Grand Theft Potato

When restriction makes people want something more, when it doesn't, and what the data says

---

This one made me laugh out loud and I've been handing it to people ever since.

Someone got ethics approval. Someone wrote a protocol. Someone sat in a room and said, out loud, to other adults, "What if we hired professional actors, sat them next to participants, and had the participants steal french fries off their plates?" Elsevier published it in Food Quality and Preference under the title "Stolen fries are spicier than justice."

That's a real title. In a real journal. With a real DOI (Skryabin, 2026).

I pulled the full text, because a forty percent effect size is usually the kind of number that gets repeated at dinner parties for a decade after the study behind it has fallen apart. This one holds up on the numbers, and its problem is somewhere else entirely.

What Skryabin actually did

The study is Valentin Skryabin, working alone, out of the Russian Medical Academy of Continuous Professional Education in Moscow (Food Quality and Preference, volume 143, 2026).

He opens by noting the folklore is everywhere. Italians say il cibo rubato ha più sapore. The Japanese have a word for it, nusumigui, secret eating. Latin American dichos put it as lo prohibido es lo más sabroso. Just about every culture on earth has independently decided that stolen food tastes better, and until now nobody had checked.

He checked. One hundred twenty adults, each of whom went through all four conditions in counterbalanced order:

Legitimate. A portion of fries that's plainly theirs. Eat them.

Gifted. A friendly confederate offers them fries. Accept and eat.

Low-risk theft. Take fries from the friendly confederate's portion while they're distracted.

High-risk theft. Same instruction, but the confederate is a different character entirely: closed posture, flat voice, minimal eye contact, answers in three words. The paper calls this one the Stern Confederate. Two professional actors played the roles, one male and one female, counterbalanced across participants, trained on video with feedback until they hit better than ninety-five percent script adherence. Somebody spent real money coaching a professional actor to be believably unfriendly about potatoes!

Every portion was identical. One hundred grams of fries (plus or minus two) from a single batch, fried at 180 degrees for three and a half minutes, served at 55 degrees in plain white containers. Palate cleansers between rounds. He even confirmed people ate the same NUMBER of fries in every condition, so nobody can argue the stolen ones just came in bigger handfuls.

The numbers

Ratings are on a nine-point scale. Same potato in every column.

Measure	Your own	Gifted	Stolen, low risk	Stolen, high risk
Taste pleasantness	6.09	7.03	7.95	8.48
Overall enjoyment	5.97	7.05	8.00	8.52
Saltiness	5.95	5.99	6.92	8.01
Crispiness	6.02	6.97	8.07	8.43
Taste intensity	6.04	6.97	8.05	8.51

High-risk theft beat legitimate ownership by 39.3 percent on pleasantness. Nearly half the participants (48.3 percent, or 58 of 120) gave stolen fries the maximum possible score of nine. Not one person gave a nine to fries that were legitimately theirs. The distributions don't overlap: your own fries scored between five and seven, stolen ones between eight and nine.

The risk gradient held too. Stealing from the stern actor beat stealing from the friendly one, on the same fry, in the same room, minutes apart.

The gifted column

Being handed fries by a friendly person beat owning them, 7.03 against 6.09. Generosity is worth something! Being handed fries still lost badly to taking them, and it lost worst of all on saltiness, where gifted fries and your own fries came out statistically identical (p = 0.686). A gift moved how much people liked the fries, and saltiness didn't budge until there was risk in it.

Skryabin put the gift condition in on purpose, as a control. He wanted to know whether the effect was about transgression or just about pleasant social contact. Warmth helps, and it moved the ratings about a point. Taking the fries moved them two and a half.

He also checked whether this was a simple halo, people liking everything more because they liked the moment more. Within each condition, the correlation between how pleasant the fries tasted and how salty they seemed was essentially zero, so the sensory channels moved independently rather than drifting up together on mood.

The guilt test

Guilt is the explanation that suggests itself: a little moral discomfort makes pleasure sharper. Skryabin tested that directly and it didn't survive.

Guilt and enjoyment were correlated across theft trials, yes. Once he accounted for which condition a person was in, though, neither guilt nor excitement predicted enjoyment on its own. When he aggregated to one guilt score per person, the association with enjoyment vanished (not statistically reliable). His own read is that guilt tracks the risk manipulation rather than causing the pleasure.

That makes context the thing doing the work: the situation the food arrives in, not the feeling the eater has about it. Context is something you can actually build. Guilt was never the target anyway.

Where the study falls short

His participants didn't steal anything. They were instructed to take fries by a researcher, in a room, as part of a study they'd consented to. A participant following a script isn't breaking a social norm the way a real thief is, and he says so plainly. He calls it "role enactment or playful transgression," and he refuses to claim it's neurologically the same thing as real rule-breaking.

Hold that next to Mann and Ward's 2001 forbidden-fruit work, which Skryabin himself cites. They found that being told to avoid a food raised desire for it, while choosing to avoid it did not, and that neither version produced overeating (Mann & Ward, 2001).

Skryabin has the same structure in reverse. His transgression also came from outside the participant. It was assigned, not chosen. It still moved the ratings almost two and a half points.

Which means the effect may not require sincerity at all. It seems to work when the rule comes from outside the person.

Hungrier participants showed a smaller enhancement, so this is about context and not appetite. Sixty percent of participants figured out roughly what the study was about. The sensitivity analysis split them: the effect came out the same size in people who guessed and people who didn't. Suspicion didn't inflate it.

The supporting research

Boothby, Clark and Bargh (2014), Psychological Science. Identical chocolate was judged more likeable and more flavorful when another person tasted it at the same moment, with no conversation and no comparing notes. When the chocolate was bitter, sharing made it less likeable (Boothby et al., 2014). Sharing AMPLIFIES whatever the experience already is, sweet or bitter.

Fisher and Birch (1999), American Journal of Clinical Nutrition. Restrict a child's access to a palatable food and their behavioral response to it goes up; selection and intake rise too, though the authors are careful to bound that second part to the restricted context (Fisher & Birch, 1999).

Jansen and colleagues (2008), Appetite. They ran it with fruit. Tell a kid they can't have the fruit and they eat more of it later, same as sweets. There was no group difference in how much the kids WANTED the fruit (Jansen et al., 2008). Restricting the fruit moved how much the kids ate without moving how much they said they wanted it. Depending on how your week is going, that's either encouraging or slightly grim.

Mann and Ward (2001), International Journal of Eating Disorders. The population: college women, tested on a food they'd named as a personal favorite.

Raghunathan, Naylor and Hoyer (2006), Journal of Marketing. Stop and say the lead author's name out loud. Raghunathan. RAH-goo-NAH-thun. It takes longer to pronounce than the finding takes to explain, which tells you most of what you need to know about how academic credit gets handed out. The finding is the unhealthy equals tasty intuition: the less healthy a food is portrayed to be, the better people expect it to taste and the more they enjoy actually eating it (Raghunathan et al., 2006). Describe a brownie as indulgent and it beats the identical brownie described as wholesome.

Bratanova and colleagues (2015), Appetite. The mirror image. Food described as ethically produced tastes better, because moral satisfaction raises taste expectations before the first bite. One caveat that matters commercially: the effect held mainly for people who already endorsed the values in question (Bratanova et al., 2015). Ethics framing works on people who care about ethics… which sounds obvious until you watch someone spend a marketing budget forgetting it.

Put all of that next to Skryabin. Taste doesn't appear to be a property of the food so much as a property of the situation. If that's true of a POTATO, which has to be the least sophisticated object in the human diet, it's probably true of everything you sell.

That's six studies and one very odd paper. If you know work that cuts against any of it, or I missed something, let me know.

What this means for four groups

Parents deciding whether to enroll their kid

A parent's read on your program comes from the situation they're standing in when they see it.

Boothby applies here, and the caveat comes with it. A parent watching class alone on a phone in the parking lot is having a measurably different experience from the same parent watching the same class from a row of chairs next to three other parents. Same instructor, same class. Remember the bitter chocolate, though: shared viewing amplifies whatever is already happening on the mat. If the class is good, seating parents together makes it better. If the class is a mess that day, you've just amplified that too.

Mann and Ward is the caution. When a parent chooses a limit ("we do two activities, that's it"), the limit holds and creates no hunger for more. When a limit gets imposed on them, wanting goes up. That's a real lever, and it gets abused.

Owners running the schools

This is the group I see getting it wrong most often, and it's costing them.

Every owner has been told to use urgency. What usually happens? You run a doors-close-at-month-end campaign, and then the doors reopen the next week because a family called and the month came in light. Mann and Ward found that a restriction only moved desire when it came from outside the person. A deadline your market has watched you move isn't coming from outside anymore. It reads as a suggestion, and the second campaign works worse than the first.

The version that works is boring and REAL. Mat space is actually capped. The beginner cycle actually starts on a date. The waitlist is actually a list. Restriction only works when it's true.

Most owners I talk to have the decorative kind and believe they have the real kind. The test is one question: would you honor that deadline if your best family called the day after?

The same mistake runs in the other direction too. In Scarcity Is a Lousy Strategist I went through what happens when an owner accepts a scarcity claim about his own market without ever putting a number on it. Same habit, pointed inward instead of at the customer, and it costs about the same.

The Gifted column ran with no scarcity in it at all, and it still beat ownership by nearly a full point on a nine-point scale. Free trial classes, a genuinely good parent viewing area, the instructor who hands a kid something on the way out. All of it is real, and you can start running any of it tomorrow.

Authors trying to get a book out

Authors have a Raghunathan problem.

Perceived cost signals perceived value, and readers seem to apply it to books the way eaters apply it to brownies. A book that took two years reads as more serious than one that took six weeks, before anyone opens it. Every shortcut a reader can perceive (the generic cover, the thin acknowledgments, the seven chapters that are obviously four chapters) lowers the taste of the book before page one. The situation the book gets read inside is part of what you're producing.

Financial professionals and business owners

Bratanova is yours, caveat included.

Every advisor has watched a client fall for a worse instrument with a better story and concluded the client was being irrational. Their actual experience of the product may genuinely have been better, because the framing arrived before the product did. A cleaner spreadsheet won't correct that. Just don't over-read it: the ethical-framing effect held mainly for people who already carried those values, so this rewards knowing what your client actually cares about.

Where the joke stops

In clinical work, manufactured scarcity isn't a tactic. It's a harm. Someone deciding whether to get help isn't a prospect with a wanting problem, and limited-spots language aimed at that decision is indefensible no matter what the conversion data says. Waitlists in a clinical practice get stated plainly and that's all.

What to do with this

Pick one restriction in your business and make it TRUE. Not a new one. A true one. The enrollment window that actually closes. The cap that's actually a cap. The deadline you'd honor if your best family called the day after. Real restriction moves wanting. Decorative restriction costs you credibility on the second run.

Then improve one situation without touching the product. Where parents sit. How a proposal arrives. What a reader sees before page one. Skryabin moved a potato 2.39 points on a nine-point scale (6.09 to 8.48) without changing one thing about the potato. That's your benchmark. Whatever your version of that is, it's cheaper than your next ad.

One last note, and it's the author's own. In his discussion of future research, Skryabin suggests testing foods more commonly associated with opportunistic theft and singles out cheese, which he notes is widely reported as the world's most-stolen food. That line is sitting in a peer-reviewed journal right now waiting for somebody to build a career on it. If someone eventually publishes on stolen cheese, I want it on record that I saw it coming!

In the meantime, if someone reaches across the table and takes a fry off your plate, let them have it. They're going to enjoy it about forty percent more than you would have, and now you know why.

One true restriction and one better situation will usually get you more growth than your next campaign, and they cost you almost nothing. You can always reach out!

---

References