# Credentials, press and reviews, in full

Source: https://edfringeguide.com/analysis/credentials
Part of: What actually sells out at the Edinburgh Fringe — https://edfringeguide.com/analysis

Awards, press quotes and reviews against Edinburgh Fringe sellout rates: which credentials predict a sellout and which do not. Snapshot pinned at 21 August 2026.

35,035 ticketed performances across 3,292 shows, 3–21 August 2026. Snapshot pinned at 21 August 2026 (2026-08-21T19:40:00.000Z).

**Every rate below is a floor**: a date counts as sold out only where the outcome was observable, and unobservable dates count as not sold out.

## What a credential in the blurb is worth

_Specific award names carry real signal. Generic “award-winning” carries almost none. Both survive controlling for venue, price and region — or fail to._

**Shows whose blurb makes each claim, and how often they sell out at least one date**

| Claim in the blurb | Shows | Ever sold out | Still there after controls? |
|---|---:|---:|---|
| Edinburgh Comedy Award / Perrier | 58 | 82.8% | Yes — strongly |
| Melbourne Comedy Festival | 41 | 78.0% | Yes — strongly |
| “By popular demand” / returning | 45 | 68.9% | Yes |
| “Sold-out run” claim | 50 | 68.0% | Yes |
| Adelaide Fringe | 68 | 63.2% | Partly |
| So You Think You’re Funny / new-act awards | 35 | 62.9% | Yes |
| Any overseas festival credit | 152 | 62.5% | Yes |
| Any award claim at all | 479 | 54.5% | Weakly |
| TV credit | 254 | 52.0% | Weakly |
| Fringe First | 29 | 41.4% | No — not significant |

Base rate: of 2,644 shows with at least three past dates, 43.6% sold out at least one. “After controls” is a Mantel–Haenszel odds ratio stratified separately by venue operator, by region and by price band — the three obvious ways a credentialled act is different anyway. The phrase patterns behind each row are pinned in scripts/analysis/analysis-page-figures.py; three of them are narrower than the ones used at the previous cutoff, so those rows moved by definition as well as by data. The rows that did not change definition — Melbourne, Adelaide, Fringe First — reproduce their earlier figures exactly.

The pattern is that naming a specific award beats claiming one. “Edinburgh Comedy Award nominee” nearly doubles the base rate and keeps most of that after you account for the better venue and higher price such an act commands. “Award-winning”, pooled across 479 shows, barely moves.

Pooled is the wrong way to read that last row, but not because one genre carries it. The bare “award-winning” claim runs at ×1.47 in comedy and ×1.96 in theatre and both intervals clear 1 — 1.11 to 1.95, and 1.30 to 2.96. Theatre’s point estimate is the larger of the two; the ratio between them is 1.33 at p = 0.26, which is not a difference this many shows can establish. So: one effect, visible in both genres, larger-looking in theatre, and weaker in either than a named award. What the pooled “weakly” verdict adds is venue operator, region and price held fixed — controls the genre split below does not apply.

**A bare “award-winning” claim, split by genre**

| Genre | Shows claiming it | Ever sold out | Same genre, no claim | ×odds vs no claim |
|---|---:|---:|---:|---:|
| Comedy | 259 | 59.5% | 49.8% of 865 shows | 1.47 (1.11–1.95) |
| Theatre | 104 | 48.1% | 32.1% of 779 shows | 1.96 (1.30–2.96) |

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: 59.5 against 49.8 is ×1.47. It used to print the genre’s overall base instead (52.0% of 1,124 comedy shows, 34.0% of 883 theatre), which includes the claimants and reconstructs to ×1.36, and was corrected on 2026-08-25. Both intervals exclude 1 and they overlap heavily; theatre’s 104 shows are thin next to comedy’s 259. Read the direction, not the second decimal, and do not read the two rows as different findings.

> **Read this as a marker, not a lever**
>
> These are counts of what the blurb says, not verified nominations, and the numbers behind the strongest rows are small — 51 shows and 40 shows. An act that can honestly write “Edinburgh Comedy Award nominee” is a different act from one that cannot, in ways no control here captures. Adding the phrase to a blurb does not add the audience.

A separate question is whether the craft of the writing — not what it claims, but how it is put together — does anything once the room is taken away. The figures below are not the floor sellout rate used everywhere else on this page: they come from a fitted model of each show’s residual, how much deeper into its central-box-office allocation it sold than comparable shows at its own venue, holding room size, price, slot, day, genre, sub-genre, origin, duration and box-office system fixed, expressed in percentage points of allocation sold. A 4-point writing effect on that scale is not the same kind of number as the 38-point spread between venue operators above or the 28-point spread by price — it is a much smaller quantity, measured after the room is already accounted for, and the two should not be read side by side as if they shared a scale.

Of everything tested against that residual — sentence length, word count, readability, second person, opening with a question, opening with a quote, the rule of three, clichés, a book-now call to action — one thing replicates: sentence length. Blurbs in the longest-sentence fifth sell about 4 points less of their allocation than the shortest fifth, and it holds inside comedy and theatre taken separately.

**Average sentence length in the blurb, by quintile — residual, not floor sellout rate**

| Quintile | Words per sentence | Shows | Mean residual (pp) | Ever sold out |
|---|---|---:|---:|---:|
| Q1 (shortest) | 3–11 | 445 | +1.4 | 56% |
| Q2 | 11–14 | 445 | +1.8 | 52% |
| Q3 | 14–17 | 445 | −0.8 | 42% |
| Q4 | 17–20 | 445 | −4.1 | 40% |
| Q5 (longest) | 20–50 | 449 | −2.4 | 36% |

n = 2,229 shows with a description and at least four past dates (985 comedy, 733 theatre), snapshot fringe-2026-08-19.db. “Mean residual” is the venue-adjusted figure defined above; “ever sold out” is the ordinary floor share, shown alongside for comparison and falling the same way, 56% to 36%, across the same quintiles. Split by genre, the worst fifth is worst in both: Q5 comedy −3.8pp against theatre −3.0pp, versus Q1 comedy +0.6pp and theatre +2.4pp.

Everything else tested is flat. Second person address (“you/your”, used in 976 of the 2,229 blurbs), opening with a question, opening with a quote, the rule of three, clichés and a book-now call to action all sit inside noise. Blurb length itself — word count, not sentence length — is flat too: a 15-word listing and a 127-word one land within 2.6 points of each other with no monotone shape, once the venue effect an earlier blurb-language pass picked up is controlled for.

One asymmetry is worth naming: printing a press quote is worth about 6 points of allocation in theatre (733 theatre shows) and nothing measurable in comedy (985 comedy shows). Theatre audiences appear to buy on the review; comedy audiences appear to buy on the act.

> **Correlational, tested against 18 features, fixed before the run**
>
> A producer who writes short sentences is not a random producer, so this cannot separate “short sentences work” from “the kind of producer who writes short sentences differs in other ways”. Eighteen boolean prose features were tested against the residual, so one or two are expected to clear the noise floor by chance alone even if none of them are real — this section publishes only the one effect that repeats across two independent genre samples. And blurbs are fixed before the run and never updated, so this is a single measurement taken before the festival, compared against an outcome measured over three weeks of it.


## Which publication’s review is worth having

_Listings advertise their press in a parseable shape — a quote, usually a star rating, then an attribution in brackets. Extracting all of it turns “do reviews help” into a question with different answers per masthead, and per genre._

**Sellout rate by the best masthead the listing can claim**

| | Sold out | Dates |
|---|---:|---:|
| National press (Guardian, Times, Telegraph…) | 36.1% | 2,435 |
| Overseas press (NYT, Rolling Stone, Aus/NZ) | 26.4% | 887 |
| Comedy trade (Chortle, Beyond the Joke) | 19.9% | 1,014 |
| Fringe review sites (Broadway Baby, ThreeWeeks…) | 18.6% | 1,614 |
| Arts & theatre trade (The Stage, The List, Time Out) | 18.3% | 1,565 |
| Scottish press (Scotsman, Herald) | 16.3% | 1,173 |
| No outlet quoted (1,545 shows — the default) | 14.5% | 16,488 |

Each show counted once, at the top of what it holds, on operators that report both outcomes. 93 named publications were extracted from the blurbs of all 4,306 shows; 1,155 shows name at least one. 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.

> **A Scotsman review and nothing better looks worse than nothing — on too few shows to say so**
>
> Raw, the direction is upward: shows whose best claim is Scottish press sell out 16.3% of their dates against 14.5% for shows quoting nobody, and in comedy alone it is 19.5% against 14.7% on the Aug-19 cut below. The negative sign appears only once venue operator is held fixed, where the odds ratio is 0.75. That point estimate is the finding. The interval around it is not: on dates it reads 0.63 to 0.90 and excludes no-effect, but 1,173 dates belonging to 99 shows are not 1,173 independent draws. Resample the shows instead — the standard this page applies in the nulls, in work-in-progress and in rooms, and the standard that withdrew the ALL-CAPS finding — and it is 0.47 to 1.15, straddling 1. Below 1 after controls, on 99 shows, not established. If it is real it is not a comment on the reviewing: the Scotsman is the most-quoted masthead in the programme, so “reviewed by the Scotsman and by no national” is close to a definition of the solid middle of the Fringe, and what would be predicting is scarcity, not prose.

**The named outlets, and what survives the corrections**

| Outlet | Shows | Dates | Raw | ×odds in venue | + genre | vs other press |
|---|---:|---:|---:|---:|---:|---:|
| Evening Standard | 10 | 116 | 52.6% | 3.72 | 2.83 | 3.19 |
| Independent | 22 | 255 | 49.0% | 3.46 | 3.91 | 2.80 |
| Telegraph | 52 | 567 | 44.1% | 2.36 | 1.78 | 2.00 |
| New York Times | 18 | 207 | 38.2% | 2.08 | 2.08 | 1.70 |
| Time Out | 24 | 303 | 35.0% | 2.03 | 1.68 | 1.69 |
| Guardian | 124 | 1,504 | 35.8% | 1.67 | 1.38 | 1.52 |
| ThreeWeeks | 57 | 508 | 25.2% | 1.55 | 1.32 | 1.24 |
| The Times | 54 | 582 | 34.7% | 1.46 | 1.37 | 1.32 |
| Chortle | 130 | 1,600 | 30.2% | 1.40 | 1.10 | 1.20 |
| The List | 73 | 920 | 25.4% | 1.32 | 1.08 | 1.05 |
| Broadway Baby | 79 | 901 | 20.9% | 1.20 | 1.09 | 0.91 |
| BroadwayWorld | 45 | 595 | 27.4% | 1.20 | 1.03 | 0.95 |
| Scotsman | 153 | 1,818 | 25.7% | 1.18 | 1.08 | 0.95 |
| The Stage | 24 | 297 | 19.5% | 0.87 | 1.04 | 0.57 |
| The Skinny | 37 | 430 | 24.7% | 0.87 | 0.69 | 0.76 |
| Fest | 25 | 297 | 22.2% | 0.71 | 0.49 | 0.56 |
| EdFringeReview | 32 | 389 | 15.7% | 0.67 | 0.62 | 0.57 |
| Everything Theatre | 15 | 170 | 7.1% | 0.35 | 0.31 | 0.22 |

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 are suppressed.

> **Three different things are happening in that table**
>
> The scarce mastheads predict: the Independent and Evening Standard review very few Fringe shows and the ones they pick run at three to four times the odds, and it holds even against other reviewed shows. The ubiquitous ones predict nothing: the Scotsman lands at 0.95 against other reviewed shows, dead flat. And the specialists predict their own speciality: Chortle’s raw 30.2% collapses to 1.10 once genre is held fixed, because Chortle reviews comedy and comedy sells out more than theatre. Only the first of the three is really about the review.

**Comedy and theatre want different press**

| Best masthead held | Comedy | ×odds | Theatre | ×odds |
|---|---:|---:|---:|---:|
| National press | 37.0% | 1.42 | 31.5% | 2.39 |
| Overseas press | 27.5% | 1.20 | 29.6% | 1.87 |
| Fringe review sites | 14.8% | 0.88 | 25.7% | 2.05 |
| Arts & theatre trade | 16.5% | 0.70 | 17.0% | 1.14 |
| Comedy trade | 19.0% | 0.77 | — | — |
| Scottish press | 19.5% | 0.85 | 9.4% | 0.57 |
| No outlet quoted | 14.7% | — | 11.4% | — |

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 Aug-21 cutoff above.

> **For comedy, only a national is worth anything**
>
> This is the room-size reversal again, in a different variable. In comedy the national press is the only tier that beats quoting nobody: the comedy trade runs at 0.77, the arts trade at 0.70 and the Fringe review sites at 0.88 — at or below a listing with no press at all. In theatre almost the opposite holds. A Fringe review site is worth 2.05 there, statistically indistinguishable from a national, and Broadway Baby alone runs at 1.67 in theatre against 0.77 in comedy. The plainest reading is that a comedy audience is largely buying a name, so only the coverage that confers fame moves anything; a theatre audience is buying that particular play, and a Fringe review site is exactly the mechanism by which an unknown play gets found.

**What a star rating is worth, by who gave it**

| | Sold out | Dates |
|---|---:|---:|
| ★★★★★ national | 39.6% | 908 |
| ★★★★ national | 35.4% | 1,238 |
| National, no stars printed | 31.4% | 1,658 |
| ★★★★★ trade | 28.8% | 1,264 |
| ★★★★★ Fringe site | 26.5% | 1,851 |
| ★★★★ trade | 23.9% | 1,756 |
| ★★★★ Fringe site | 14.3% | 1,256 |
| No outlet quoted | 14.5% | 16,488 |

Stars are read from the asterisks printed immediately before the attribution, which is the convention every listing uses.

> **A four-star from a national is nearly a five-star; a four-star from a Fringe site is nothing**
>
> Dropping from five stars to four costs a national outlet 4.2 points and costs a Fringe review site 12.2 — from 26.5% down to 14.3%, close to the rate for listings quoting no press whatsoever (14.5%). A national quote with no rating printed at all still runs at 31.4%. At national level the decision to review is the scarce thing and the rating is a detail; on sites that review most of what they are sent, only the top mark separates anything.

**What sort of sellout each kind of press goes with**

| Best masthead held | Sells out ≥1 date | Share of its dates | Whole run gone | Median lead | Median room |
|---|---:|---:|---:|---:|---:|
| National press | 74.2% | 33.2% | 6.9% | 35 h | 102 |
| Overseas press | 61.0% | 24.7% | 1.7% | 33 h | 100 |
| Arts & theatre trade | 58.7% | 17.3% | 2.8% | 19 h | 74 |
| Comedy trade | 54.3% | 18.5% | 1.4% | 7 h | 56 |
| Fringe review sites | 54.1% | 16.9% | 0.9% | 18 h | 70 |
| Scottish press | 50.6% | 15.4% | 1.3% | 20 h | 80 |
| No outlet quoted | 38.8% | 12.6% | 1.9% | 32 h | 60 |

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 Aug-21 cutoff above.

National press buys breadth; the comedy trade buys the last hour. Three in four shows holding a national quote sell out at least one date and they sell out a third of their run, with nearly seven percent going completely clean — three to seven times any other group. A Chortle-quoted show is a different animal: an £11 ticket in a 56-seat room that goes with a median seven hours left on the clock, against 35 for a national-press show.

> **This one survives the controls that killed the award nomination**
>
> Holding venue operator fixed, the national-press advantage is ×1.77; adding room size, ×1.72; adding genre, ×1.49; adding price band as well, ×1.38 with an interval from 1.19 to 1.60. It halves and it stays — unlike the award-nomination effect, which vanishes once the same controls are applied together. A single masthead is weaker: the Guardian on its own finishes at ×1.20 with an interval that barely clears 1.00, so the defensible claim is about the tier and not about any one paper. And it is still an association. A blurb is fixed text, so none of this can show a review landing mid-run and tickets moving the next day; what it measures is the standing an act already had before the festival opened. Dated reviews can be tested that way, and are, in the section below: nothing moves. This step-wise control chain is not part of press-figures.json and is still read off fringe-2026-08-19.db.


## A review lands, and the tickets do not move

_Everything above is blurb text — what a show advertises, and therefore the standing it already had. Reviews carry a date, so they can be tested the other way round: seats moving in the days after the first review against the days before. Nothing measurable happens, and the star count does not order._

For each show’s first professional review, the seats moving in the four days after are compared with the four days before, and that difference is taken against control shows matched on genre and venue tier over the same calendar days, none of them reviewed inside the window. The control group is what makes the number mean anything: a show sells faster as its dates approach and as the festival goes on, so without it every review scores as a hit.

**Change in selling speed after the first review, against matched controls**

| Reviews rated | Events | Points/day | 95% interval | Placebo |
|---|---:|---:|---:|---:|
| Every first review | 256 | −0.04 | −0.51 to +0.43 | −0.59 |
| Five stars | 20 | +0.14 | −1.46 to +1.74 | +1.28 |
| Four stars | 67 | +0.44 | −0.52 to +1.40 | −1.54 |
| Three stars | 20 | +0.33 | −1.48 to +2.13 | −1.76 |
| No star stated | 149 | −0.33 | −0.94 to +0.27 | −0.26 |

Points of the central box-office allocation sold per day, averaged across the show’s still-future performances so a 400-seat room does not outvote a 40-seat one. Snapshot fringe-2026-08-21.db, the same cutoff as the rest of this page — the reviews table was added to that export for this question. The 6–10 August outage days are excluded from every window mean, not zero-filled — three have no history rows at all and the fourth is a catch-up reading that would otherwise read as a demand spike. “Placebo” is the identical calculation run one window earlier, before the review existed; it should sit near zero, and a placebo as large as the estimate means the design is reading a trend rather than an event. 256 of 1,082 first reviews are usable at this window.

Zero, with an interval running from −0.51 to +0.43 points a day. Over the four days that bounds any effect at roughly ±2.0 points of a show’s house — under one seat a night in a 50-seat room. It holds inside every cut tested: top-tier venues, mid, everything else, and comedy on its own.

> **Narrow the window and it turns positive — that is mean reversion, not press**
>
> At ±2 days the headline reads +0.88 points a day with an interval clear of zero, and every star group lifts with it. The placebo moves in step: those same shows were selling 0.63 points a day BELOW their controls in the days before the review, and three-star shows 2.68 below. A review lands on a show at a moment its sales have dipped, and the dip corrects. The placebo is the same size as the estimate, so on this data the two cannot be told apart, and the honest reading is the wider window’s zero.

> **Three stars scores as well as five, at every window**
>
> At ±2 days the ordering is three stars +3.60, four +2.06, five +1.76 — inverted. At ±4 it is flat. If a review sold tickets through its verdict, five stars would beat three; a flat-or-inverted ordering is what a noisy measure looks like. This is also the one comparison here that is close to clean, because critics choosing which shows to review is a selection problem, while the rating they give a show they have already chosen is much less so.

Audience reviews do nothing either, on four times the sample: −0.11 points a day across 1,094 events, interval −0.53 to +0.30, and a placebo of +0.33 — bigger than the estimate itself, though still small next to the professional placebos above, and pointing in neither direction. Audience reviews arrive because people came. They do not bring more.

> **What would overturn this**
>
> The cutoff, first. This snapshot stops on 21 August against a festival running to the 31st, so every review from the final ten days is unusable and the ones that survive are the early ones — a rebuild at festival end roughly doubles the sample and is the single thing worth doing before believing any of it. Beyond that: only each show’s FIRST review is used, so breadth of coverage is untested; “a review landed” is binary, with no measure of how many; and which outlet reviewed the show is recorded but not yet tested. Sales are read from a daily history, so a morning review moving that evening’s house is invisible at this resolution.



## Method, in full

The single largest limitation on this chapter is that press is read from blurb text, not from a press archive. A blurb is fixed text written before the festival opened, so what it measures is the standing an act already had, not a review landing in week one and tickets moving in week two. Nobody prints a bad review in their own listing either, so every outlet’s population is selected upward before any measurement happens. Where a claim on these pages can be tested the other way round — reviews carry a date, blurbs do not — it is, and the answer is in the review section: nothing moves.

Two figures in this chapter come from an earlier snapshot than the rest of the page and say so inline rather than being quietly restated: the comedy/theatre press split and the breadth cut. They were computed against a cut two days earlier and have not been rerun, and the honest handling of a figure that has not been refreshed is to date it, not to present it as current. The tier and outlet figures beside them are current.

Odds ratios in this chapter are stratified, not raw. A claim appearing more often at one operator would otherwise measure that operator, so the comparison group for any claim is shows at the SAME operator without it, and the strata are combined rather than pooled. Where an interval is quoted it resamples shows and not dates, for the same reason as everywhere else: a run of twenty dates is close to one observation repeated, and a date-level interval on it comes out about half the width it should be.