The Widmark formula is an equation forensic toxicologists use to estimate blood alcohol concentration from how much a person drank, their body weight, a biological distribution factor, and how much time has passed. Written out, it reads: BAC = (A / (W × r)) × 100 − (β × t). Swedish researcher Erik Widmark developed it in the 1930s, and it remains the standard tool for reconstructing BAC after the fact, especially when no chemical test was taken at the moment that matters legally.
The left side of the equation calculates a theoretical peak BAC, assuming all the alcohol was absorbed instantly and none has been metabolized. The right side subtracts what the liver has processed in the hours since. The output is a percentage, the same format you see on a breath or blood test.
What Each Variable Means
- A is the mass of alcohol consumed, in grams.
- W is total body weight, also in grams.
- r is the Widmark distribution ratio, which accounts for how alcohol spreads through the body based on sex and body composition.
- β (beta) is the hourly alcohol elimination rate.
- t is the number of hours between the end of drinking and the moment you want to estimate.
The formula requires alcohol in grams, not drinks or ounces. In the United States, one standard drink contains about 14 grams of pure alcohol, which corresponds to roughly 12 ounces of 5% beer, 5 ounces of 12% wine, or 1.5 ounces of 40% spirits.1Centers for Disease Control and Prevention. About Standard Drink Sizes For drinks outside those profiles, you multiply the volume in milliliters by the alcohol-by-volume percentage (as a decimal), then multiply by ethanol’s density of 0.789 g/mL. A 16-ounce craft beer at 7.5% ABV, for example, contains considerably more alcohol than two standard drinks.
The distribution ratio reflects how much of your body actually absorbs alcohol. Alcohol is water-soluble, so it concentrates in lean tissue rather than fat. A person with more body fat has less total body water, and the same dose produces a higher concentration. Commonly cited averages are 0.68 for men and 0.55 for women.2Indiana University Indianapolis. BAC Calculator More precise BMI-adjusted formulas exist: for men, r = 1.0181 − (0.01213 × BMI); for women, r = 0.9367 − (0.01240 × BMI).3PMC (PubMed Central). Alcohol Calculations and Their Uncertainty An overweight man might carry an r closer to 0.60, which meaningfully changes the result.
The elimination rate (β) is typically cited as somewhere between 0.015 and 0.017 g/100mL per hour, roughly one standard drink per hour. Individual rates vary by a factor of three to four depending on genetics, sex, age, food intake, liver health, medications, and drinking history. Heavy drinkers develop metabolic tolerance and clear alcohol faster. People with liver disease clear it slower. Even circadian rhythm plays a role.4National Center for Biotechnology Information. Alcohol Metabolism
A Worked Example
Take a 180-pound man who drinks four standard beers over an evening, and estimate his BAC two hours after his last sip.
Four standard drinks at 14 grams each equals 56 grams of alcohol. Body weight converts to 81,648 grams (180 × 453.6). Using 0.68 for r and 0.016 for β:2Indiana University Indianapolis. BAC Calculator
BAC = (56 / (81,648 × 0.68)) × 100 − (0.016 × 2)
The denominator comes to 55,521. Dividing 56 by that gives 0.001008. Multiplying by 100 produces 0.1008, the theoretical peak. Subtracting the elimination (0.032) leaves roughly 0.069%. Just under the common 0.08% threshold. Change a single input, though — a fifth beer, a lighter frame, a lower r — and the number crosses the line. The equation is that sensitive.
The Absorption Problem
Widmark’s equation assumes the alcohol is already in the bloodstream. In reality, absorption takes time, and that gap is one of the formula’s biggest blind spots.
On an empty stomach, BAC usually peaks 30 to 60 minutes after the last drink, with spirits absorbing faster than beer or wine.5PMC (PubMed Central). Absorption and Peak Blood Alcohol Concentration After Drinking Beer, Wine, or Spirits Food changes the picture entirely. Carbohydrates in the stomach can cut peak BAC to less than a quarter of what it would reach on an empty stomach, because food slows gastric emptying.6PMC (PubMed Central). Alcohol in the Body Someone who ate a full meal before drinking may not reach peak BAC for well over an hour.
If the formula is applied before absorption is complete, the result overstates how much alcohol is actually circulating. That matters whenever the time window between drinking and the moment of interest is short.
How Much Uncertainty the Formula Carries
Every Widmark estimate comes with error, and the size of that error depends on how well the inputs are known. There is no fixed margin that applies to all calculations. The coefficient of variation has to be worked out case by case through error propagation, accounting for the uncertainty in each variable.3PMC (PubMed Central). Alcohol Calculations and Their Uncertainty
The inputs themselves have documented variability. Published figures put the coefficient of variation at roughly 22% for the elimination rate, 9% for the distribution ratio, and 3% for a beverage’s stated alcohol strength.3PMC (PubMed Central). Alcohol Calculations and Their Uncertainty Earlier work suggested an overall coefficient of variation of about 21% for a forward Widmark calculation, meaning a reported BAC of 0.08% could realistically fall between roughly 0.063% and 0.097%. In a courtroom, that spread is often the whole case.
Working the Formula Backward: Retrograde Extrapolation
In most DUI cases, the chemical test is not taken at the moment of driving. A blood draw an hour or two after a stop measures BAC at the draw, not behind the wheel. Retrograde extrapolation runs the Widmark formula in reverse: start with the measured BAC and add back what the liver would have processed in the elapsed time (β × t).7PubMed. Retrograde Extrapolation of Blood Alcohol Data – An Applied Approach
Say a blood test two hours after a stop reads 0.06%. At a 0.015 per-hour elimination rate, the back-calculated BAC at the time of driving would be about 0.09%. That reconstructed value, not the actual tested one, is what the prosecution puts in front of the jury.
Courts generally accept retrograde extrapolation as scientifically reliable, but the prosecution still has to show that the calculation rested on adequate facts. An expert who assumes the driver was past peak absorption needs evidence to support that assumption, whether from the driver’s own statements, the officer’s observations, or a documented drinking timeline.
The Rising BAC Defense
Retrograde extrapolation only works if BAC was falling during the window between driving and testing. The rising BAC defense attacks that premise directly. If the last drink came close to driving, BAC may still have been climbing at the moment of the stop, and the test taken later could capture a peak that had not yet been reached while the person was behind the wheel.
Absorption can take 15 to 90 minutes depending on stomach contents, beverage, and individual physiology. Someone who finishes a drink, drives immediately, and gets pulled over ten minutes later may have had a BAC well below 0.08% at the wheel, even if a breath test 45 minutes after the stop reads over the limit. Standard retrograde math makes the gap worse by adding elimination time on top.
Making the argument work requires a tight timeline: the exact time of the last drink (receipts or card records help), the time and content of the last meal, the precise time of the stop, and the precise time of the test. Without that documentation, the defense stays speculative.
Getting the Numbers in Front of a Jury
Before any Widmark-based estimate reaches the jury, the judge decides whether to admit it. Federal Rule of Evidence 702 requires the party offering expert testimony to show that the opinion rests on sufficient facts, reliable methods, and a sound application of those methods to the case.8United States Courts. Federal Rules of Evidence
The Daubert framework guides that reliability analysis. Courts consider whether the method has been tested, whether it has been peer-reviewed and published, its known or potential error rate, whether standards govern its use, and whether it has broad acceptance in the field.9Legal Information Institute. Daubert v. Merrell Dow Pharmaceuticals, 509 U.S. 579 The formula itself clears those hurdles easily, with nearly a century of literature behind it. The contested ground is the application: whether the inputs the expert chose fit this particular person, whether the elimination rate was defensible, and whether absorption was accounted for.
Testimony that plugs generic averages into every slot and reports a single precise number invites challenge. Stronger testimony presents a BAC range and quantifies its uncertainty. The 2023 amendment to Rule 702 reinforced that the proponent carries the burden of showing reliability to the court.8United States Courts. Federal Rules of Evidence
Why Small Errors Produce Large Swings
The equation is deceptively simple, and the number it produces looks more precise than it is. Because the formula multiplies and divides estimates rather than adding them, small input errors compound. Miscounting a single drink in a four-drink scenario shifts the numerator by 25%. Using a population-average r instead of a BMI-adjusted value can move the denominator by 10% or more. Picking an elimination rate of 0.012 instead of 0.018, both within the normal human range, changes the subtracted term by 50%.
The weakest link is almost always the drinking history. Police reports commonly describe “a few beers” or “some drinks at dinner.” An expert building a Widmark calculation from that kind of record is guessing at the single most important variable. The output is only as good as the inputs, and real-world inputs are rarely as clean as a textbook example.