for each service after the first, a tenth of a point for each dollar the household spends in a year, 2 for each set-up step and 5 for each game that is not on the household's main service. The model tries every combination of services that covers all 82 games and uses the one with the lowest score, so adding a free way to watch can never make the score worse. Each service counts as two steps, one to get it and one to set it up on the TV. When paying by the month for only the months with games is cheaper than a full year, the model uses the monthly price.

If the only service that carries a household's games has no public price, that household gets no score and the screen says why. A household that cannot watch every game with any combination of services has no score either.

Blank means the service has no such price or has not published one, and hovering over the word says which. The friction score and the calendar are for one household at a time, and each team has its own list of example households.

There are nine assumptions. Seven have a slider, and the other two are fixed ranges for now. The table shows where each one starts, how far it can move and where it comes from, and the notes under the table say what is weak about each one. Anything you change is saved in the link.

Households that sign up to pay. No hub exists yet, so there is no real sign-up rate to use. The three comparisons are a team stream whose games are also free over the air, one baseball team's stream, and cable-bundle access from before 2023.

Games kept free over the air. The free games are spread evenly through the season, which is not how a real broadcast schedule would look. Each team's slider stops at its number of local games, meaning the games that are not exclusive to a national network.

Hub price, per season. No price has been announced for the hub. The slider starts at what Jazz+ charges today.

Teams in the hub. Both reported figures come from Sports Business Journal by way of another outlet's summary, and the straight line between them is my guess.

Under-35s vs. everyone else. This depends on the fan survey, which has not run. Until it does, households headed by someone under 35 are assumed to sign up at the same rate as everyone else.

Hub pay cut for the simulcast. No public deal says how much hub money a team would give up to keep games on a local station. The range is my judgment.

Over-the-air ad revenue. I could not find a public figure for any team. One outside estimate exists for the Jazz, and I did not use it.

Team stream subscriptions. The range is based on one public count from Mar 11, 2024. The top of the range is my judgment.

Platform costs. This is my judgment. Streaming vendors do not publish what they charge.

The two fixed ranges are stream subscriptions, at 20,000 to 60,000 a season, and platform costs, at 15% to 30% of the stream's price. The hub's season price starts at what Jazz+ charges today.

This table has every sourced number the model or a page uses, with its grade, its source and the date I read it. Where the value is blank, I looked for the number and could not find it, and the note says where I looked.

The hub is reported to start in 2027-28. That schedule does not exist yet, so the model uses the 2026-27 schedule and prices in its place. The league scheduled 80 of each team's 82 games in August 2026 and will set the other 2 after the NBA Cup group stage. I count those 2 as local games in early December and do not make up dates or opponents for them.

A game on ABC, ESPN, NBC, NBCSN, Peacock or Prime Video is exclusive to that network, so the team's local outlet does not carry it. NBC games also stream on Peacock, and NBCSN games are Peacock exclusives that also air on cable, according to NBC's August 2026 release. I count an NBA TV game as also airing locally unless a source says otherwise, because the league's NBA TV page allows local broadcasts. The Jazz have 2 of these games, the Timberwolves 6, the Celtics 4, the Suns 5 and the Knicks 8. Those for the Suns are documented game by game on NBA.com. The schedules come from ESPN's public schedule data, which I checked against the league's and the networks' releases. Each team's data file lists its sources and notes what I could and could not confirm.

The Jazz market is a working list of 60 counties. No published source lists them, and I left out the parts of Montana, Oregon and Washington that were added to the Jazz+ territory in October 2025. The other teams' territories are graded in the markets section, where a C means the list is a working definition.

Translators in Emery, Grand and Millard counties may carry KJZZ without being credited to it. That puts real Utah coverage somewhere between the 97.6% this method gives and the whole state, which is what the Jazz have claimed. Juab County's center of population is right on the edge of the Nephi translator's contour, and I count it as uncovered.

The hub's pool, its team count and the two public rights fees all come from Sports Business Journal reporting that I read in other outlets' summaries, not in the original articles. The DAZN per-team offer is also secondhand, from spring 2026, and I do not have its exact date. The Padres subscription count comes through Sports Media Watch, and I did not save the link.

When a service has not published a 2026-27 price, its price is blank, and so is the friction score for any household that needs it. NBC's late Tuesday games air over the air only on Mountain and Pacific stations, but the model counts every NBC game as over the air for every team. The NBC Sports Boston add-on requires a Peacock plan, and in a month with no national game on Peacock the model only counts the price of the add-on. KARE 11 announced 15 Wolves games, and every schedule source puts five of them on national telecasts, so the Wolves' free nights may be overstated by as many as four. The fan survey has not run, so the under-35 multiplier is 1.0× by assumption. None of this is a prediction of what any team will decide.

Model v0.3 · Oct 7, 2026. A household's calendar now counts every game it can get with an antenna as free, including the national games on NBC and ABC. Those games used to be marked as national exclusives, so an antenna household's calendar showed fewer free games than the count above it. The hybrid can only keep as many games free as a station could carry, which is the number of games that are not exclusive to a national network. The slider now stops at that number, and an older link that asks for more is capped at it. Before this, reach kept rising past that point. The would-pay figure in the map key now counts only the counties in its stream-only row. For a team with some free nights, it used to count the whole market. When a station carries only some of a team's games, the sentence at the top of the results now counts the rest of the local games as the stream's. It used to count every other game of the season, including national games. When no games are kept free, the hybrid now matches the hub on every figure, so there is no pay cut.

Model v0.2 · Oct 6, 2026. A household whose only way to watch has no public price now gets no friction score. It used to get a score that treated the stream as free. A team with no station has nothing to simulcast, so its hybrid is the same as the hub, with no free nights, no pay cut and no over-the-air ad revenue. A station that carries only some games now counts as free on those nights only, averaged over the season the same way the hybrid is. If a stream has no public subscriber count, today's revenue is left blank. It used to borrow the Jazz+ range. In the hybrid, households with no way to watch are now averaged over its nights, like its other figures. Added the Timberwolves, Celtics, Suns and Knicks alongside the Jazz.

Model v0.1 · Oct 5, 2026. The first version of the model. It covers reach as a free part and a paid part, the access each county gets, a hub pool split that runs from market size to equal shares, a friction score based on the lowest-scoring set of services that covers every game, and findings written from templates.

This product uses the Census Bureau Data API but is not endorsed or certified by the Census Bureau. The data pipeline reads the ACS table files from census.gov, which do not need a key. On every build it also checks each variable's label against the Census API. If the Census moves a variable, the build fails, so a number cannot change without anyone noticing.

Over-the-air contours come from the FCC's TV Query and its TV service contour files. Station licenses, communities and facility IDs are also from the FCC. I use RabbitEars.info only to identify which station a translator repeats.

Team and partner logos are here to identify who the analysis is about, and they belong to their owners. Countyline is independent. The about page has the full disclosure.