Daniel's been thinking about the socially deviant hobby of hunting down the worst-reviewed restaurants in a city. You know, the ones with hundreds of reviews and two and a half stars, the places everyone warns you about. He points out that what these spots usually share is a certain tourist trap vibe — the sidewalk hustler, the menu in six languages with flags, the laminated photos of food that looks nothing like what arrives. But here's the part that stuck with me: a few doors down, you'll sometimes find a place that simply doesn't bother with any of that. No English menu, no greeter, and yet it's the one that's actually extraordinary. Daniel's question is how an entire area tips from the second kind of place into the first. How do ordinary restaurants, hotels, whole streets fall into the tourist trap sinkhole and pull each other down with them — and can that ever go the other direction?
The thing is, we tend to talk about tourist traps like they're a moral failing. A lazy chef, a greedy owner, someone who decided to stop trying. But that's not what's happening. A tourist trap is an equilibrium. It's what a market produces when you pump ten thousand hungry people a day through a three-block radius, none of whom know anything about the local food, most of whom will never come back, and all of whom are making decisions based on the same handful of signals on their phone.
So it's not a bad restaurant. It's a rational response to a specific set of conditions.
No, wait. It's both. It is a bad restaurant. The food is mediocre, the service is indifferent, the prices are high. But the mediocrity isn't an accident. It's the output of a system that selects for mediocrity. And that system has three or four distinct mechanisms that lock into place one after another.
Walk me through them.
First mechanism is what behavioral economists call an information cascade. Picture this: you're standing near the Colosseum. You're hungry. You open Google Maps and search "best pizza near me." What do you see? The places with the most reviews. Not the highest rated — the most reviewed. There's a place with four stars and two thousand reviews, and a place with four point seven stars and forty reviews. Which one feels safer?
The two thousand reviews. Even if the average is lower, the sheer volume feels like a consensus.
Because we use review count as a proxy for reliability. The BBC Travel ran a piece on this in twenty twenty-one — they talked to psychologists who study tourist decision-making, and the finding was that when people are in an unfamiliar environment with limited time, they don't optimize for quality. They optimize for certainty. They want to know what they're getting. And a restaurant that's been reviewed by two thousand strangers feels more knowable than one reviewed by forty, even if the forty are all raves.
So the algorithm sees high engagement, serves it to the next tourist, who goes there, leaves another review, and the gap widens.
That's the cascade. Early traffic creates more traffic, regardless of whether the food is any good. And here's where it gets perverse: the tourists who go to the high-volume spot are exactly the people least equipped to evaluate it. They don't know what the local food is supposed to taste like. They're comparing it to the version they get at home. So they might leave a four-star review for a pasta that an Italian would send back, because to them it was fine. It was familiar. The cascade doesn't just amplify volume — it degrades the quality of the signal.
So the review system that's supposed to surface good food is actually surfacing familiar food.
And familiar food is what tourist trap kitchens learn to make. Pre-prepared sauces, frozen pizza bases, the same ten dishes every tourist recognizes. You don't need a good chef. You need a reliable one.
That's mechanism one. What's the second?
Rent. The Eater piece from twenty eighteen — and this is one of the best things written on this — they traced what happened around Piazza Navona in Rome over about fifteen years. When the tourist numbers climbed, landlords realized they could charge more. Why rent to a family-run trattoria doing forty covers a night on thin margins when you can rent to an operator who'll turn tables three times an evening serving pre-made lasagna at eighteen euros a plate?
And the family place can't compete on revenue per square foot.
Can't even get close. The trattoria has a chef who makes everything from scratch, they buy fresh ingredients, they probably own their recipes. Their margins are maybe eight, ten percent. The tourist trap is running margins closer to thirty percent because the food cost is negligible and the turnover is fast. When the landlord triples the rent — and rents around Piazza Navona did roughly triple over that period — the trattoria is mathematically eliminated. It's not that they failed. It's that the numbers stopped working.
So the rent gradient acts as a filter. It doesn't make anyone serve bad food. It just makes it impossible to serve good food and still pay the lease.
Unless you own the building. That's one of the few escape hatches, and we'll get to that. But first, mechanism three, which is where it gets really interesting. Once an area has tipped — once the cascade is running and the rents have filtered out the local places — the remaining businesses face a new problem. They have to signal to tourists that they're open for business. And tourists don't just walk into any door. They're anxious. They're in a foreign country, they don't speak the language, they're afraid of ordering something they won't like, they're afraid of being ripped off, they're afraid of looking stupid.
So the six-language menu with the flags isn't just tacky decoration.
It's a risk-reduction tool. Every flag on that menu is saying "we speak your language, we know what you expect, nothing here will surprise you." The sidewalk greeter — the guy with the laminated photos — he's not just hustling. He's performing approachability. He's closing the gap between a nervous tourist and an unfamiliar doorway. The problem is that these signals become mandatory. A restaurant that refuses to display them becomes invisible. Tourists walk past it because it doesn't look like it's for them.
Even if the food is better.
Especially if the food is better. The signals that attract tourists are the same signals that warn off locals. A place with no English menu, no photos, no greeter — that's where the locals eat, because those signals say "this place doesn't care about tourists." And the tourists interpret the same signals as "this place doesn't want me here."
So the signaling arms race is self-reinforcing. The more tourist-coded a street becomes, the more every business on it has to adopt the same signals to survive, which makes the street even more tourist-coded.
And the few holdouts — the bakery that doesn't advertise, the wine bar with the hand-written Italian-only menu — they survive because they've got something the system can't touch. Usually property ownership, or a local clientele that predates the tourism wave and isn't going anywhere.
That's the Piazza Navona pattern. The Eater piece mentioned holdouts.
A bakery that's been there for forty years, owns the building, serves the same neighborhood families who've been coming for three generations. They don't need a single tourist to walk through the door. The rent gradient can't touch them because they're not paying rent. And the signaling arms race doesn't matter because they're not competing for tourist attention. They're playing a completely different game.
So those three mechanisms — the information cascade, the rent gradient, the signaling arms race — they don't operate in sequence. They feed each other.
They're interlocks. The cascade drives foot traffic. Foot traffic drives rents. Rents filter out anyone who can't monetize tourist volume. The survivors adopt tourist signals. The signals attract more tourists. And around it goes.
Which means by the time a street looks like a tourist trap, it's been a tourist trap for years. The visible signals are the last thing to arrive, not the first.
That's the thing most people get wrong. They think the laminated menu and the greeter are the cause. They're the symptom. The cause is a thousand individual decisions made by tourists, landlords, platform algorithms, and business owners, each one rational in isolation, all of them producing a collective outcome nobody wanted.
Including the tourists.
Especially the tourists. Nobody travels to Rome hoping to eat mediocre pasta. They end up there because the system funnels them there.
So let's talk about what happens next. Daniel's question wasn't just how these ecosystems form — it was whether they can go the other direction. Can a tourist trap street un-trap itself?
The short answer is yes, but only when something breaks one of the interlock mechanisms. And the most interesting reversals have come from changes to the information layer, because that's the easiest part of the system to disrupt.
Give me an example.
Google Maps in twenty twenty-two updated their ranking algorithm. They started deprioritizing raw review count in favor of recency and what they called authenticity signals — things like whether reviews came from people who'd actually been to the place recently, whether the reviewer had a history of reviewing similar venues, whether the review text showed signs of being genuine rather than copy-pasted. The effect wasn't huge, but it was real. Places that had been coasting on old, high-volume reviews started slipping. Places with fewer but more recent, more detailed reviews started rising.
So the cascade can be interrupted at the platform level.
It can be. But platform changes are slow and corporate and not exactly designed to solve the tourist trap problem. The more interesting disruption is coming from the demand side. Daniel's hobby — seeking out the worst-rated spots — is part of a broader subculture of travelers who are deliberately inverting the usual signals.
The anti-tourist influencers.
Right. You've got TikTok and YouTube creators whose whole thing is going to the lowest-rated restaurant in a city and filming the experience. Some of them have real audiences. And what they're doing, whether they know it or not, is creating a counter-signal. They're telling the algorithm "low rating plus high volume is interesting." They're training a small but growing number of travelers to read one-star reviews as a recommendation rather than a warning.
Which is a form of information arbitrage. If everyone else is filtering for four stars and above, the three-star zone is underpriced attention.
And the food isn't necessarily worse. That's the part that keeps getting missed. The BBC piece made this point explicitly: tourists rate down restaurants that don't meet their expectations of familiarity. A place that serves tripe or sea urchin or a pasta shape they've never seen — that gets three stars from tourists, not because it's bad, but because it's unfamiliar. The same restaurant might be a local institution. The rating isn't measuring quality. It's measuring how much the food resembles what the tourist expected to eat.
So the one-star review circuit isn't just contrarianism. It's a way of reading against the grain of the platform's incentives. You're using the rating system to find the places the rating system is worst at evaluating.
And if enough people do it, the cascade starts to weaken. Not collapse — the rent gradient and the signaling arms race are still there — but the information layer gets noisier, and that noise creates space for different kinds of businesses to survive.
Let's talk about the businesses that resist without needing a counter-culture to save them. You mentioned property ownership.
That's the big one. If you own your building, you're immune to the rent gradient. You can run a forty-seat trattoria with eight percent margins forever, because your costs are fixed and you're not competing with a tourist trap operator who'll pay triple the lease. The NYT ran a piece last year on avoiding tourist traps, and the one factor that kept coming up in every city they looked at was property ownership. The places that survived touristification were almost always the ones that owned their walls.
What else?
Local density. If your neighborhood still has a residential population — not just Airbnbs, but actual people who live there and need somewhere to eat on a Tuesday night — you've got a fallback customer base. Tourists are gravy, not survival. The places that get completely captured are the ones in areas where the residential base has been hollowed out by short-term rentals. No locals, no local business.
And the third factor?
Identity capital. This is the intangible one. A restaurant with a forty-year reputation, a James Beard award, a chef who's been written up in every food magazine — that place doesn't need to signal to tourists because its name is the signal. Tourists who care about food will find it regardless of whether there's an English menu in the window. Trattoria Da Cesare al Casaletto in Rome is the classic example. No English menu, no website, four point eight stars from a fraction of the reviews of the tourist places, but every single review is from someone who knew what they were looking for.
So identity capital acts as a filter in the other direction. It selects for the tourists who are already informed.
And it repels the ones who aren't. Which is exactly what you want if you're trying to run a serious restaurant in a tourist zone. You don't want ten thousand walk-ins. You want forty people a night who booked three weeks in advance and will pay whatever you charge because they came for your food, not your proximity to a landmark.
There's a knock-on effect here that I think is worth pulling on. You mentioned that a single tourist trap degrades the reputation of the whole street. The NYT piece called it a contagion of mediocrity.
Yeah, this is the negative externality problem. A good restaurant two doors down from a tourist trap gets lumped into the same mental bucket. Travelers look at a street with five laminated menus and one unmarked doorway and they think "tourist area, avoid." The good restaurant gets punished for its neighbors' choices.
And the punishment is algorithmic as well as psychological. If the whole street gets tagged as touristy in enough reviews, the platform starts downranking everything on it.
Which accelerates the homogenization. The good restaurant either adopts tourist signals to survive — at which point it stops being good — or it dies. The contagion metaphor is a bit dramatic, but the mechanism is real. Mediocrity spreads because the presence of mediocrity changes the conditions for everyone nearby.
So we've got formation, we've got resistance, we've got the possibility of reversal. Let me ask you the thing Daniel's prompt is really getting at, which is the appeal. Why does seeking out the worst-rated experience feel like a more authentic way to travel?
I think there are two things happening. One is the obvious contrarian pleasure — you're doing something most tourists aren't doing, and that feels clever. But the deeper thing is that the worst-rated spots are often the most legible. A three-star restaurant with five hundred reviews tells you exactly what it is. There's no marketing, no curation, no performance. The one-star reviews are a documentary of everything that went wrong, and reading them is like reading the unedited transcript of a hundred strangers' bad decisions.
It's honest in a way that a five-star, professionally photographed, influencer-endorsed restaurant isn't.
The five-star place is selling you a story about yourself. "You are the kind of person who discovers hidden gems." The two-star place isn't selling anything. It's just there, failing publicly, and there's something almost endearing about that.
Almost.
I said almost. I'm not eating the lasagna.
The other thing Daniel's getting at — and I think this is the part that doesn't get said enough — is that the worst-rated circuit is a way of seeing the city beneath the city. The Rome that exists for Romans, not the Rome that exists for the ten thousand people who walk through Piazza Navona every day. When you deliberately go to the place with no English menu and a three-point-two rating, you're not just eating. You're crossing a boundary.
And the boundary is real. The tourist economy and the local economy are not the same economy. They occupy the same physical space, but they're serving different people with different expectations, different budgets, different relationships to the food. A restaurant that serves locals is part of a community. A restaurant that serves tourists is part of a supply chain. The worst-rated circuit is a way of finding the community businesses that got caught in the supply chain's net.
By accident.
By being in the wrong place when the tourism wave hit. Or by refusing to change when everyone else did. Either way, they're not playing the game, and the rating system punishes them for it.
What happens when the information landscape shifts again? You mentioned the Google Maps update. You mentioned the anti-tourist influencers. But we're recording this in August twenty twenty-six, and there's a new variable that wasn't in any of those articles.
AI-generated reviews.
Yeah. If every review can be faked — and the tools to generate plausible, human-sounding reviews at scale are getting better every month — what happens to the information cascade? The whole system depends on reviews being a somewhat reliable signal. If the signal goes to zero...
Then tourists revert to older heuristics. Word of mouth. Guidebooks. Physical queues. The BBC piece actually flagged this as a possible future — that as trust in platforms erodes, people go back to the signals that predate the platforms. A line of locals outside a door. A recommendation from a friend who's been there. A restaurant that's been in the same family for three generations and doesn't need to advertise.
Which would favor exactly the kinds of businesses that are currently getting crushed by the tourist trap ecosystem.
It might. If the information cascade breaks badly enough, the advantage flips. The place with two thousand reviews starts to look suspicious, because who's to say any of those reviews are real? The place with forty reviews from accounts that look like actual humans starts to look trustworthy. And the place with no online presence at all but a line out the door at eight p.m. — that place doesn't need reviews.
The line is the review.
The line has always been the review. We just forgot for a while.
Hilbert: They didn't forget.
Sorry?
Hilbert: The tourists. They didn't forget. They never learned. Most of them, this is their first time in Rome. They don't know what a good line looks like. They don't know what the food is supposed to taste like. They're making it up as they go, and the platform is the only thing they've got. That's not a failure of judgment. That's what being a tourist is.
Hilbert: I did six months in Venice in two thousand nine. Menu translator. That was the job title. A restaurant group had four places around San Marco, and my job was to take the Italian menu, run it through Google Translate, and then fix the output so it sounded appetizing in English, French, German, Japanese. I was the guy adding the flags.
Hilbert: The thing you missed — and you got close, but you missed it — is that the multilingual menu isn't just a signal. It's a promise. The promise is: nothing here will surprise you. Tourists are terrified of surprise. They're terrified of ordering something they can't identify, something that arrives looking wrong, something they paid thirty euros for and can't eat. The six-language menu is a contract. It says we will give you exactly what you expect, in a form you recognize, at a price you can see before you commit. That's not a trap. That's what they want.
Hilbert: The places Daniel goes looking for — the worst-rated, the no-English-menu, the locals-only — those places break the contract on purpose. They serve food that is unfamiliar. Tripe. Lampredotto. Whatever the chef felt like making that morning. Tourists hate that. They rate it one star because it violated the promise. The promise was never made, but they thought it was. They thought every restaurant in Italy was there to serve them.
The rating is measuring gap between expectation and delivery, not quality.
Hilbert: Always has been.
Hilbert: The restaurant group I worked for had a secret menu. Same kitchen, same ingredients, same chef. One menu for tourists — the six-language thing with the photos. Another menu for locals — Italian only, different dishes, never offered to anyone who walked in speaking English. The tourist carbonara was fine. The local carbonara was the real thing. Same guy made both. He knew what each customer wanted, and he gave it to them.
Hilbert: The line between trap and authentic — it's not where you think it is. The same kitchen can be both. The same chef can be both. The only difference is which menu you get handed.
That's... I hadn't thought about it that way. The kitchen isn't the variable. The customer is.
Hilbert: The kitchen adapts. Always has. You serve what people will pay for. If they'll pay for familiarity, you serve familiarity. If they'll pay for tripe, you serve tripe. Most places near landmarks, the tourists outnumber the locals twenty to one. What do you think the kitchen is going to optimize for?
The tourist trap isn't a failed restaurant. It's a restaurant that succeeded at reading its market.
Hilbert: Succeeded at reading its market and failed at being interesting. Both things can be true.
Hilbert: The group I worked for closed three of the four places by twenty thirteen. The one that survived was the one farthest from the square, the one that accidentally got a local following because it was on a street Venetians actually used. They dropped the tourist menu after a while. Didn't need it anymore. The locals came for the real food, and the tourists who found it were the kind who wanted the real food. The cascade ran the other way.
Hilbert: Took about four years.
Reversal is possible. It just takes time and a shift in who's walking through the door.
Hilbert: A landlord who doesn't triple the rent the moment he smells money.
The rent gradient is the thing that keeps coming back.
Hilbert: Always does. Information changes fast. Rents change slow. You can fix the reviews and still lose the restaurant if the landlord wants thirty percent more next year.
Hilbert: Anyway. That's what I had.
The secret menu detail changes how I think about the whole signaling arms race. It's not that tourist trap restaurants are incapable of making good food. It's that they've decided it's not what their customers are paying for.
Which makes the worst-rated circuit even more interesting. Daniel's not just finding bad restaurants. He's finding restaurants that might be serving the local menu to the locals and the tourist menu to him — and the rating is the result of other tourists who got the tourist menu and didn't know there was another one.
The question is how you get handed the right menu.
I think you walk in and don't speak English.
Or you walk in and order the thing you can't pronounce. The thing without a photo. The thing the waiter raises an eyebrow at.
The thing that might be tripe.
The thing that might be tripe. Exactly.
Here's the misconception I want to land on. People think tourist traps are bad restaurants that trick tourists. They're not. They're rational businesses responding to a market where the customers don't know what good looks like, can't tell the difference, and won't be back either way. The trap isn't a scam. It's an equilibrium.
The corollary: the worst-rated spots aren't the worst food. They're the worst at managing tourist expectations. Which is a completely different thing, and often a sign that you're about to eat something interesting.
The open question I'm left with — and this is the thing to watch over the next few years — is whether AI-generated reviews break the information cascade badly enough to reset the whole system. If tourists can't trust the platforms, they go back to the signals that predate the platforms. Word of mouth, physical queues, the restaurant that's been there for forty years. Those signals favor the holdouts, the resisters, the places that never needed tourists in the first place.
Which would be a strange kind of justice. The information layer that created the tourist trap gets so polluted that it ends up destroying the tourist trap. The holdouts inherit the street.
Daniel's hobby might turn out to be ahead of its time.
As always. Thanks to Hilbert Flumingtop for producing, and for the secret menu story that I'm going to be thinking about every time I walk past a six-language menu from now on.
This has been My Weird Prompts. If you want to hear more about the hidden economics of places you're told to avoid, our episode on the one-star review circuit goes deeper into the methodology. And if you've got a weird prompt for us, send it to show at my weird prompts dot com.
We'll be back soon.
