The Cocktail Problem: Splitting a Bill With Drinks

Who actually pays for the cocktails when you split a bill evenly?

Everyone does — including the people who didn’t drink. An even split takes the most lopsided line on the bill and smears it flat across the table, so the few who ordered cocktails pay less than they should and everyone else covers the difference. The bigger the drinks order, the bigger that hidden transfer.

Put real numbers on it. Six friends, a $224 dinner before tax and tip. Two of them each had two espresso martinis at $14 apiece — $56 of the bill is their drinks alone. The other four skipped the cocktails; one of them just drank water. The food, near enough, was even at about $28 a head. Now split the whole thing six ways:

The water-drinker hands over a $9 surcharge for drinks they never touched, and each cocktail drinker pockets about $19. Four people quietly chip in to discount two people’s espresso martinis. Add tax and a tip on top — both calculated on the inflated total — and the gap only widens, because the water-drinker is now tipping on someone else’s vodka.

Why are cocktails the biggest source of unfairness on the bill?

Because the cocktail is the single highest-variance line on the menu: it is expensive, it is optional, and it lands on some people and not others. VinePair, reporting on-premise data, put the average cocktail at $12 a drink after a $1 rise; the $14 craft cocktail used throughout this article is the worked example Evergreen prices out, not a measured craft average — Evergreen prices a neat pour higher still, at $18. Either way a cocktail is among the priciest single things a person orders, so the spread between the heaviest and lightest drinker is wider than for almost anything else on the table. Food everyone shares; cocktails split the room.

It is also the item the menu is built to sell. A $14 craft cocktail costs the bar roughly $2 in ingredients, which works out to about an 85% gross margin — against the 30–40% a kitchen makes on food — though the two are measured differently, so read the gap as a wide one rather than an exact ratio. Whatever drives the sequencing, the pattern at the table is familiar: the drinks list arrives first, the server asks about another round, and the easiest yes on the table is also the one that does the most damage to an even split.

splitty’s receipts aren’t an outlier here, either. Independent on-premise data from CGA by NielsenIQ, reported by VinePair, put the espresso martini in the top 10 cocktails in the United States as of that tracker — a few years old now — though nationally the margarita still outsells it by a wide margin. Where splitty’s tables differ is degree: on these particular receipts the espresso martini ranks third among all items, which says as much about who reaches for splitty as it does about America. Two snapshots are not a trend line, and neither source tracks where this goes next. What both do show is the thing that matters for a split: the cocktail is on the table, and it is not cheap.

Sources: Evergreen, “How Much Profit Does a Bar Make Per Drink? The Complete 2025 Analysis” (2025) — margins are gross, before labor and overhead; VinePair, “The Espresso Martini Is One of the Top 10 Cocktails in U.S.” (citing CGA by NielsenIQ on-premise data).

Why do cocktails keep getting ordered after the bill is already big?

One likely reason: the next drink is the most vivid thing in the room, and the bill is the least. Psychologists Claude Steele and Robert Josephs called this alcohol myopia: alcohol impairs perception and thought in a way that makes the immediate, salient cue loom large while the distant one fades. The espresso martini on the menu is immediate and appealing. The total it adds to a check you won’t see for another hour is exactly the kind of distant consequence the model says drinkers stop attending to.

So “one more” can keep winning without anyone being reckless. Nobody has run that sequence as an experiment on restaurant drink orders — but on the myopia account, at no point in it is the eventual bill the loud thing in the room. By the time the check lands, the drinks have stacked up — and the person who paced themselves on water watched the whole thing happen stone-cold sober.

Source: Steele, C. M., & Josephs, R. A. (1990). “Alcohol Myopia: Its Prized and Dangerous Effects.” American Psychologist, 45(8).

Isn’t this just the diner’s dilemma?

It is — cocktails are the textbook case of it. Natalie Glance and Bernardo Huberman, who wrote up social dilemmas in Scientific American in 1994, set out the “unscrupulous diner’s dilemma” like this: a group agrees to split the check evenly, then each person decides individually whether to order the lobster or the hot dog. Game theory classes it as an n-player prisoner’s dilemma — everyone, acting reasonably, ends up worse off than if they had all held back.

The cocktail is the lobster. When the bill is split six ways, the true cost of your $14 espresso martini feels like about two dollars — your one-sixth of it — so ordering it is an easy call. But everyone at the table runs the same math. That is the model’s prediction rather than a measured effect on cocktail orders — but it is the incentive the even split creates. It doesn’t just divide the drinks unfairly; it rewards ordering more of them.

Source: Glance, N. S., & Huberman, B. A. (1994). “The Dynamics of Social Dilemmas.” Scientific American.

When is splitting drinks evenly actually fine?

When everyone drank about the same, splitting evenly is the fair answer — and pulling out a calculator to itemize two-dollar differences would just be rude. The cocktail problem only shows up when the drinks are both expensive and unevenly ordered. Use that as your test, not a blanket rule.

Doesn’t covering more of the tip even it out?

Only by accident — and it is worth knowing why, because this is the fix people actually reach for. Nobody wants to itemize a receipt in front of friends, so the cocktail drinker offers the socially graceful version instead: I’ll get the tip. It sounds like restitution. It is priced off a number that has nothing to do with the debt, so sometimes it undershoots, sometimes it overshoots, and nobody at the table can tell which.

Tipping extra sends the money to the server, not the table

Start with the version that transfers nothing at all. If you leave a bigger tip and the table still splits the total evenly, the extra money goes to the server. The water-drinker’s share does not fall — it rises, because they are now splitting a larger tip too. Generosity to the restaurant is not compensation to the table.

Covering the tip transfers real money — just not the right amount

The version that does move money between diners is “I’ll cover the tip, you split the rest.” That is a real payment from the drinkers to everyone else. The problem is its size. It is a percentage of what the whole table ate, and it has nothing to do with what anyone drank.

Back to the six friends and the $224 dinner. The two espresso-martini drinkers ran up $56 of cocktails, so each of the other four is carrying $9.33 they did not order — that is just $56 divided by six. Now tip 19.4% — the average on card and digital payments at US full-service restaurants in the first quarter of 2025, per LendingTree’s read of Toast platform data. That is $43.46 on a $224 check, and a tip the drinkers absorb is a tip nobody else has to split — so each of the four non-drinkers keeps their sixth of it, $43.46 ÷ 6 = $7.24. That covers about 78% of the $9.33 debt. Close enough that everyone goes home feeling square.

Now change one thing that has nothing to do with the drinks — how much food everyone else ordered.

Illustrative arithmetic on this article’s running example — six diners, $56 of cocktails on two of them, food divided evenly, tip applied to the whole check. Every figure in the table is the same two steps: 19.4% of the check divided by six, set against $56 divided by six. Tip rate: LendingTree, “Full-Service Restaurant Tips With a Card Average Nearly 20%” (2025) — 19.4% at full-service restaurants paid by card or digital payment, Q1 2025.

Same drinkers, same drinks, same gesture — and the correction swings from about half the debt to more than all of it. The lever moving it is the food bill, which the cocktail drinkers neither ordered nor control. On the big-dinner night the water-drinker quietly comes out ahead. On the light night they absorb more than half the loss and are told the matter is settled.

Source: James Konow, “Fair Shares: Accountability and Cognitive Dissonance in Allocation Decisions”, American Economic Review 90(4), 2000.

The tip line already tracks the drinks

There is a second reason the tip line is a poor instrument here: it does not sit still while the drinks arrive. Michael Lynn and Michael Sturman analysed 275 dining occasions recorded by 51 diners and found that meals where the diner drank alcohol carried tip percentages roughly 6 points higher, in a model that also accounted for service ratings, bill size and party size. It is a small, self-reported, correlational study, so it shows an association, not a cause, and 51 people is not the country. But it is the direction worth knowing: on that evidence, drinks may raise the rate the tip is struck at as well as the subtotal it is struck on — and under an even split, the water-drinker is exposed to both.

Source: Michael Lynn & Michael Sturman, “Tipping and Service Quality: A Within-Subjects Analysis”, Journal of Hospitality & Tourism Research (2010).

And nobody at the table can check it

The last problem is that nobody can check it. A proportional split is auditable — everyone can look at the receipt and find their own line. “I got the tip” is an announcement, not a settlement. The person it is meant to compensate never learns whether $7.24 covered a $9.33 debt, because neither number is ever said out loud. Whatever the gesture is worth, it cannot be confirmed by the only person with standing to confirm it — which leaves the imbalance exactly as unexamined as the even split that created it.

If you truly will not itemize, the honest version of the gesture is one subtraction rather than a flourish: work out what the drinks add per head — $56 across six people is $9.33 — and hand that over directly. That is the same arithmetic the tip trick was invented to avoid, which is the tell. Somebody has to price the drinks. The only question is whether you price them on purpose or hope the tip line does it for you.

How do you split a bill fairly when only some people had cocktails?

Pull the drinks out first, then split what’s left. The whole trick is to stop treating a $14 cocktail like it belongs to the table when it belongs to one person. Three moves handle every version of it.

Source: LendingTree, “Full-Service Restaurant Tips With a Card Average Nearly 20%” (2025), drawing on the Toast Restaurant Trends Report.

How does splitty handle the cocktail problem?

splitty is built so the priciest, most lopsided line on the bill lands on the right people without anyone having to argue for it. Scan the receipt and every item starts split across the whole table; from there, the drinks are a few taps, not a negotiation.

That is the whole point: the cocktail is the most expensive thing people forget they’re splitting, so the fix is to make it the easiest thing to un-split. The psychology of why even splits feel unfair runs deeper than drinks, and the same habit scales to whatever the whole bill turns out to be — but the drinks are where it starts, and where it costs the most.