Investing Unpacked · No. 07 · Recoupment math, not gross
This week’s Pulse found $14 of the $17 that Halloween took off every seat back on it: revenue per available seat went from $103.10 to $117.00, the houses from 87.8% to 92.1% full. The 31 shows at the same number of performances grossed 14% more, with price 59% of the gain: their average rose 8%, attendance 6%, and 27 of 31 took a higher average. Capacity and the average moved together this week; the tell is the week they move apart. The table shows the average moving. It cannot show why. A discount lowers it, a cheaper mix of seats lowers it, and so does a seat given away, because the League has counted complimentary tickets in attendance since 2009. This issue is about the case the table hides best: a house that stays full while the gross falls, and what that does to the weeks a show needs to pay its investors back.
What the table counts
Until the spring of 2009 the Broadway League published paid attendance. From the week ending May 31, 2009 it has published total attendance, every seat with a person in it, including the thousands given away each year to awards voters, the industry and the press. The same release moved the gross from net to gross gross, with the credit-card and group commissions left in. The League's own definitions are short: attendance represents paid and complimentary tickets; average paid admission is total gross divided by total attendance; capacity is attendance divided by the seats on sale.
So a complimentary ticket does two things to the public figures. It raises capacity by one seat, and it lowers the average by counting as a zero. It does nothing to the gross, and in our model it does nothing to the costs either: the running costs are charged by the week, on every seat in the house, sold, given away or empty. The one public figure it cannot move is revenue per available seat, which is the gross divided by the seats on sale, or the average times capacity, the two effects cancelling.
Average paid admission = gross ÷ total attendance (a comp counts as $0)
Revenue per available seat = average paid admission × % capacity
Three weeks of the house musical make the point. In week A, 80% of the house pays budgeted prices and the rest is empty. In week B, the same 80% pay and ten points of the house are complimentary. In week C, every seat in a 90% house is sold, at 11.1% off. On the public table week A is the one that looks empty, and week B looks exactly like week C. On the P&L the three are the same week.
Three weeks, one P&L · the house musical, $1.25M a week when full at budgeted prices
| A · empty | B · comped | C · 11.1% off | |
|---|---|---|---|
| Paid seats | 80% | 80% | 90% |
| Capacity | 80% | 90% | 90% |
| Average paid admission | $130.21 | $115.74 | $115.74 |
| Gross | $1.00M | $1.00M | $1.00M |
| Per available seat | $104.17 | $104.17 | $104.17 |
| Investors' week | $132,705 | $132,705 | $132,705 |
| Weeks to recoup | 113.0 | 113.0 | 113.0 |
Revenue per available seat is $104.17 in all three. That is the number the Pulse ranks by, and this is why: it is the only public figure that does not care how the seats were filled.
What a point of paper costs, in weeks
Weeks to recoup is the capitalization divided by the weekly profit that pays it back, and the weekly profit is what is left of the gross after fees, the theatre, the running costs and the royalty pool. A seat given away adds nothing to the gross, so each point of the house that is complimentary rather than paid comes straight off the profit, and off the profit it comes with the operating leverage this series keeps finding: the costs do not move, so the profit moves more than the gross.
A 90%-full house, with 0 to 20 points of it complimentary · the house musical
| Comped | Avg. paid | Investors' week | Weeks | Added |
|---|---|---|---|---|
| 0 points | $130.21 | $196,168 | 76.5 | — |
| 5 points | $122.97 | $164,437 | 91.2 | +15 |
| 10 points | $115.74 | $132,705 | 113.0 | +37 |
| 15 points | $108.51 | $100,973 | 148.6 | +72 |
| 20 points | $101.27 | $69,242 | 216.6 | +140 |
Comped points come out of the paid share: 10 points comped is 80% of the house paid. Weeks to recoup at the paid share, and the weeks added against the all-paid week.
76.5 → 113
Weeks to recoup · ten points of a 90% house comped
+37 weeks
Added to recoupment · capacity unchanged at 90%
$104
Per available seat · the same in every version of the week
Ten points of paper in a 90% house take the investors' week from $196,168 to $132,705, a third of it, and the weeks to recoup from 76.5 to 113.0. The capacity figure does not move. The average does, by 11.1%, which is the only trace the table keeps. And the cost rises with every point: the first point of the house given away adds 2.6 weeks, the tenth adds 5.2, the twentieth adds 18.2, because each one lands on a thinner margin than the last. At twenty points the show is 70% paid, and recoupment takes nearly three times as long.
Why a producer fills the house anyway
Some of the paper is in the contract. Under the Tony rules, every producer of an eligible production must invite each eligible voter, about 830 people, free of charge, and keep at least 32 paid performances open to them; the press and the industry come on top. Those seats are a cost of the season, and in the model they are what every comp is: attendance that earns nothing.
The rest is a choice. Papering services hand complimentary tickets to members who pay a membership fee and a few dollars a ticket, and nothing to the production; their busiest weeks are press performances and Tony voting time; their members are told to keep the listings private. The reason is the one the services themselves give: a full house plays better, and the people in it talk. The reason the listings stay private is, we think, the other half of the trade: a seat given away is a seat the market now knows was for sale.
The cleaner version of the same decision is preview pricing. Shucked opened in April 2023; its first three preview weeks played to 99%, 100% and 98% capacity, with orchestra seats from $69 and mezzanine seats from $29 and without papering; its producer said the show had capitalized an extra million dollars of preview losses it would not have carried with a star in the cast, and called it a marketing expense. In the model the trade reads the same way: a house that is 100% full at 30% off earns the week of a house 70% full at budgeted prices, and at 40% off it earns the week of a house 60% full, which is a loss. The difference between a producer and a problem is whether the shortfall was budgeted and the capitalization carries it.
The tell, on this week's table
The public table cannot show comps, but it shows attendance and gross week over week, and the two together imply what happened to the average. When attendance holds and the gross falls, the average fell: by discounts, by a cheaper mix of seats, or by seats given away, and the table cannot say which. That divergence is the tell. The street showed it as a whole in the week ending October 18, when attendance rose 4% and the gross fell 4% because the holiday price came off, and it shows it one show at a time every week.
Implied change in the average = (1 + gross change) ÷ (1 + attendance change) − 1
Attendance flat and gross down: the average fell, by discount, mix or paper
Attendance against gross · the 32 shows running in the weeks ending October 25 and November 1, 2026
| Show | Attend. | Gross | Avg. | Per seat |
|---|---|---|---|---|
| The street · 32 shows running both weeks | −5% | −14% | −10% | $102 |
| Schmigadoon! | 0% | −27% | −27% | $82 |
| The Fantasticks (4 → 7 perfs) | +63% | +20% | −26% | $75 |
| & Juliet | −14% | −35% | −24% | $74 |
| The Imaginary Invalid | +9% | −13% | −21% | $62 |
| Just in Time † | −12% | −30% | −21% | $87 |
| The Outsiders | −19% | −34% | −19% | $74 |
| Hadestown | −5% | −22% | −18% | $103 |
| Operation Mincemeat † | −17% | −32% | −18% | $87 |
| Harry Potter and the Cursed Child | −1% | −18% | −17% | $126 |
| Buena Vista Social Club † | −8% | −24% | −17% | $81 |
| Downstate (6 → 8 perfs) | +23% | +4% | −15% | $46 |
| SIX | −18% | −29% | −14% | $54 |
| The Lost Boys † | 0% | −14% | −14% | $91 |
| Oh, Mary! | −1% | −14% | −14% | $183 |
| The Lion King | −3% | −16% | −13% | $116 |
| Wicked | −3% | −15% | −12% | $74 |
| The Book of Mormon | −11% | −21% | −12% | $85 |
| Two Strangers † | −24% | −33% | −11% | $47 |
| Aladdin | −7% | −17% | −11% | $64 |
| The Great Gatsby | −14% | −23% | −10% | $72 |
| Stranger Things | −2% | −11% | −10% | $55 |
| Maybe Happy Ending † | −11% | −18% | −8% | $79 |
| Hamilton | −1% | −8% | −7% | $219 |
| School Girls | −2% | −9% | −7% | $73 |
| A Few Good Men † | 0% | −7% | −7% | $145 |
| Wanted | −13% | −18% | −7% | $61 |
| MJ | −2% | −7% | −5% | $146 |
| Chicago | −19% | −23% | −5% | $91 |
| Paranormal Activity † | 0% | +2% | +2% | $116 |
| Other Desert Cities † | 0% | +4% | +4% | $301 |
| The Rocky Horror Show | +2% | +9% | +7% | $122 |
| 860 | +1% | +12% | +11% | $151 |
† OFFSCRIPT is an investor or producer. Attendance and gross are week-over-week changes; avg. is the implied change in the average paid admission; per seat is the week's gross ÷ seats on sale. The Fantasticks and Downstate added performances.
Thirty of the 32 shows played the same number of performances both weeks. Twelve of the 30 held attendance within two points of the week before, and eight of the twelve took an average at least 5% lower, by 5% to 27%: the same seats, fewer dollars. The other four held attendance and took a higher average, which is the opposite tell, and the better one. Across all 32, 28 took a lower average and 24 of those also sold fewer seats; the Halloween week emptied the table from the bottom, as the Pulse found, and took price off the part that stayed full. Harry Potter and the Cursed Child is the clean case the Pulse already named, as an inference from the table: attendance down 1%, the average down 17%, so its whole decline was price. The table says that much and no more. Whether a show's lower average was a sale, a cheaper mix of seats or a block of complimentary tickets is a question for the weekly statement its investors receive, not for the public table, and it is the first thing to ask.
What to ask
01
Paid attendance, every week.
The public figure stopped separating paid from complimentary in 2009. The weekly statement should not. Read capacity on paid seats, and read the comps as their own line.
02
The comp budget, by category.
Voters, press, the industry and house seats are a cost of the season; papering is a decision. Ask how many seats each takes, who approves them, and what the offering budget assumed.
03
The preview plan.
If the house will be filled at a low average on purpose, ask what the plan is budgeted to cost and whether the capitalization carries it. A producer who can answer in dollars has made a decision; one who cannot has made a hope.
04
The average net of comps.
Gross divided by paid attendance, against the average the offering assumed. It is the price the market is actually paying, and the public average understates it by exactly the complimentary share of the attendance.
05
Weeks to recoup, on paid seats.
Rerun the recoupment schedule on paid capacity, not on capacity. In the model, ten points of paper in a 90% house turn 76 weeks into 113.
Watch for it
01
Seats come back before price.
Halloween took the street to 87.8% capacity and a $117.49 average. We expect the week ending November 15 to be fuller than that while the average is still below the October 25 week's $129.18. Wrong if capacity is at or under 87.8%, or the average is back at $129.18 or more. Settles on that table.
02
The opening lifts the average.
Wanted opens November 8. Its last full preview week averaged $74.31 a ticket at 81.6% capacity. We expect its average paid admission in the week ending November 15, its first full week after opening, to be above $74.31. Wrong if it is not. Settles on that table.
03
Fewer shows show the tell.
Eight shows held attendance within two points and took an average at least 5% lower in the Halloween week. We expect fewer than eight in the week ending November 15. Wrong at eight or more. Scored by count; the shows are not named.
Notes
Illustrative figures for a hypothetical show, not any specific production: the musical this series runs its numbers on (Nos. 03 to 06 use it too), in a 1,200-seat house playing eight performances a week, grossing $1.25 million a week when full at budgeted prices (a $130.21 average); 9.5% of gross deducted to reach net receipts; the theatre paid 6.5% of net receipts plus a $15,000 flat rent; $625,000 of weekly running costs, the rent and house charges included; a royalty pool of 40% of weekly operating profit before recoupment, with $22,500 of weekly minimums; $15 million to recoup, and no tax credit counted; so the investors receive $196,168 a week at 90% full and the show breaks even at 61.2% of the house paid. In this issue the seats in the house are split into paid and complimentary: a complimentary seat counts in attendance and capacity, adds nothing to the gross, and costs nothing, since the running costs are fixed and the deductions are a share of gross. Educational only; not an offer to sell, or a solicitation of an offer to buy, any security.
Sources: the League's definitions of attendance, average paid admission and capacity, The Broadway League, Statistics: Broadway in NYC, 2025–2026 season; the 2009 change from paid to total attendance and from net to gross gross, Playbill (May 28, 2009), BroadwayWorld (May 29, 2009) and the League's weekly grosses page; the average's treatment of complimentary tickets, Broadway News (June 11, 2025); comps and papering, amNewYork (November 3, 2026); papering services and their fees, BroadwayWorld (September 19, 2022) and Play-By-Play's own terms; Tony voters' free invitations, The Tony Awards, Rules and Regulations 2025–2026, Rule 2(a)(vi), and the count of voters, BroadwayWorld (May 17, 2025); Shucked's preview pricing, Broadway News (March 29, 2023); reading the average before the capacity, Ken Davenport (January 5, 2008). Every figure in the table traces to the Broadway League's grosses for the week ending November 1, 2026 and the week ending October 25, 2026, as read in Pulse Nos. 06 and 05; the street's October 18 week as read in Pulse No. 04.
OFFSCRIPT is a producer or investor in nine of the 32 shows in the table: Buena Vista Social Club, Just in Time, Two Strangers, Operation Mincemeat, The Lost Boys, Maybe Happy Ending, Paranormal Activity, Other Desert Cities and A Few Good Men. They appear as rows, marked †, and nowhere else in this issue. Observations on other productions are inferences from published League data, not reported facts.
Next week in Unpacked: touring and global. Subsidiary rights: what a Broadway run keeps earning after it closes, and who it is paid to.
Investing Unpacked takes one mechanism of the Broadway business apart each week, alongside The Pulse, OFFSCRIPT's weekly read on the grosses. How we read them: our methodology.
This issue on Instagram: the narrated reel. Every issue: @offscript.theatricals.
Revised November 13, 2026: the opening brought to this week’s Pulse (week ending November 8); no other figure changed.