BAC Calculator

What Is BAC?

BAC is a concentration, not a quantity. That single distinction explains almost everything people find counterintuitive about drinking: why two friends who drank identically test differently, why the same three drinks are unremarkable on one body and dangerous on another, and why nothing you do after the fact changes the number except waiting.

What the percentage actually measures

The figure is grams of ethanol per 100 milliliters of blood, and the percent sign is doing unusual work — it is a weight-per-volume percentage, not a share of your blood. A reading of 0.08% is 0.08 g/100 mL, which the same laboratory might report as 0.08 g/dL or 80 mg/dL. Much of Europe drops the decimals and uses per mille, where the identical sample reads 0.8‰. Whichever notation appears, the underlying quantity is the same, and this calculator prints it in the US convention to three decimal places.

Roadside instruments do not measure blood at all. They measure ethanol in exhaled breath and convert to a blood figure using a fixed blood-to-breath partition ratio — 2,100:1 in the US evidential-testing convention. That ratio is a population average like every other constant on this page, and the variation between individuals around it is one of the standard grounds on which breath results get challenged in court. It is a further reason to treat any single number, from a machine or from arithmetic, as an estimate with a margin.

Why the same drinks produce different numbers

Alcohol dissolves in water and is excluded from fat and bone, so the volume it spreads through is roughly the body's water — and a bigger person has more of it. Set the calculator to zero hours so no metabolism has happened yet, feed it the same four standard drinks, and change nothing but the body:

Identical alcohol, a spread of nearly three to one: 0.076% for the 240-lb man and 0.224% for the 100-lb woman. The 200-lb woman's 0.112% is already over the 0.08% driving limit, and the 100-lb woman's 0.224% falls in the band this calculator labels severe impairment — an aggravated-DUI level in many US states. Matching someone drink for drink is not a fair contest, and this table is the reason.

Ethanol also carries about 7 kcal per gram, which makes one standard drink 98 kcal before any mixer — four is 392 kcal and six is 588 kcal. To see where that fits your day, use the Calorie Calculator.

The legal numbers, and who each one applies to

There is no single American BAC limit. There are four, and which one binds you depends on your age, your license and the state you are in.

The federal commercial rule is stricter than the number alone suggests: 49 CFR 382.201 bars a driver from performing safety-sensitive functions at 0.04 or greater, and 49 CFR 392.5 separately prohibits any measured alcohol concentration at all while operating a commercial vehicle. For drivers under 21, 23 U.S.C. 161 conditions federal highway funds on states treating 0.02% or greater as driving under the influence, which is why every state has a zero-tolerance law.

None of these is a safety threshold. They are the points at which the offense becomes provable without argument about your behavior, which is what "per se" means. Impairment charges below the per se limit exist in every state, and the Utah Highway Safety Office reports an average BAC of 0.14 among DUI arrests in its 2024–25 state fiscal year — nearly three times the state's own limit, and a reminder that the people the law catches are mostly not the ones hovering at the line.

The 0.08% that also defines binge drinking

NIAAA defines binge drinking as a pattern that brings BAC to 0.08% or higher, and notes that for a typical adult this corresponds to five or more drinks for a man, or four or more for a woman, in about two hours. That definition is a BAC statement first and a drink count second — the drink count is just the arithmetic worked backwards for an average body.

This calculator reproduces it almost exactly. Enter a 180-lb man with four drinks over two hours and it returns 0.071%, below the line; the fifth drink takes him to 0.096%, above it. Enter a 160-lb woman with three drinks over two hours and it returns 0.075%; the fourth takes her to 0.110%. NIAAA's high-intensity drinking category, at twice the binge threshold, is where the numbers stop being abstract: ten drinks in two hours puts that same man at 0.222%, and eight puts that woman at 0.251%.

How Do You Calculate BAC?

The Widmark equation is a century old and still the backbone of nearly every BAC estimate published anywhere. It has two moving parts — how much alcohol went in, and how much body there is to dilute it — and one subtraction for the alcohol your liver has already removed.

The Widmark formula, written out

BAC (%) = ( A ÷ ( m × r ) ) × 100 − β × t

A = grams of alcohol   = standard drinks × 14
m = body mass in grams = kilograms × 1,000
r = distribution factor = 0.68 male, 0.55 female
β = elimination rate   = 0.015 percentage points per hour
t = hours since the first drink

Pounds to kilograms: lb × 0.45359237

Each term has a physical meaning worth holding on to. A is the alcohol itself, counted in grams rather than glasses because glasses are not standardized. The product m × r is the part of your body alcohol actually occupies: multiply an 81.6-kg man by 0.68 and you get 55.5 kg of alcohol-distributing tissue, essentially his body water. Divide grams by grams and multiply by 100 and you have a percentage by mass, which is close enough to the mass-per-100-mL convention that the whole field treats them as one.

The final term is the only thing that ever reduces the number. β × t is a straight-line subtraction: the liver clears alcohol at a near-constant rate rather than a proportional one, so BAC falls by the same amount every hour regardless of how high it started. That is why doubling your drinks doubles your peak but far more than doubles your time to sober.

Counting standard drinks correctly

The drinks field is where accuracy is won or lost, and it does not mean glasses. NIAAA defines a US standard drink as any beverage containing 0.6 fluid ounces or 14 grams of pure alcohol — a 12-oz beer at 5% ABV, a 5-oz glass of wine at 12%, or a 1.5-oz shot of spirits at 40%. NIAAA makes the trap explicit: a 12-oz bottle at 10% ABV is two standard drinks, not one. For anything else, the arithmetic is one line.

Standard drinks = ( volume in fl oz × ABV ) ÷ 0.6

Read down the standard-drinks column and the scale of the usual undercount is obvious. A bottle of wine shared between two people is 2.75 standard drinks each, not "a couple of glasses". Outside the US the accounting differs again: the NHS defines one UK unit as 10 mL or 8 g of pure alcohol, so a US standard drink is 1.75 UK units, and the UK's 14-units-a-week guideline is 8 US standard drinks.

Step by step

• Convert your weight to kilograms if you entered pounds: multiply by 0.45359237. A 180-lb man is 81.6466 kg.
• Convert kilograms to grams by multiplying by 1,000, because the alcohol is being counted in grams and the units have to match.
• Multiply that body mass by the distribution factor r — 0.68 for men, 0.55 for women — to get the mass alcohol actually spreads through.
• Count your standard drinks honestly, using the table above for anything that is not a 12-oz beer, a 5-oz wine or a 1.5-oz shot, and multiply by 14 for grams of alcohol.
• Divide the alcohol grams by the distributing mass, then multiply by 100. That is your peak estimate before any metabolism.
• Subtract 0.015 for every hour since the first drink. The calculator never returns a negative number — once the subtraction exceeds the peak it reports 0.000%.

The calculator does all six steps from the raw inputs, so nothing here needs doing by hand. Working through it once is still worth the two minutes, because it shows exactly which step your own uncertainty sits in — and for most people it is step four, not the model.

Worked example: 180-lb man, four drinks, two hours

The canonical case, and the one where the estimate lands in the most dangerous part of the range — under the legal line but well past sober.

Weight    180 lb × 0.45359237     = 81.6466 kg = 81,646.6 g
Alcohol   4 drinks × 14 g          = 56 g
Body      81,646.6 g × 0.68        = 55,519.7 g of distributing mass
Peak      ( 56 ÷ 55,519.7 ) × 100  = 0.10086%
Burn-off  0.015 × 2 h              = 0.030%
BAC       0.10086 − 0.030          = 0.07086%

Result returned by this calculator

0.071% BAC (estimate)

The calculator's own supporting line reads "Widmark: 4 standard drinks (14 g alcohol each) ÷ (82 kg × 0.68) − 0.015%/h × 2 h", and it classifies the result as "Impaired — above the 0.05% limit used in Utah and much of the world". Two facts sit inside that one number: he is under the 0.08% line in 49 states, and he is over the limit that would apply to him in Utah, over the commercial limit by a factor of nearly two, and more than three times the threshold that applies to a driver under 21.

Worked example: 140-lb woman, three drinks, two hours

Fewer drinks, a smaller body, and a lower distribution factor — the combination that puts a woman over the line on a drink count most people would call moderate.

Weight    140 lb × 0.45359237     = 63.503 kg = 63,503 g
Alcohol   3 drinks × 14 g          = 42 g
Body      63,503 g × 0.55          = 34,926.7 g of distributing mass
Peak      ( 42 ÷ 34,926.7 ) × 100  = 0.12025%
Burn-off  0.015 × 2 h              = 0.030%
BAC       0.12025 − 0.030          = 0.09025%

Result returned by this calculator

0.090% BAC (estimate)

Three drinks, two hours, and she is over the 0.08% line the man above cleared on four. One detail worth knowing if you check the tool's working: the supporting line rounds her weight to "64 kg" for display while the calculation uses the exact 63.503 kg. Redo the arithmetic with the rounded 64 and you get 0.089% — a thousandth adrift, and the gap is display rounding, not an error in the estimate.

Where the two constants come from, and how far they move

The distribution factors of 0.68 and 0.55 are Widmark's own, and the 0.015%/h elimination rate is the standard population midpoint. Both are defensible averages and neither is a property of you. Seidl and colleagues measured body water and blood water in 256 women and 273 men at Erlangen-Nürnberg and found the ratio ran from 0.44 to 0.80 in women and from 0.60 to 0.87 in men. Holding the drinks and the hours fixed and moving only r shows what that costs:

The man's plausible range straddles the 0.05% Utah limit; the woman's straddles the 0.08% one. Neither range is exotic — both endpoints came out of a real population sample.

The elimination rate has a similar story. Jones's evidence-based survey in Forensic Science International reviewed the published drinking studies and concluded that the physiological range runs from 10 to 35 mg/100 mL per hour — 0.010% to 0.035% — with 15 mg/100 mL/h remaining a good average for moderate drinkers, the figure this calculator uses. The paper is more specific than that headline: after drinking on an empty stomach the rate tends to fall in the 0.010–0.015% band, in non-fasted subjects 0.015–0.020%, and in alcoholics during detoxification, when the microsomal CYP2E1 pathway is upregulated, 0.025–0.035%. It also notes the decline is slightly steeper in women, tracking liver weight relative to lean body mass.

Put numerically: take a 180-lb man on six drinks four hours in. This calculator returns 0.091%. At the slow end of the measured range he would be at 0.111%, at 0.020%/h at 0.071%, and at the detoxification rate at 0.011%. The same 0.151% peak takes 15.1 hours to clear at 0.010%/h and 4.3 hours at 0.035%/h. Nobody knows their own β, which is precisely why waiting on a calculated clearance time is not a plan.

BAC Chart by Body Weight and Drinks

Every cell below came out of this calculator at three decimal places, exactly as it prints them. All figures assume two hours have passed since the first drink, so 0.030 percentage points of elimination are already subtracted — which is why a light body on one drink can read 0.000%. Find your weight column, read down to your drink count, and treat the cell as the center of a band rather than a measurement.

BAC chart for women, two hours in

Same two hours, same standard drinks, distribution factor 0.55 instead of 0.68:

Compare like with like across the two tables. A 160-lb woman on four drinks reads 0.110%; a 160-lb man on four drinks reads 0.083%. Identical bodies by the scale, a 33% higher estimate, and the whole of the difference is the distribution factor.

Peak chart: the same drinks at zero hours

Set the hours field to 0 and the elimination term disappears, leaving the raw Widmark peak. This is the number the tables above are counting down from, and it is the one that matters if the drinking happened fast.

Two hours of elimination is worth 0.030 percentage points and no more. That is a large fraction of a two-drink evening and a rounding error on a six-drink one — which is the arithmetic behind the advice to slow down rather than to wait it out afterwards.

BAC chart in kilograms

The same tables for readers entering metric weights. Switch the weight unit to kilograms and these are the figures the calculator returns two hours after the first drink.

Alcohol is a diuretic, and the next day's fluid target is worth setting deliberately rather than by thirst — run it through the Water Intake Calculator.

The tool accepts 35 kg to 300 kg, so the metric tables could run further in both directions; these six columns cover the great majority of adults. One standard drink at 70 kg or above has already been fully eliminated by the two-hour mark, which is the only reason those cells read 0.000%.

Metric chart for women

A 45-kg woman reaches 0.140% on three drinks in two hours — a level this calculator labels severe impairment and an aggravated-DUI range in many states. A 110-kg man reaches 0.026% on the same three drinks, and a 95-kg man 0.035%. The tool applies no judgment to either figure; the tables are here so you can see the distance before you need it.

How many drinks reach 0.08% and 0.05%

Read the chart the other way round and it answers the question people actually arrive with. These are the first whole standard drink at which the calculator's two-hour estimate reaches each threshold — not a permission slip, and not a number to aim at.

For a 100-lb woman the same drink crosses both thresholds at once: her second standard drink takes her from 0.026% to 0.082% in one step, skipping the 0.05–0.08 band entirely. That is the practical meaning of a small distribution volume — there is no gentle middle to feel your way through.

Every figure here assumes two hours elapsed and honestly counted 14-gram drinks. Change either assumption and the whole table shifts: at zero hours the 180-lb man's fourth drink is already 0.101%, not 0.071%, and if his "drinks" were 16-oz 7% pints the fourth one is 1.87 standard drinks each and the arithmetic is somewhere else entirely.

How the number falls with time

Elimination is the one term that ever reduces the estimate, and it moves at 0.015 percentage points an hour whatever the starting level. Holding a 180-lb man at five standard drinks and moving only the hours field:

Five drinks, and the estimate is still above 0.08% three hours after the first one. Because the fall is linear rather than proportional, the clearance time scales directly with the peak — which is why the last drink of the night is the expensive one.

Alcohol also shortens and fragments the back half of the night, so if the plan involves an early start, set the wake time deliberately with the Sleep Calculator.

Running several sessions through the calculator at quarter-hour steps gives the timings below. Read them as what this model says, not as a schedule: the same peak clears in less than half the time at the fast end of the measured elimination range and in half again as long at the slow end.

Hours to fall below each threshold

The bottom row is the morning-after case in numbers: a 140-lb woman on six standard drinks is, by this model, still above 0.08% almost eleven hours after her first drink. Stop at midnight and the arithmetic still has her over the limit at breakfast. That is where most next-day DUI arrests come from, and no amount of sleep changes the subtraction.

How to Read Your Result

The calculator returns four lines: an estimate, the arithmetic behind it, an impairment band, and a caveat it prints every single time. Each is doing a different job.

The four lines it gives you

• The headline is the estimate to three decimals, tagged "(estimate)" — for a 180-lb man on four drinks two hours in, "0.071% BAC (estimate)". The third decimal is there because the impairment bands are only 0.03 wide, not because the model is accurate to a thousandth.
• The second line restates the inputs as arithmetic: "Widmark: 4 standard drinks (14 g alcohol each) ÷ (82 kg × 0.68) − 0.015%/h × 2 h". This is where you check that the sex, weight unit and hours went in the way you meant. The kilogram figure is rounded to a whole number for display while the calculation uses the exact conversion.
• The third line is the impairment band, in words rather than numbers — the six bands are listed below. It is the part of the output that is making a claim rather than reporting arithmetic.
• The fourth line never changes and never goes away: real BAC swings widely around any average-body estimate, and this is not a tool for deciding whether to drive. It is printed on every result deliberately, including the ones that read 0.000%.

Two inputs are rejected rather than calculated. A body weight outside 35–300 kg (77.2–661.4 lb) returns "Enter a valid body weight", and a drink count of zero, a negative, or anything above 30 returns "Enter standard drinks (1–30)". Both come back as a 0 with an instruction instead of a number.

The six impairment bands, and what triggers each

The band is assigned from the estimate alone, on fixed thresholds. The right-hand column is the point at which a 180-lb man reaches each one with no time elapsed, to give the ranges a scale you can picture.

The bands are cliffs, not gradients: 0.079% and 0.081% are indistinguishable physiologically and land in different sentences. Read the band as a label for a neighborhood, and remember that the neighborhood itself may be one or two hundredths off given the constant ranges above.

The lowest band is worded the way it is on purpose. NIAAA notes that even small increases in BAC decrease motor coordination and cloud judgment, and the top band is not rhetorical either — NIAAA describes alcohol overdose as the point at which areas of the brain controlling breathing, heart rate and temperature begin to shut down, with the gag reflex among the functions lost. If a calculator is returning a number in that range about a real person, the response is 911, not arithmetic.

Does one drink an hour hold you flat?

It is the most common piece of folk pacing advice, and the calculator settles it. One standard drink adds a fixed amount to your BAC; one hour removes 0.015 percentage points. Whether you hold steady depends entirely on which is bigger.

Every row is positive, so the answer is no for everybody in the table — but the rate of failure differs enormously. Feeding the calculator n drinks at n−1 hours, a 180-lb man drinking one an hour reads 0.025% after his first, 0.056% after his fourth and 0.097% after his eighth, crossing 0.08% on his seventh. A 140-lb woman on the identical schedule reads 0.040% after her first, 0.090% after her third and 0.216% after her eighth. She is over 0.08% at the point where he is still at 0.046%.

The useful version of the rule is the last column. One standard drink takes a 180-lb man about 1 hour 40 minutes to clear and a 100-lb woman about 3 hours 45 minutes. Pace to that and the estimate genuinely stays flat; pace to the folk version and it climbs all night.

What lowers the number and what only feels like it does

Time is the only entry on the first list. The subtraction in the formula is metabolism, almost all of it oxidation by alcohol dehydrogenase in the liver, and it runs at its own pace. NIAAA is direct about the alternatives: cold showers, hot coffee and walking do not reverse the effects of alcohol and can make things worse — they change alertness, not blood chemistry, which is a genuinely dangerous combination in someone deciding whether to drive.

Food belongs on the second list with an asterisk. Eating before or during drinking slows gastric emptying and flattens the peak, which is real and protective, and Jones's survey found elimination running faster in non-fasted subjects than in fasted ones. Neither effect removes alcohol you have already absorbed, so eating after the fact does nothing for a number that is already high. NIAAA adds a warning the arithmetic here cannot show at all: BAC can keep rising after someone stops drinking or loses consciousness, because alcohol still in the stomach and intestine continues to enter the bloodstream.

Limits: When This Calculator Does Not Apply

The Widmark equation is a good model of a simple thing. Below is an honest accounting of what it leaves out, ordered by how much damage each omission does — because a tool that only tells you its answer is less useful than one that tells you how wrong the answer can be.

It assumes every drink is already in your blood

This is the largest structural simplification and it has no input to control it. Set the hours to 0 with four drinks and the calculator reports 0.101% — as though all 56 grams crossed the gut wall the instant the fourth glass was empty. In reality absorption takes roughly 30 to 90 minutes and is slower on a full stomach, so an estimate taken early in a session runs high against a blood test at the same moment.

The mirror image is more dangerous. If you are still absorbing when you check the number, your real BAC is heading up while the calculator's is coming down at 0.015 an hour. Posey and Mozayani's review of BAC estimation models exists largely because of this gap: they summarize five published models and propose a sixth that incorporates absorption data from the gastrointestinal tract, and observe that these documented improvements to Widmark are frequently ignored in practice. This calculator is one of the tools ignoring them, which is worth knowing when you read its output.

Two fixed constants, two wide measured ranges

Recapping the numbers from the formula section, because they are the difference between a point and a band. Moving r across the range Seidl and colleagues measured takes a 180-lb man on four drinks from 0.049% to 0.084%, and a 140-lb woman on three drinks from 0.053% to 0.120%. Moving the elimination rate across the range Jones documented takes the same man on six drinks four hours in from 0.011% to 0.111%.

The tool cannot narrow those ranges, because it is not given the inputs that would — body-fat percentage, height, drinking history, liver health, whether you ate. Sex and weight are proxies for body composition and nothing more. Two 180-lb men, one lean and one not, have materially different body-water fractions and will get the same estimate from this page.

The input error that dwarfs the model error

All of the above is second to the drinks field. Miscounting is not a small perturbation; it is a multiplier on the largest term in the equation, and the calculator has no way to detect it.

In the first row the undercount is worth 0.088 percentage points — larger than the entire spread the distribution factor produces, and larger than the 0.08% legal limit itself. Anyone using this tool seriously should spend their effort on the drinks table above, not on worrying about the model.

What the inputs accept, and one quirk to know about

• Body weight is accepted between 35 kg and 300 kg, which is 77.2 lb to 661.4 lb. Anything outside returns "Enter a valid body weight" — 77 lb is rejected, 77.2 lb calculates.
• Standard drinks are accepted from just above 0 to 30, and decimals are allowed: 4.5 drinks for a 180-lb man at two hours returns 0.083%. Zero and anything above 30 return "Enter standard drinks (1–30)".
• The hours field is text and strips everything that is not a digit or a decimal point, so "3 hours" is read correctly as 3. Leave it blank, or type something with no digits at all, and it is treated as 0 hours — the most conservative reading, since no elimination is subtracted.
• That same stripping has one trap: a comma decimal is not understood. Typing "1,5" for an hour and a half becomes 15 hours, and a 180-lb man on four drinks drops from 0.071% to 0.000%. Use a period.
• There are no inputs for height, body fat, age, food, medication, drinking history or the times of individual drinks, so nothing on this page can account for any of them.

The calculator also fills in defaults where a field is left alone: hours starts at 2, and if the sex selector is somehow empty the formula falls back to the male distribution factor of 0.68, which produces the lower of the two possible estimates. Check the second output line to confirm which r was used.

It has no legal or evidentiary standing

Working a BAC backwards from a drinking account is called retrograde extrapolation, and it is contested even among forensic toxicologists. Posey and Mozayani note that many practitioners are reluctant to perform these calculations and that some jurisdictions expressly prohibit them. What courts use is an evidential breath test or a blood draw, analyzed under a chain of custody, not a web page.

That cuts in both directions, and the unhelpful direction is the one people forget. A calculator reading 0.06% is not a defense, because the measured value can be higher. It is also not a clearance: impairment charges below the per se limit exist in every state, the commercial limit is 0.04%, and for a driver under 21 the threshold is 0.02% — a figure a single drink clears on every body in every table on this page.

Where the model does not belong at all

• Anyone under 21 in the US, for whom the relevant threshold is 0.02% and the useful count is zero — the arithmetic here has nothing to add.
• Pregnancy, where no amount of alcohol is established as safe and a concentration estimate is beside the point.
• Anyone taking medication that interacts with alcohol. NIAAA singles out opioids and sedative hypnotics, including common sleep aids and benzodiazepines, and notes that even over-the-counter antihistamines with alcohol can be dangerous — combinations that can produce an overdose at moderate amounts of alcohol this calculator would score as unremarkable.
• Anyone with liver disease, and anyone whose drinking history places their elimination rate far from the population midpoint in either direction.
• Any situation where someone might be in danger. If a person is confused, hard to rouse, vomiting, breathing slowly or irregularly, cold or blue-tinged, that is an emergency call, not an input set.
• Anyone in recovery, or anyone for whom a page that makes drink counts look like a budget is unhelpful.

Nothing here is medical or legal advice, and the calculator knows nothing about you beyond four fields. Its honest use is to make the shape of alcohol accumulation visible — how fast it climbs, how slowly it falls, and how much of the answer depends on a body you cannot change on the night.