How the blood alcohol calculator works

Three kinds of numbers

• Measurable facts. The grams of alcohol in a drink follow from its volume and strength. Unit conversions are exact.
• Model assumptions. Distribution volume, elimination rate and absorption speed come from published population data.
• Individual uncertainty. Your own values are unknown. The range shows how much that matters; it cannot remove it.

Step 1: grams of alcohol in a drink

Alcohol (g) = volume (mL) × ABV (%) ÷ 100 × 0.789 g/mL. The factor 0.789 g/mL is the density of ethanol used in forensic practice; pure ethanol has a density of 789.24 kg/m³ at 20 °C, the temperature at which alcoholic strength by volume is defined in the EU.

Example: a 500 mL beer at 5 % ABV contains 500 × 0.05 × 0.789 = 19.7 g of alcohol. The alcohol in a cocktail is the sum of the alcohol in its alcoholic ingredients; ice, juice and mixers add volume but no alcohol.

Sources: AAFS Standards Board, Best Practice Recommendation for Performing Alcohol Calculations in Forensic Toxicology; International Organization of Legal Metrology, International Alcoholometric Tables (OIML R 22).

Step 2: from grams to concentration

Concentration (g/L) = absorbed alcohol (g) ÷ (V_d (L/kg) × body weight (kg)). V_d is the volume of distribution: how many litres of blood-equivalent each kilogram of body weight offers the alcohol. It is lower for bodies with a smaller share of water. Dividing g/L by 10 gives the familiar g/100 mL (% BAC).

This is the mass-per-volume form given in ANSI/ASB Best Practice Recommendation 122. No blood-density factor is applied, and the values are not mixed with Widmark’s original factors, which were defined against grams per kilogram of blood.

The ranges are the fixed ranges of ASB 122. The central values for male and female are population medians reported by Maskell and colleagues (2023); for “Not sure” it is the midpoint of the sex-independent range. The standard notes that these refer to sex assigned at birth and may be affected by gender-affirming hormone therapy. Body composition varies within each group, which simple categories cannot capture.

Sources: AAFS Standards Board, Best Practice Recommendation for Performing Alcohol Calculations in Forensic Toxicology; Maskell PD, Population ranges for the volume of distribution (Vd) of alcohol for use in forensic alcohol calculations.

Step 3: absorption takes time

Alcohol does not reach the blood the moment it is swallowed. In the model each drink enters a gut pool evenly while it is being drunk, and that pool empties into the body at a rate proportional to what is left in it (first-order absorption). The central absorption constant is 2.5 per hour, with a range of 1 to 6 per hour.

This is the weakest part of any such model. Published absorption constants run from about 1.0–1.7 per hour in venous blood to about 5–8 per hour after rapid drinking on an empty stomach, and food, drink type and the individual all change it. The range is meant to span “full stomach” to “empty stomach” without asking you; no food multiplier is applied.

As a plausibility check, simulating the protocol of Mitchell and colleagues (0.5 g/kg drunk over 20 minutes on an empty stomach) gives a peak at 45 minutes with fast absorption, 70 minutes with the central value and 103 minutes with slow absorption. The study observed mean peaks at 36, 54 and 62 minutes for vodka and tonic, wine and beer. A match in timing for one protocol does not show that the model is accurate for you.

Sources: Martin E, The pharmacokinetics of alcohol in human breath, venous and arterial blood after oral ingestion; Rangno RE, Ethanol 'dose-dependent' elimination: Michaelis-Menten v classical kinetic analysis; Uemura K, Individual differences in the kinetics of alcohol absorption and elimination: a human study; Mitchell MC, Absorption and peak blood alcohol concentration after drinking beer, wine, or spirits.

Step 4: elimination, once for the whole body

The body removes alcohol at a roughly constant rate while enough is present. The model subtracts 0.15 g/L per hour (0.015 g/dL per hour) in the central scenario and uses a range of 0.10 to 0.25 g/L per hour (0.010 to 0.025 g/dL per hour), the minimum range set by ASB 122.

• Elimination applies to all the alcohol in the body together. It is not subtracted separately from each drink, which would count it several times.
• The level cannot go below zero. A sober gap between two drinking sessions does not create credit that is later deducted from new drinks.
• Drinks can be entered in any order; the result depends only on their times.

Sources: AAFS Standards Board, Best Practice Recommendation for Performing Alcohol Calculations in Forensic Toxicology; Jones AW, Evidence-based survey of the elimination rates of ethanol from blood with applications in forensic casework.

Worked example

An 80 kg man (male option) drinks two 500 mL beers at 5 % ABV, the first from 20:00 to 20:30 and the second from 20:30 to 21:00.

1. Alcohol: 2 × 19.725 g = 39.45 g.
2. Theoretical maximum if everything were absorbed at once and nothing eliminated: 39.45 ÷ (0.71 × 80) = 0.69 g/L (0.07 %). With V_d from 0.58 to 0.83 L/kg it is 0.85 to 0.59 g/L.
3. The simulation, with gradual absorption and elimination running throughout, gives a lower curve:

Near zero: the central scenario reaches the model’s zero about 4 hours 40 minutes after the first sip, around 00:40. The fastest scenario gets there around 22:25 and the slowest around 04:35. The calculator rounds these outwards to quarter hours and shows “around 00:45, somewhere between 22:15 and 04:45”.

You can check the slow end by hand: 39.45 g ÷ (0.10 g/L per hour × 0.58 L/kg × 80 kg) = 8.5 hours.

What the sensitivity range is and is not

The range is the envelope of nine scenarios: the central one, plus every combination of low and high distribution volume, low and high elimination rate, and slow and fast absorption. Drinks entered with “I don’t remember exactly” add scenarios in which they sit earlier or later inside their time window, at five sampled positions each.

• It shows how far the result moves when the assumptions move within published ranges.
• It is not a 95 % confidence interval. The corner combinations are deliberately extreme and no probability is attached to them.
• It is not a guaranteed upper or lower limit. Some people fall outside the published ranges, wrong inputs (a stronger drink, a larger pour) move everything, and values between the sampled corners can fall slightly outside the band.

Searle’s review of alcohol calculations concludes that their uncertainty is not a fixed percentage and grows in relative terms as the level falls, and that an error range must always be quoted.

Sources: John Searle, Alcohol calculations and their uncertainty; AAFS Standards Board, Best Practice Recommendation for Performing Alcohol Calculations in Forensic Toxicology.

Near zero, and why 0.00 is never shown

Straight-line elimination stops being a good description at low levels: ASB 122 states that below 0.020 g/dL the elimination rate may not be linear. Below that level the calculator shows “below 0.02 %” instead of a figure, and it never shows a rounded 0.00.

“Near zero” is the point where a scenario’s line reaches zero and less than 0.1 g of alcohol is still waiting to be absorbed, with no later rise. Because real elimination slows at the very end, the last traces can take longer than a straight line suggests. If the slowest scenario has not reached zero within 5 days, the calculator says so instead of giving a time.

What the model leaves out

No multipliers are applied for food, coffee, water, sleep or “holding your drink”: there is no sound basis for turning them into numbers for one person. These things do affect real blood alcohol and are outside the model:

• Food in the stomach, which slows absorption and lowers the peak.
• Medicines that change absorption or metabolism.
• Genetic differences in the enzymes that break down alcohol.
• Liver disease, and the faster elimination seen in heavy drinkers.
• Height, age and body composition beyond the three body-water options.
• Drinks that were stronger or larger than what was entered.

Sources: National Institute on Alcohol Abuse and Alcoholism, The Basics: Defining How Much Alcohol is Too Much; Cederbaum AI, Alcohol Metabolism.

Units

Internally the calculator works in grams per litre of blood. % BAC is g/L ÷ 10, mg/100 mL is g/L × 100, and per-mille by mass (g/kg) is g/L ÷ 1.055, the density of whole blood in g/mL. The standard drink definition you choose changes how an amount of alcohol is described, never the grams used in the calculation; see the standard drink calculator.

Sources: Ialongo C, The enzymatic analysis of alcohol (ethanol) in serum and plasma with the alcohol dehydrogenase reagent: focus on intra-analytical and post-analytical aspects.

Model version

Model version 1.0.0. The parameters were checked against the sources below on 6 November 2026.