It is the 78th minute, the bar is three-deep, and someone has just put another round on the tab — the fourth? the fifth? — for a group that started at four people and is now somehow seven. Ninety minutes later the bartender slides over a single number: $312. Nobody ordered “$312.” Everybody ordered a drink at a time, on a tab no one was reading. Now you have to split it.
That moment is when a bar tab shows what it actually is: not a bill you split, but a bill built to resist splitting. A dinner check arrives once, itemized, early enough that you can still do arithmetic. A running tab inverts all three of those. It accumulates for hours, it hides its total behind the bar, and it itemizes only at the end — after the night has quietly dismantled everyone’s ability to use the itemization. The reasons it’s hard are structural and chemical, not a failure of anyone’s honesty.
There is a tell in the shape this bill takes. When a tab finally prints, it lands as a column of drinks itemized all at once — the night’s only written record, produced at the exact moment the people who ordered can no longer attribute it. That printed receipt is the artifact splitty is built to read: it works from what the bill prints, not from the table’s memory of the night. Everything below is one question seen from that vantage — why this particular receipt fights back, and what actually divides it.
It is not an abstract problem for us. Ten of the twelve most-ordered items across splitty’s own US-leaning receipts are drinks — the single most-ordered cocktail, the espresso martini, ranks second among all items, ahead of every food on the list. A tab is mostly drinks, and drinks are exactly what get ordered one at a time, all night long.
Source: splitty first-party data — most-ordered items across splitty’s US-leaning scanned receipts (2026).
What makes a bar tab different from a dinner check?
A dinner check is closed and itemized; a bar tab is open and accumulating. At dinner you order once, the kitchen prints a line for every dish, and the check lands while you can still read it. A tab is the reverse on every axis. You order over hours instead of in one sitting. The total stays hidden behind the bar instead of arriving on paper. And the receipt itemizes only at the very end — after the drinks it paid for have degraded your ability to make sense of it. The dinner check is a snapshot taken while you’re sober and itemized by the kitchen. The tab is a four-hour exposure you only develop at last call.
Those three differences — time, visibility, and state — are why every tactic that works on a dinner check fails on a tab. You can’t “just remember what you had” across four hours. You can’t eyeball a total you were never shown. And you can’t do clean arithmetic in the condition the tab itself put you in.
Why does buying rounds make a tab so hard to split?
Because a round replaces accounting with a promise. The instant someone says “I’ll get this one,” a set of individual purchases becomes a reciprocity ledger: I get this round, you get the next, nobody itemizes. Sima Riazi and Sarah MacLean, who interviewed 60 young adults in Melbourne in 2016, found that round-buying is understood as “an important way of demonstrating connections with friends” — a bonding ritual, not a payment method. The whole point of a round is that you don’t count. Counting would be missing the gesture.
That works until it doesn’t, and the breaking point is specific. Riazi and MacLean found that “round buying was less likely to occur in groups larger than 4–5 people, as beyond this participants found it difficult to control the reciprocity of round buying, cost, and consumption of alcohol.” Above five drinkers, the ledger stops closing: you buy a round of seven, two of them never buy back, and the reciprocity that made it feel fair silently inverts. Worse, the round itself inflates the number you’ll have to divide. “Overwhelmingly,” the authors reported, participants believed round buying “increases their overall alcohol intake through implicit and explicit pressures to consume more.”
The round does three things at once: it erases the itemization (nobody’s tracking), it only stays fair in small groups (4–5 people), and it drives everyone to drink more than they would alone — so the tab you eventually split is both bigger and blurrier because of how it was bought.
Why can’t anyone remember what they ordered?
Because alcohol attacks memory directly, and it gets worse the more you drink. Jacob Orquin, Heine Jeppesen, Joachim Scholderer, and Curtis Haugtvedt tracked attention and recall under intoxication in a 2014 study and found a steep drop: “consuming one or two beers reduces brand recall from 40 to 36% while being heavily intoxicated further reduces brand recall to 17%.” They measured recall for advertised brands, not bar orders, but the underlying mechanism is general — alcohol produces “a narrowing of visual attention to the most salient features…shallow processing of conceptual information…and memory loss.”
Apply that to a tab. The only place the itemization lives while the tab is open is inside people’s heads — who had the IPA, who switched to cocktails, who only drank the first round. That record is exactly the “conceptual information” alcohol processes shallowly and then loses. By last call, the data you’d need to divide the tab fairly has been erased from the only storage it ever had. The honest reconstruction at the end of the night isn’t lying; it’s genuinely gone.
Source: Jacob L. Orquin et al., “Attention to advertising and memory for brands under alcohol intoxication,” Frontiers in Psychology (2014).
Why does the running total disappear as the night goes on?
Because alcohol narrows your attention to whatever is immediate, and a fresh drink is immediate in a way a four-hour total never is. Claude Steele and Robert Josephs named this effect alcohol myopia in 1990: a “state of shortsightedness in which the immediate aspects of the experience have a disproportionate effect on behaviour.” Their own phrase for it is missing the forest and only dimly seeing the trees. The drink in front of you is a tree, vivid and present. The accumulating tab is the forest — abstract, deferred, and therefore exactly the kind of cue alcohol pushes out of view.
This is why “keep an eye on the total” is advice that cannot survive contact with the night. You keep ordering against a number you have structurally stopped perceiving. The tab doesn’t run away from you because you’re careless; it runs away because the same chemistry that makes the next round appealing also removes the total from your field of vision. The bigger the tree, the dimmer the forest.
Source: Claude M. Steele & Robert A. Josephs, “Alcohol Myopia: Its Prized and Dangerous Effects,” American Psychologist (1990).
Does drinking actually change how you split?
Yes — it makes you more impatient and less generous toward the very people you’re splitting with, right when fairness needs patience and generosity most. Luca Corazzini, Antonio Filippin, and Paolo Vanin ran a placebo-controlled experiment at roughly the legal driving limit and found that “alcohol intoxication increases impatience and makes subjects less altruistic.” Notably, it did not make people bigger risk-takers — the robust effects were impatience and reduced altruism, not recklessness. Translated to a tab: you want the next round now, and you care a little less about whether the split lands fairly on everyone else.
Frank Schilbach’s 2019 field experiment with 229 workers shows how deep the money effect runs. Simply increasing sobriety “raised savings by 50 percent,” and the majority of participants paid real money to commit themselves to staying sober — revealed demand for a check on their own intoxicated financial decisions. People know, on some level, that drinking erodes money discipline, and they’ll buy a guardrail against it.
Impairment scales with the bill.
Every drink that grows the tab also shrinks the table's collective ability to divide it. By the time the tab is biggest and hardest to split, the group is least equipped to do it — drunker, more impatient, less generous, with the itemization already erased. The difficulty and the size rise together, which is the opposite of how a dinner check behaves.
Sources: Corazzini, Filippin & Vanin, PLOS ONE (2015); Frank Schilbach, “Alcohol and Self-Control,” American Economic Review (2019).
Who actually overpays when you split a bar tab evenly?
The light drinker, the early leaver, and the round-skipper — every time. An even split isn’t a neutral default; it’s a transfer from the people who consumed less to the people who consumed more, and a tab maximizes the gap because consumption and time both vary so widely. Take that $312 tab across six people. Split evenly, it’s $52 each. But the friend who had two beers and left at halftime drank maybe $16 of it, while the one who stayed to close and found their way to the espresso martinis — the most-ordered cocktail on splitty’s own receipts, and one of the priciest lines on any tab — drank closer to $80.
Source: splitty first-party data — most-ordered items across splitty’s US-leaning scanned receipts (2026).
That $36 doesn’t vanish — it moves from the early leaver to the closer. Do it every weekend and the lightest drinker is quietly subsidizing the heaviest, which is the same math that makes the sober friend pay for everyone’s drinks and makes cocktails the most unfair line on the bill. The even split feels generous in the moment. It’s generous with someone else’s money.
How do you split a bar tab fairly?
Decide the method before the first round, keep the tab on one card, capture each round as it lands, and itemize from the printed receipt before anyone leaves. The fair split itself is the same as anywhere — each drink to whoever drank it, shared things and tax and tip in proportion. What’s different about a tab is that time and memory are working against you, so the job is front-loaded into a few habits rather than left to end-of-night arithmetic you’ll be in no state to do.
Agree the rule before round one
“By what we drink” or “even, we’re all matching” — decided sober, it’s a ten-second conversation. Decided at close, it’s an argument, because by then everyone is myopic and impatient. Set the rule while judgment is still cheap.
Put the whole tab on one card
One open tab beats six separate ones: the bar prefers it, and it gives you a single itemized receipt at the end instead of a scattered mess. Whoever holds the card is the banker, not the payer — everyone settles up with them later.
Capture each round as it happens
A quick photo of the round, or a note in your phone, is the memory the night is about to erase. This is the step that beats the recall problem: write it down while you still can, because by last call the “who had what” data is gone from everyone’s heads.
Itemize from the printed tab, not from memory
When the tab closes, the receipt lists every drink — the record your heads no longer hold. Assign each drink to whoever had it, then split shared bottles, snacks, tax, and tip in proportion to what each person ran up. The paper is the source of truth, not the table’s recollection.
Settle before anyone leaves
The early leaver is both the most expensive person to chase and the one an even split overcharges most. Square up while the group — and the receipt — is still in one place. A debt collected at the table gets paid; a debt mailed the next morning often doesn’t.
When is splitting a bar tab evenly actually fine?
When everyone drank about the same, stayed about as long, and the group is small enough that nobody’s tracking — split it evenly and move on. The even split is unfair only in proportion to the variance at the table. Four friends who matched each other pint for pint and walked out together have almost no variance, and itemizing their tab would be pedantic, not fair. The even split is the right tool there: fast, friction-free, and close enough to accurate that the difference isn’t worth a conversation.
The decision rule: low variance at the table (everyone drank similarly, nobody left early) → split evenly and move on. High variance (a sober friend, a late arrival, someone who found the cocktail menu at 11pm, an early leaver) → split by what each person had. The mistake is using one method for both.
How does splitty handle a bar tab?
splitty can’t watch an open tab fill up — it isn’t the bartender’s register, and it doesn’t keep a running IOU over the night. What it does is take the one artifact the evening finally produces, the printed itemized tab, and turn it back into a fair, per-person split in the state you’re actually in by closing time: scan it, and every drink on it becomes something you can assign.
That is the whole design bet, and it falls out of the thesis: by closing time the receipt is the only sober witness at the table — the one record the night didn’t erase. So splitty trusts the paper instead of the people, which is exactly backwards from how the table wants to do it and exactly why it works.
The honest limit: if the bar prints a single lump-sum line with no items, there’s nothing to itemize — splitty reads what’s on the receipt, it doesn’t invent line items. That’s exactly the night that step 3 saves: the rounds you photographed are the itemization the receipt didn’t give you.
FAQ
Frequently asked questions
01 Should the heavy drinker pay more on a bar tab?
If the goal is fairness, yes. An even split charges the person who had two beers the same as the person who ran cocktails until close, which moves money from the light drinker to the heavy one. Splitting by what each person actually drank — drinks assigned to whoever had them, with shared items, tax, and tip in proportion — is the fair version. The exception is a low-variance table where everyone drank about the same; there, an even split is close enough that itemizing isn't worth the friction.
02 Is it rude to skip buying a round?
It depends on what kind of group you're in. Riazi and MacLean (2016) found round-buying is treated as a way of demonstrating connection, not just a payment method, so opting out can read as opting out of the bonding gesture rather than just saving money. But the same research found the system breaks down above 4-5 people, where reciprocity stops closing fairly anyway. In a big group, it's reasonable to skip rounds and pay for what you drink — that's often the fairer arrangement, not the rude one.
03 How do you split a bar tab when someone leaves early?
Settle with them before they go. The early leaver is the most expensive person to chase later and the one an even split overcharges most, because they consumed less and for less time. Either total what they actually had and collect it at the table, or — if the tab is still open — note their drinks so their share can be pulled out when it closes. Squaring up while the group is still together is the single most reliable way to avoid an unpaid share.
04 Should you split a bar tab evenly or by what everyone drank?
Match the method to the variance. Low variance — everyone drank about the same and stayed about as long — split evenly, it's fast and close enough to fair. High variance — a sober friend, a late arrival, a heavy closer, an early leaver — split by what each person had. The even split is only unfair in proportion to the variance at the table, so the skill is reading that spread, not defaulting to one method for every night.
05 How do you keep track of a big group's bar tab?
Put the whole thing on one card and capture each round as it happens with a quick photo or phone note. One open tab gives you a single itemized receipt at the end instead of a scattered mess, and the photos preserve the who-had-what data that alcohol erases over the night. When the tab closes, itemize from the printed receipt rather than the table's memory — the paper is accurate when nobody at the table is.
