What actually sells out at the Edinburgh Fringe

Cells marked “redacted” are measurements of how much of a show’s central box-office allocation sold, temporarily withheld pending a review of data use with the festival organisers. The remaining columns are unchanged.

• Traverse Theatre12 shows≥ 89.6% n=270
• Monkey Barrel157 shows43.5% n=1,893
• Pleasance266 shows32.1% n=4,819
• The Stand110 shows≥ 25.4% n=638
• Summerhall80 shows24.5% n=1,708
• Underbelly192 shows19.0% n=3,645
• Assembly238 shows15.8% n=4,124
• theSpace480 shows13.3% n=4,523
• Gilded Balloon203 shows10.6% n=3,897
• ZOO51 shows9.9% n=928
• Hoots217 shows7.7% n=4,014
• Laughing Horse / Free Fringe223 shows≥ 6.9% n=4,570
• Greenside196 shows6.5% n=2,323
• Just The Tonic224 shows5.2% n=3,060
• C venues117 shows3.2% n=1,098
• Braw Venues57 shows≥ 2.8% n=1,025

Floor rate: a date counts as sold out only where we could see it, so every bar is a lower bound. For most operators the ceiling is only a few points higher, because they sell through their own ticketing as well as the Fringe’s and both outcomes are visible. The four marked ≥ have no ceiling at all — see below. Twelve timed-entry and walk-around listings are excluded here, as exhibitions and events already were — see “What these numbers cannot tell you”. They moved six of these bars, and the other ten not at all. Five moved by points: Underbelly from 23.5% to 19.0%, Summerhall from 27.8% to 24.5%, C venues from 6.4% to 3.2%, Assembly from 13.5% to 15.8% and Pleasance from 30.3% to 32.1% — the last of those ARCADE’s 432 slot dates at 9.7%, which had been holding Pleasance down. The sixth barely moves and is listed for completeness: Hoots from 7.9% to 7.7%, on 103 dates.

A +4.89-point gap on shows the ranking never saw, with a 95% interval of +2.03 to +7.71 resampling SHOWS rather than dates — dates inside one show are not independent and a date-level interval here would be about half the width it should be. Repeating the whole split ten times gives a mean gap of +5.06pp and it never changes sign. This is the one table on the page built on a random split; the split is seeded, so a refresh that moves this number means the data moved.

• Under 50 seats18.0% n=12,426
• 50–9913.7% n=20,499
• 100–19918.7% n=11,704
• 200–39925.9% n=2,863
• 400+35.0% n=2,263

Capacity is the room’s, from the venue’s own seating data, and is known for 96.5% of dates. Excludes the twelve shape-detected timed-entry listings — see “What these numbers cannot tell you”. They were worth ten points to the 200–399 band on their own, which used to top this chart at 36.0%.

Excludes dates sold only through the Fringe’s central box office, which reports sellouts but not misses and is 26.0% of comedy against 12.6% of theatre — pooling the two without that exclusion understates comedy. The 400+ theatre cell is 107 dates; treat it as a direction, not a number.

Minimum 40 dates per cell. Theatre falls in four operators of four, and it is the half of this table that survived the full season intact. Comedy rises in three of the four — Assembly, Underbelly and Gilded Balloon — and is flat at Pleasance. An earlier version of this table reported comedy running the other way at Gilded Balloon, its small rooms beating its big ones by nine points; that was entirely an artefact of the faulty post-performance readings described in “What these numbers cannot tell you”, and is withdrawn — corrected, its small comedy rooms lose to its big ones by 4.5 points. So the pooled comedy chart above is carried by the room bands across the whole festival rather than by any one chain, and the “opposite directions” headline is properly a claim about theatre.

Cells marked “redacted” are measurements of how much of a show’s central box-office allocation sold, temporarily withheld pending a review of data use with the festival organisers. The remaining columns are unchanged.

Shows with at least three past dates, as the pooled table above. The third column is the comparison group the odds ratio actually uses — shows in the SAME genre making no award claim — so the row reconstructs: 62.7 against 51.1 is ×1.60. It used to print the genre’s overall base instead, which includes the claimants and reconstructs to the wrong ratio; corrected on 2026-08-25. Both intervals exclude 1 and they overlap heavily; theatre’s 116 shows are thin next to comedy’s 284. Read the direction, not the second decimal, and do not read the two rows as different findings.

Cells marked “redacted” are measurements of how much of a show’s central box-office allocation sold, temporarily withheld pending a review of data use with the festival organisers. The remaining columns are unchanged.

• National pressGuardian, Times, Telegraph…33.6% n=3,808
• Overseas pressNYT, Rolling Stone, Aus/NZ25.4% n=1,427
• Comedy tradeChortle, Beyond the Joke20.4% n=1,589
• Fringe review sitesBroadway Baby, ThreeWeeks…16.8% n=2,572
• Arts & theatre tradeThe Stage, The List, Time Out16.6% n=2,435
• Scottish pressScotsman, Herald14.2% n=1,830
• No outlet quoted1,809 shows — the default13.3% n=25,894

Each show counted once, at the top of what it holds, on operators that report both outcomes. The seven bars are 39,555 dates; the population behind them is 39,683. 1,809 shows in it name no recognised outlet at all, and the six quoting tiers hold 810 shows between them. The 128 dates that are in the population and not on the chart are an eighth tier — “Personal / broadcast”, where the best thing a listing can claim is a named critic or a broadcaster rather than a masthead — and it is the highest rate on the board, 52.3% over 8 shows, interval 43.7 to 60.8. It gets no bar for the reason Wales and Northern Ireland get no bar in the origin chart further down: eight shows and an interval seventeen points wide cannot be ordered against anything drawn here, and a bar above national press would assert exactly that ordering. It is named here rather than folded into a neighbouring tier, because folding it would move some other tier’s rate without saying so. Read every row as “shows that advertise this outlet”, never “shows this outlet reviewed” — nobody prints a bad review, so each population is self-selected upward, and none of it is checked against a press archive.

Odds ratios against listings quoting no recognised outlet, Mantel–Haenszel, stratified by venue operator and then also by genre. The last column compares each outlet only against shows quoting some OTHER publication, which removes the has-any-press advantage — it is the column that separates a masthead worth having from one that merely accompanies a well-booked show. Outlets under 10 shows or 60 dates in the population cannot appear here at all; of the 43 that clear that floor this table shows 18, chosen to span the range rather than by any rule — press-outlets.py prints all of them and is the list to read if you want an outlet this one leaves out. The omitted rows include several with odds above the Guardian’s, among them the BBC at 3.46 and Theatre Weekly at 2.40.

Comedy 9,253 dates, theatre 8,406, on operators that report both outcomes. Odds ratios within venue operator, against listings quoting no outlet. Theatre’s comedy-trade row is suppressed — too few dates to publish. Snapshot fringe-2026-08-19.db — this cut is not part of press-figures.json and has not been rerun at the season-final cutoff above.

• ★★★★★ national80 shows36.1% n=1,419
• ★★★★ national125 shows33.5% n=1,948
• National, no stars printed157 shows29.2% n=2,622
• ★★★★★ trade109 shows26.6% n=1,975
• ★★★★★ Fringe site173 shows24.9% n=2,967
• ★★★★ trade153 shows22.7% n=2,671
• ★★★★ Fringe site115 shows12.8% n=1,956
• No outlet quoted1,809 shows13.3% n=25,894

Stars are read from the asterisks printed immediately before the attribution, which is the convention every listing uses. The n is DATES; the show count is beside it because dates inside one run are not independent observations, and every bar here rests on between 80 and 173 shows. “National” is twelve titles from the Guardian to the Sun and they do not behave alike — the outlet table above ranks the ones with enough shows to rank, and the Evening Standard and the Independent sit at the top of it, not the papers with the most listings.

The three national titles with enough listings to split — Guardian 138 shows, The Times 61, Telegraph 61 — as a share of each group’s dates that sold out. The cells behind these nine numbers hold between 7 and 64 shows: the Guardian’s four-star cell is 64 shows and its five-star cell is 11, so read the shape of the table and not any single figure. Split the tier by masthead AND star and thirty-one cells appear, of which six hold twenty shows or more — which is why this is a caveat on the chart above and not a ranking of newspapers.

Breadth measured on shows with at least five past dates. Lead time is the median hours before curtain at which the live monitor caught the sellout, across sellouts it timed. Snapshot fringe-2026-08-19.db — this cut is not part of press-figures.json and has not been rerun at the season-final cutoff above, which also means it is unaffected by the closing-week sampling change described in the appendix: every date in it was observed on the ten-minute cadence.

Cells marked “redacted” are measurements of how much of a show’s central box-office allocation sold, temporarily withheld pending a review of data use with the festival organisers. The remaining columns are unchanged.

• £1–66.4% n=7,040
• £6–107.6% n=8,832
• £10–1415.6% n=17,435
• £14–1821.3% n=11,858
• £18+33.7% n=2,709

Cheapest listed price per date; bands include their upper figure, so a £10 ticket is in “£6–10”. That is worth stating because £10 alone is 4,499 dates — put it in the band above instead and the £6–10 band falls to 4,333 dates at 7.4% rather than 7.9%. Free-but-ticketed dates are excluded: over half of them are one show, and they behave nothing like the rest of the programme. Exhibitions, events and the twelve shape-detected timed-entry listings are excluded too — see “What these numbers cannot tell you”.

Excludes dates sold only through the Fringe’s central box office, for the same reason as the room-size split above — comedy is 26.0% system-1 against theatre’s 12.6%, and without the exclusion theatre’s dearest cell is largely Traverse, whose rate is a floor with no ceiling. Comedy climbs more than nine-fold end to end and never once turns down; theatre ends under three times where it starts, and by way of one reversal at the cheap end. Both are still associations — the adjustment below applies to each of them.

Variable-price shows only, measured against each show’s own minimum price — so a show that never played a preview has its midweek price as its baseline, which pushes the Wednesday and Thursday rows mechanically towards zero. Separately, 859 shows played 5–6 August at a median £5.00 below their own later price. Do not read the sold-out column as an effect of the price: Saturday is dear AND full for reasons the weekend table already covers.

• Assembly98.0% n=5,058
• Gilded Balloon98.0% n=3,890
• Underbelly98.0% n=3,865
• Pleasance97.0% n=5,195
• Summerhall87.0% n=1,990
• C venues85.0% n=1,238
• ZOO85.0% n=928
• Monkey Barrel78.0% n=1,884
• Greenside42.0% n=2,304
• Just The Tonic33.0% n=3,052
• Hoots18.0% n=4,046
• theSpace14.0% n=4,502
• Laughing Horse / Free Fringe2.0% n=4,533
• Braw Venues2.0% n=1,025

A share of dates, not a sellout rate — the bars here are not comparable in meaning to any other chart on this page, only to each other. Yield management is close to a perfect operator attribute: the Big Four run it on 97% or more of their dates and the free-fringe and theatre-block operators barely at all. Which means any date-level price analysis of this festival is quietly an analysis of four venue chains.

Every operator the query returns, ranked on the last column: a row needs 60 variable-price dates in both its Mon–Thu and its Fri–Sun cell, and all seventeen clear it on the full season. The preview cell needs the same 60 dates, which is why some chains show no preview row — they do not price previews separately. On the differential the steepest weekend house is Hoots, which charges four times as much extra on a Friday as on a Tuesday off a low base; on the largest absolute premium it is Summerhall at +£6.17, and Summerhall’s £6 is flat between Mon–Thu and Fri–Sun rather than every day of the week — its previews carry none of it. Traverse, Paradise, theSpace and Greenside are inverted, charging more on weekdays than at the weekend. Gilded Balloon’s tariff is flat and small. One shared convention plus independent house tariffs is what a market looks like; identical calendars would have been the interesting finding, and they are not what is there. This table published eight rows until 2026-08-25, with no stated rule and with both ends of the ranking among the ones it dropped, under a footnote calling Pleasance — sixth here — the clearest weekend house.

290 rooms with at least 30 priced dates; the comparison group is the 19,260 dates priced exactly at their room’s median. The raw rate column falls above 1.25× and the odds column rises, because the shows priced far above their room’s norm sit in cheap rooms. The apparent ceiling is room composition. As everywhere else on this page, none of this is a lever: producers price to the demand they expect.

Clean window is 12–19 August, which excludes both the monitoring outage and the 2-for-1 days. The strata are severely unbalanced — inside £10–14 there are 4,753 discounted dates against 288 undiscounted ones — so the sub-rows are themselves fragile. This is the one section on the page that has not been recomputed at the season-final cutoff: it is a warning about a variable rather than a finding from it, and rerunning a composition artefact on more dates makes it a bigger composition artefact, not a truer number.

• Children’s shows127 shows19.6% n=1,795
• Comedy1,423 shows16.5% n=22,955
• Opera179 shows15.4% n=2,013
• Theatre1,095 shows13.9% n=14,955
• Cabaret150 shows13.9% n=2,034
• Music381 shows13.5% n=2,352
• Spoken word141 shows12.6% n=500
• Circus124 shows8.9% n=1,579

Exhibitions and events are excluded: walk-in and timed-entry formats barely use the booking channel, so their rate measures the format, not demand. Since 1 September the same exclusion is made by SHAPE as well as by genre — twelve listings running five or more performances on one calendar date, filed under theatre, music and comedy, so the genre test missed them. They were the difference between theatre at 17.3% and theatre at 14.6%, and between music at 17.4% and music at 13.8%; comedy’s rate does not move. See “What these numbers cannot tell you”.

Cells marked “redacted” are measurements of how much of a show’s central box-office allocation sold, temporarily withheld pending a review of data use with the festival organisers. The remaining columns are unchanged.

• Australia22.8% n=2,733
• Scotland18.8% n=8,398
• England16.0% n=21,727
• Rest of UK13.2% n=653
• Ireland13.1% n=1,769
• USA10.7% n=7,933
• Elsewhere10.0% n=4,970

An origin gets its own bar at 60 shows or more. That is what selects these five and not a list anyone chose: Ireland clears it at 93 shows and 1,769 dates, and the next origins down are Canada at 47 shows and New Zealand at 34. Ireland was inside “elsewhere” until 27 August and was much the largest thing in it — bigger than Canada, New Zealand and Wales together. From the same date no UK nation sits in “elsewhere” either: Wales and Northern Ireland are below the floor, so what is left of the UK once Scotland and England are drawn goes in a “rest of UK” row rather than a global one — a residual named as a residual, not two unrelated nations bracketed as though they were one. They are not drawn separately because they cannot be. Wales is 32 shows at 11.0%, interval 4.0 to 19.4; Northern Ireland is 15 shows at 18.9%, interval 4.2 to 38.5 — a bar at 18.9% would rank it above England off an interval that cannot order it against anything on this chart. Both intervals resample shows rather than dates. “Elsewhere” now means everywhere outside the UK, Ireland, Australia and the United States. Exhibitions, events and twelve shape-detected timed-entry listings are excluded — see “What these numbers cannot tell you”. Scotland is the bar they moved most, and it reads 18.8% without them: one Scottish listing playing the Scottish Parliament banked 461 sold-out slots, 4.6% of every sellout at the festival.

Cells marked “redacted” are measurements of how much of a show’s central box-office allocation sold, temporarily withheld pending a review of data use with the festival organisers. The remaining columns are unchanged.

Median run: 8 dates against 21. The two right-hand columns are the whole finding — the date column says a work in progress was nearly seven points more likely to go, the show column says it was four points LESS likely to ever go, and neither is wrong.

Mantel–Haenszel, on 1,121 dates from 118 shows. A show-clustered interval on the raw seven-point gap runs +0.3 to +15.4 points — wide, and only just clear of zero. It is not one venue’s house style either: Hoots 365 dates, Monkey Barrel 263, Laughing Horse 145, Just The Tonic 70.

Every row is a Mantel–Haenszel odds ratio against the rest of the driver population, held inside the strata its own row names — five rows controlled five different ways, so the column saying which is not decoration. Twenty hypotheses at the usual threshold buys about one false positive by chance, so nothing here was promoted on a date-level p-value: each was re-tested either with an interval that resamples SHOWS rather than dates, or on a held-out set of shows. The two candidates that died at exactly that step are in this table.

Cells marked “redacted” are measurements of how much of a show’s central box-office allocation sold, temporarily withheld pending a review of data use with the festival organisers. The remaining columns are unchanged.

• 7pm19.5% n=4,349
• 8pm17.0% n=3,882
• 6pm16.8% n=4,069
• 3pm16.7% n=3,974
• 11am15.7% n=2,674
• 5pm15.6% n=4,041
• 9pm15.4% n=3,788
• 11pm15.4% n=1,392
• 4pm14.6% n=3,978
• 1pm14.0% n=3,708
• 2pm12.7% n=4,011
• 10pm12.5% n=2,582
• Noon11.8% n=3,779

UK start time, 11am to 11pm. Earlier slots are a different market — a 9am show is not competing with an 8pm one — and are left out rather than ranked against them. Exhibitions, events and the twelve shape-detected timed-entry listings are excluded — see “What these numbers cannot tell you”. They used to put 3pm on top of this chart at 19.2%; on the corrected population 3pm is fourth at 16.7% and the peak is 7pm.

Same population as the pooled chart above. The 11pm theatre cell is 145 dates — the only row here thin enough to ignore.

The clean window is 11–30 August: it drops the 6–10 August monitoring outage at the start and the censored final day at the end, and on the completed season it holds three Saturdays and three Sundays rather than one of each. That is what settles the open question, and it settles it in the wrong direction for the phrase “the weekend”: Sunday runs +0.4 points above the weekday average, which is to say level with it. Saturday is the weekend.

Same clean window, excluding dates sold only through the Fringe’s central box office. Against its own weekday baseline comedy’s Saturday is worth +18.5 points and its Sunday +0.8; theatre’s Saturday is worth +6.4 and its Sunday +2.9. Comedy’s weekend is Saturday and nothing else. Theatre’s is smaller and does include the Sunday — a difference in shape rather than in size, and on three Sundays rather than the single one the mid-festival snapshot could see.

• ZOO68.5% n=92
• The Standfloor-only venue54.9% n=162
• Underbelly54.2% n=694
• Monkey Barrel46.1% n=824
• Just The Tonic38.1% n=160
• Traverse Theatrefloor-only venue33.9% n=242
• Assembly32.2% n=653
• Pleasance31.5% n=1,548
• theSpace25.8% n=600
• Gilded Balloon41.3% n=414
• Summerhall23.6% n=419
• Hoots18.7% n=310
• Laughing Horse / Free Fringe11.8% n=314
• Greenside11.3% n=151

n is the operator’s sellouts, not its dates. Bar scale runs to 70%, not 100% — this chart is not comparable in length to the sellout-rate charts above. Fourteen of nineteen operators are shown: those with fewer than 80 sellouts across the season are omitted, and so is the “Other / independent” bucket, which is several hundred unrelated rooms rather than one operator. The twelve shape-detected timed-entry listings are excluded here as they are from the other operator league on this page — a slot format’s “reopening” is its release schedule, not a box office putting returns back on sale, and Underbelly’s reopen rate was reading 41.6% with one of them in it against 54.2% without. Their departure also takes C venues below the eighty-sellout floor, from 80 to 35, so it is no longer drawn. Every operator, on either measure, is in the explorer.

See this on every row in the data explorer →

Every headline finding on this page survives the stricter definition. What does not survive is the size of the venue spread at the top of the table. The percentage rows are generated from the snapshot and asserted against it. The two ×odds rows are stratified odds ratios owned by scripts/analysis/press-outlets.py: its reached-zero column is refreshed at this cutoff, its stayed-gone column has not been rerun since 21 August and is marked in the cell rather than presented as current.

Cells marked “redacted” are measurements of how much of a show’s central box-office allocation sold, temporarily withheld pending a review of data use with the festival organisers. The remaining columns are unchanged.

Cells marked “redacted” are measurements of how much of a show’s central box-office allocation sold, temporarily withheld pending a review of data use with the festival organisers. The remaining columns are unchanged.

The top 100 within each genre’s own catalogue, not the festival’s top 100 filtered down to a genre — those answer different questions. Four theatre sellouts in five belong to under a tenth of theatre shows. Part of the 30.7-point gap between the two rows is arithmetic, because a fixed 100 is 9.1% of theatre’s catalogue and only 7% of comedy’s: cut both genres at 7% and theatre’s top 77 shows still carry 73.6% of its sellouts against comedy’s 49.5%, a gap of 24.1 points. Like-for-like it is a shade under twenty-five points, not thirty-one. Either way the pooled headline understates how far theatre’s whole sold-out category rests on a small hit list, and is roughly right for comedy, which is why nobody caught it.