Blood Alcohol Level Calculator

The one thing this must never decide

Do not use this, or any calculator like it, to work out whether you can drive. That is the single sentence this page exists to deliver, and everything below is really just the evidence for it.

The reasons are not squeamishness. They are arithmetic. The formula underneath carries a published uncertainty of about 21 percent, it cannot see most of the things that actually move your blood alcohol, and it assumes every drink hit your bloodstream the instant you ordered it. A number here that reads comfortably under a limit could genuinely be over it. Nobody has ever explained that to a police officer successfully, and more to the point, impairment starts well below any legal threshold anyway.

So what is it for? Understanding. It is a decent way to see roughly how alcohol behaves in a body, why two people drinking identically end up in different states, and how brutally slowly it leaves. That is worth knowing. It is just not a permission slip.

What the Widmark formula actually does

The sum here is nearly a century old and still the forensic standard, which tells you something about how little the underlying biology has changed. Erik Widmark, a Swedish scientist, worked it out in the 1920s by giving controlled doses of alcohol to hundreds of people and measuring what happened.

It runs in two steps. First, work out how concentrated the alcohol is once spread through the water in your body:

BAC = [ grams of alcohol ÷ (body weight in grams × r) ] × 100

Then subtract what your liver has removed since you started:

− (0.015 × hours since your first drink)

Two constants are doing the work. r is the fraction of your body that is water-containing tissue, taken as 0.68 for men and 0.55 for women. 0.015 is how many percentage points of BAC your liver clears per hour, which is an average rather than a personal figure.

Before any of that, the tool converts your drinks into grams of pure alcohol, because that is the only unit the formula understands. A drink's volume times its strength times 0.789, the density of ethanol, gives the grams. This is why the calculator asks for the size and the ABV of each drink rather than just how many you had. A pint of 3.5 percent bitter and a pint of 8 percent IPA are not the same event.

A worked example

An 80 kg man drinks two 12 oz beers at 5 percent ABV, and checks two hours after his first one.

Alcohol in each beer = 12 oz × 29.57 ml × 5% × 0.789 = 14 g
Two beers = 28 g
Peak BAC = 28 ÷ (80,000 × 0.68) × 100 = 0.0515%
Minus two hours of elimination = 0.0515 − 0.03 = 0.021%

That 14 g figure is worth remembering: a 12 oz beer at 5 percent is almost exactly one US standard drink. It is a useful yardstick for reading the rest of this.

Why the same two beers hit two people differently

Now run the identical drinks through a 60 kg woman, checking at the same two hours.

More than two and a half times higher, from the same two drinks. Some of that is weight, and some is the r factor, and the r factor is the part worth understanding.

Alcohol distributes into body water, not body tissue generally. Muscle is roughly three quarters water; fat holds essentially none. Men average a higher proportion of water-containing tissue, so the same alcohol is diluted into a bigger tank. Women average less, so the same alcohol arrives more concentrated. It is not about tolerance or how anyone feels; it is a dilution problem.

Which also means the 0.68 and 0.55 are population averages that will not fit you exactly. A very muscular man might genuinely sit at 0.70 or above, an obese man nearer 0.60, an athletic woman well above 0.55. The calculator has no way of knowing which you are, and that is one of several reasons the answer has slack in it.

The 21 percent problem

Here is the number that should govern how much faith you put in any of this. When researchers have worked through the uncertainty in a Widmark calculation properly, accounting for the variability in the r factor, in the drink volume, and in the stated strength, the coefficient of variation comes out around 21 percent.

Made concrete: an estimate of 0.050 percent is really somewhere in the region of 0.040 to 0.060 percent. And that is the well-behaved version, with honest inputs and no complications. It is not a confidence interval you can shave.

Twenty-one percent is fine for understanding how alcohol works. It is useless for deciding anything near a threshold, which is exactly the decision people want to make with it.

What the formula cannot see

The calculator knows four things about you: your weight, your sex, what you drank, and when you started. Here is a partial list of what it does not know and cannot account for:

• Food. Eating properly before drinking can cut your peak blood alcohol substantially and delay it by an hour or more, because a full stomach slows what reaches your small intestine. Two identical drinkers, one who had dinner and one who did not, are not in the same state.
• Medication. Plenty of drugs interact with alcohol or affect how it is processed, some seriously.
• Your liver. The 0.015 per hour is an average. Real elimination runs from roughly 0.010 to 0.020, which is a twofold spread across people.
• Body composition. Covered above: the r value is an average, not your value.
• Pace. Four drinks over four hours and four in one hour produce very different curves, and the formula treats them nearly identically.
• Tolerance. This one deserves care. Tolerance changes how drunk you feel. It does not change your blood alcohol at all, and it does not restore your reaction time. Feeling fine is not evidence.
• Whether your inputs were honest. Home pours are famously generous, and a "glass of wine" is not a unit of measurement.

You can still be going up after your last drink

A quiet flaw worth knowing about, because it produces exactly the wrong kind of error.

The formula assumes all your alcohol was absorbed instantly, the moment you started. Real absorption takes time, typically half an hour to an hour and a half to reach peak, longer with food. So your actual blood alcohol lags behind, climbs while you are sitting there, and can peak after you have stopped drinking entirely.

Which means the most dangerous moment is often not when you feel worst. You can finish your last drink, feel level, run a calculation that reassures you, and be measurably more impaired forty minutes later. The number was not lying about the arithmetic. The arithmetic was just describing a moment that had not arrived yet.

Only time lowers it, and nothing else does

This is the most useful thing on the page after the warning at the top.

Your liver clears alcohol at a roughly fixed rate, about 0.015 percentage points an hour, and that rate is not negotiable. It is one of the few processes in the body that just refuses to be hurried. So from a BAC of 0.08 percent, getting to zero takes over five hours. Not "a couple of coffees". Five hours and change, whatever you do in them.

Things that do not speed it up, despite everything you have ever been told:

• Coffee. Makes you an alert drunk person. Caffeine does nothing to your blood alcohol, and the alertness is arguably worse, because it removes the tiredness that might otherwise have stopped you doing something stupid.
• Cold showers. Make you a wet drunk person.
• Food afterwards. Eating before drinking helps by slowing absorption. Eating after the alcohol is already in your blood does nothing to remove it.
• Exercise, fresh air, sleep. Sleeping passes time, and time is the active ingredient, but you do not metabolise faster while asleep. Waking up early after a heavy night can leave you over a limit at breakfast, which is how a surprising number of morning-after drink-driving charges happen.
• Water. Worth drinking for the hangover. Does not dilute your blood alcohol meaningfully.

The beer or wine dropdown does nothing

Worth being straight with you about a quirk of this tool. Each beverage has a "Type" dropdown offering beer or wine, and the calculation ignores it entirely.

That is less of a problem than it sounds, because there is nothing for it to do. Once you have told the calculator the size and the ABV of a drink, it knows everything that matters. Alcohol is alcohol; 14 g of ethanol behaves identically whether it arrived as beer, wine, or anything else. The type would only matter if the tool were using it to guess a typical strength on your behalf, and it is better that it asks you instead, since real drinks vary far more than category averages suggest.

So set the type to whatever you like and put your care into the two boxes that do the work: the actual size in ounces, and the actual ABV printed on the label.

Legal limits are a ceiling, not a target

Limits vary by country and often within them. Much of Europe and Australia sit at 0.05 percent, the United States generally at 0.08, and plenty of places set lower or zero limits for new drivers, professional drivers, and anyone under a certain age. If you are trying to work out your local number, look it up rather than assuming.

But the framing itself is the trap. A legal limit is the point at which the state has decided to prosecute you. It is not the point at which you became worse at driving. Meaningful impairment of judgment, reaction time and coordination begins well below any of these thresholds, and in many places you can be charged with impaired driving under the limit if you are visibly affected.

The only figure that carries no risk at all is zero, and given a 21 percent uncertainty band, zero is also the only number this calculator can tell you with any confidence.

When it stops being about the number

If you are looking at someone who has drunk too much, put the calculator away. These are emergency signs, and they need an ambulance, not arithmetic:

• Unconscious, or conscious but impossible to rouse properly.
• Breathing that is slow, shallow, or irregular.
• Cold, clammy, pale or bluish skin.
• Vomiting while not fully awake, which carries a genuine risk of choking.
• Seizures, or deep confusion about where they are.

Do not let someone "sleep it off" in that state. Blood alcohol can keep climbing after the last drink, so a person who seems merely very drunk at midnight can be in serious trouble by one. Put them on their side, stay with them, and call for help. Nobody has ever regretted calling an ambulance for this.

Questions people ask

Can I use this to check if I am under the limit to drive?

No. The estimate carries roughly 21 percent uncertainty, cannot account for food, medication, your actual metabolism or your real body composition, and assumes instant absorption when you may still be rising. A reading of 0.05 could genuinely be 0.06. If you have been drinking, do not drive, and use the money you saved on the calculator for a taxi.

How long until my BAC reaches zero?

Divide your estimated BAC by 0.015. From 0.08 percent, that is more than five hours. This is why the morning after matters: stop drinking at 2am at a high level and you can still be over a limit at 8am, feeling merely rough rather than drunk. Nothing shortens it.

Why does the calculator give a woman a higher result for the same drinks?

Because alcohol spreads into body water, and women average a lower proportion of water-containing tissue, so the same alcohol arrives more concentrated. It is a dilution effect rather than anything about tolerance. In our example, identical drinks gave 0.021 percent to an 80 kg man and 0.055 percent to a 60 kg woman.

I have a high tolerance. Does that lower my BAC?

Not by a fraction. Tolerance changes how drunk you feel, not how much alcohol is in your blood or how impaired your reactions are. A breathalyser cannot tell how much practice you have had. If anything, tolerance is a hazard here, because it removes the feedback that would otherwise tell you to stop.

Does coffee or a cold shower sober you up?

No. Neither touches your blood alcohol. They can make you feel more alert, which is genuinely worse, because it disguises impairment that is still entirely present. Only time removes alcohol, at roughly 0.015 percentage points an hour, and nothing accelerates it.

References

Where the figures come from. The formula's structure, the conversion of a drink's volume and strength into grams of ethanol using a density of 0.789, and the coefficient of variation of roughly 21 percent on a calculated blood alcohol concentration are set out in the analysis of alcohol calculations and their uncertainty published in Medicine, Science and the Law. The original derivation of the formula, including the distribution factor and the linear elimination rate the tool applies, is Widmark's own work.

1. Alcohol calculations and their uncertainty. Medicine, Science and the Law. https://pmc.ncbi.nlm.nih.gov/articles/PMC4361698/
2. Widmark EMP. Die theoretischen Grundlagen und die praktische Verwendbarkeit der gerichtlich-medizinischen Alkoholbestimmung. Berlin: Urban & Schwarzenberg; 1932.