How these estimates work
Every number on this site comes from one century-old formula, a handful of published population ranges, and a set of deliberately conservative choices. This page states all of them, so you can judge the estimates instead of trusting them.
The Widmark formula
In 1932 the Swedish physician Erik Widmark published the model that still underlies forensic alcohol calculation.1 In the form this site uses:
BAC(t) = A / (W × r × 10) − β × t
Ais the mass of pure alcohol you drank, in grams. The engine works in grams throughout; standard drinks are just packaging (14 g in a US drink, 8 g in a UK unit, 10 g in most EU guidelines).2Wis body weight in kilograms.ris the Widmark distribution ratio: the effective fraction of your body that alcohol distributes into. It tracks body water, which is why it differs by sex and build.βis the elimination rate: how fast BAC falls per hour once your liver is working through the queue.tis hours since your last drink.
The ranges, and where they come from
A calculator that picks single values for r and β is choosing your physiology for you. This site runs the formula across published population ranges instead, and reports the whole interval:
| Parameter | Range used | Basis |
|---|---|---|
| r, men | 0.6 to 0.76 | Widmark's men averaged 0.68; modern surveys spread roughly ±0.08 around it.3, 4 |
| r, women | 0.49 to 0.61 | Women average 0.55 with a similar spread: less body water per kilogram, so the same drinks land higher.3, 4 |
| r, unspecified | 0.49 to 0.76 | Skip the question and the engine spans the full adult range: honesty by widening, not by guessing. |
| β (elimination) | 0.012 to 0.02 %/hour | Jones's evidence survey: social drinkers cluster in 0.012 to 0.020, averaging near 0.015.3 |
The sober-by window you see is the formula evaluated at the corners of these ranges. The early edge assumes a high-water, fast-clearing body; the late edge a low-water, slow-clearing one. The headline always quotes the late edge, because the cost of being wrong is not symmetric.
Deliberately conservative choices
- The clock starts at your last drink. By default the model treats every gram as absorbed, and none eliminated, until you stopped: the most cautious reading, and an overstatement for long sessions. If you also give it your first-drink time, it credits the alcohol cleared while you were drinking, with each edge of the range using its own elimination rate so the slow-clearing corner stays a consistent slow clearer. Skip the field and you keep the overstatement.
- Recent drinks flag as "still rising." BAC peaks 30 to 90 minutes after the last drink, later on a full stomach.6 Within that window the calculator refuses to pretend the peak has passed.
- The headline is the late edge. "Probably sober by 2pm" on this site means even the slow corner of the model has cleared, not the average.
What the model cannot know
Genetics of alcohol dehydrogenase, liver health, medications, body composition beyond the sex averages, how much was really in each glass, carbonation, food timing, and fatigue all move real BAC in ways no formula reads.4, 5 Impairment adds a second layer: judgment and reaction time can stay degraded even as BAC approaches zero, and sleep deprivation stacks on top.
This is why the honest chain of authority runs: a blood test measures, a decent fuel-cell breathalyzer estimates closely, a formula estimates roughly, and how you feel barely counts as evidence. This site lives on the third rung and says so on every result.
Never drive after drinking. If you are wondering whether you are sober enough to drive, the answer is no.
Frequently asked questions
How accurate is a BAC calculator?
Under lab conditions, Widmark-style estimates commonly land within about 0.01 to 0.02 percentage points of a measured BAC, and outliers are worse. In real life, unknown pour sizes and drink strengths add error on top. That is a substantial margin at exactly the values people care about, which is why this site only publishes ranges and puts the conservative edge in the headline.
Why does the calculator show a range instead of a number?
Because the two dominant unknowns, the body-water ratio and the elimination rate, each vary by tens of percent across healthy adults, and no formula can know where in the range you sit today. A single number would be an invented precision. The range is what the evidence actually supports.
Why does the estimate start the clock at my last drink?
By default the model assumes all your alcohol is absorbed and none eliminated until the moment you stopped drinking, then eliminates from there. Real bodies eliminate during the night too, so for long sessions this overstates BAC and pushes the sober-by window later. That is a deliberate conservative choice: when a simplification must err, this site errs toward "assume more impaired." The optional first-drink field softens exactly this: it credits elimination during the session itself, at each range edge's own published rate, which tightens long-night estimates while staying honest.
Why will this site never show my exact BAC or a green light?
Because it cannot know either. An estimate that displays 0.049 next to a green checkmark is telling a story with two lies in it: the decimal implies precision that does not exist, and the color implies permission that no calculation can grant. Ranges, muted colors, and a permanent never-drive line are the honest presentation of what this math can and cannot do.
Try it against your own skepticism: the sober-up estimator and the standard drink calculator both run exactly the math described here, nothing more.
Sources
- Widmark, E.M.P. (1932). Die theoretischen Grundlagen und die praktische Verwendbarkeit der gerichtlich-medizinischen Alkoholbestimmung. (The original distribution-ratio model; reviewed in Jones 2010 below.)
- NIAAA. What is a standard drink? (14 g of ethanol per US standard drink.)
- Jones, A.W. (2010). Evidence-based survey of the elimination rates of ethanol from blood; with applications in forensic casework. Forensic Science International 200:1-20.
- Norberg, Å., Jones, A.W., Hahn, R.G. & Gabrielsson, J.L. (2003). Role of variability in explaining ethanol pharmacokinetics: research and forensic applications. Clinical Pharmacokinetics 42:1-31.
- Cederbaum, A.I. (2012). Alcohol metabolism. Clinics in Liver Disease 16:667-685.
- Mitchell, M.C., Teigen, E.L. & Ramchandani, V.A. (2014). Absorption and peak blood alcohol concentration after drinking beer, wine, or spirits. Alcoholism: Clinical and Experimental Research 38:1200-1204.
- NIAAA. Alcohol metabolism.