How does BarCount work?
BarCount estimates your blood alcohol content (BAC) directly on your smartwatch — using the Widmark model, the Watson formula, and a transparent elimination range grounded in forensic guidelines. This page explains what the app does, which formulas it uses, how accurate the estimate is, and why your data never leaves your watch.
1. What BarCount does — in one sentence
BarCount logs your drinks with a single tap on your Apple Watch or Wear OS smartwatch, calculates a real-time estimated blood alcohol content from them, and tells you when you'll likely be back at 0.00 — all offline, no account, no cloud.
2. The three phases of alcohol pharmacokinetics
Every BAC calculation — including BarCount's — models three sequential processes in the body:
- Absorption: Alcohol moves from the stomach and small intestine into the bloodstream. Depending on stomach contents, drink type, and drinking pace, this can take between 15 minutes and up to 2 hours after the last sip.
- Distribution: Alcohol spreads through the body water. How much body water you have depends on weight, height, age, and sex.
- Elimination: The liver breaks down alcohol at a roughly linear rate — typically 0.10 to 0.25 ‰ per hour, depending on the person.
BarCount represents all three phases as a timeline: every drink is an event with a timestamp, and the app computes the currently active amount of alcohol at every moment.
3. The formulas — explained transparently
3.1 Pure alcohol per drink
The first step is always to derive grams of pure ethanol from volume and alcohol content (ABV):
The factor 0.789 g/ml is the density of ethanol (NIST). European prevention literature commonly approximates this as 0.8 — the difference is in the low percentage range and is dwarfed by the biological uncertainty anyway.
3.2 The Widmark formula for the theoretical peak
The classic Widmark formula gives the maximum theoretical BAC without elimination:
The factor r is the reduction / distribution factor. Classic values are 0.68 for men and 0.55 for women — they represent the average share of body water.
3.3 The Watson formula for a personal distribution volume
BarCount additionally uses the Watson formula to estimate total body water (TBW) from age, height, weight, and sex. This produces a more individual distribution factor than the flat Widmark r and noticeably improves estimate quality — especially for people whose body composition deviates from the "average".
3.4 Elimination over time
Once alcohol is in the bloodstream, the liver breaks it down in an approximately linear fashion:
Forensic guidelines (NIST/OSAC) cite β = 0.010 – 0.025 g/dL per hour as the typical range — converted, roughly 0.10 – 0.25 ‰ per hour. BarCount uses the middle of this range as its point estimate.
4. Which parameters BarCount needs
For the calculation to work, BarCount collects your body profile once and then records each drink as an event. Here's the overview:
| Parameter | Used for | Required? |
|---|---|---|
| Weight (kg) | Core variable of distribution | Yes |
| Sex | Proxy for body water share | Yes |
| Age | Watson formula (body water) | Recommended |
| Height (cm) | Watson formula (body water) | Recommended |
| Drink: timestamp | Timeline, absorption/elimination | Yes |
| Drink: volume (ml) | Ethanol dose | Yes |
| Drink: ABV | Ethanol dose | Yes (preset default possible) |
| Food before/during drinking | Extends absorption phase | Optional |
5. Standard drinks: how many grams of alcohol are really in there?
For the calculation to be correct, BarCount needs to know how much pure alcohol actually ends up in your drink. The following table is based on the standard glasses defined by the German Federal Institute for Public Health (BIÖG, formerly BZgA) and adds typical bar drinks. All values are guidelines — the app lets you tweak every preset individually.
| Drink | Serving | ABV | Pure alcohol |
|---|---|---|---|
| Beer (small) | 0.3 l | ~5 % | ~12 g |
| Beer (large) | 0.5 l | ~5 % | ~20 g |
| Shandy (Radler) | 0.5 l | ~2.5 % | ~10 g |
| Wine / sparkling (small) | 0.125 l | ~12 % | ~12 g |
| Wine / sparkling (large) | 0.25 l | ~12 % | ~24 g |
| Port / sherry | 0.1 l | ~15 % | ~12 g |
| Shot (spirit) | 4 cl | ~38 % | ~12 g |
| Jägermeister (small shot) | 2 cl | 35 % | ~5.5 g |
| Jägermeister (shot) | 4 cl | 35 % | ~11 g |
| Gin & Tonic / Vodka Lemon | 4 cl spirit | ~40 % | ~12 g |
| Aperol Spritz (original) | 90/60/30 ml | Prosecco 11 %, Aperol 11 % | ~13 g |
| Mojito (IBA) | 45 ml rum | ~40 % | ~14 g |
| Alcopop | 0.33 l | ~6 % | ~16 g |
| Non-alcoholic beer | 0.5 l | up to 0.5 % | up to ~2 g |
Note: In Germany, drinks labeled "alcohol-free" are legally allowed to contain up to 0.5 % ABV. Drinking many of them still adds up to measurable grams of ethanol over the course of an evening.
6. Step-by-step example: a typical bar evening
An 80 kg man drinks over three hours:
- 8:00 pm — 0.5 l of beer (5 %) → ~19.7 g ethanol
- 9:00 pm — 0.2 l of wine (12 %) → ~18.9 g ethanol
- 10:00 pm — 4 cl of spirit (38 %) → ~12.0 g ethanol
Using r = 0.7 and β = 0.15 ‰/h, BarCount arrives at a point estimate of roughly 0.25 ‰ at 11:00 pm. By that point, the beer has been almost fully eliminated; wine and the spirit contribute most of the remaining value.
In food mode, BarCount extends the absorption window (because the stomach is full), which flattens the peak and makes the overall curve more realistic.
7. How accurate is the estimate really?
Honest answer: a BAC estimate is always an estimate, not a measurement. Reasons:
- Absorption varies. Depending on stomach contents, daily condition, and drink type, the peak can occur anywhere from 30 minutes to 2 hours after the last sip.
- Elimination rates vary from 0.10 to 0.25 ‰/h. Below ~0.2 ‰, elimination is no longer necessarily linear.
- The distribution volume is known to span about 0.40 – 0.80 l/kg — a single fixed value is forensically inappropriate.
- Medication, illness, tolerance, and even hormonal cycles have measurable effects.
That's why BarCount never pretends to show an "exact" value. The number on the dial is orientation — not permission for anything.
8. Binge drinking and special risk groups
The US NIAAA defines binge drinking as a drinking pattern that produces a BAC of ≥ 0.08 g/dL (~0.8 ‰) — typically 4+ drinks for women or 5+ drinks for men within about two hours. In this range, the uncertainties of the estimate grow noticeably: late drinks push the peak significantly later, and small errors in ABV or glass size have a much larger effect.
Pregnancy and breastfeeding
During pregnancy, alcohol is never safe: mother and child reach the same blood alcohol level within minutes. The recommendation from the BIÖG is unambiguous: complete abstinence. BarCount does not soften this message and should not be used to rationalize "occasional" drinks.
Medication and illness
Anyone on medication should check beforehand whether drinking alcohol is allowed at all. Liver conditions, metabolic disorders, and certain medications can significantly alter elimination rates — effects that BarCount cannot model.
9. Why your data never leaves your watch
BarCount is consistently local-first. Concretely:
- All personal data (age, weight, height, sex, drink logs) lives exclusively on your smartwatch.
- No account, no registration, no email.
- No cloud, no sync, no backup on remote servers.
- No analytics, no tracking, no crash reports containing user data.
- No internet connection required — the app works fully offline.
Health-related data like alcohol logs is particularly sensitive. The simplest and safest protection is not to transmit it in the first place. Details are in the privacy policy.
10. Legal context: drink-drive limits in Germany
BarCount is not legal advice — as general orientation, here are the key limits in Germany:
- 0.5 ‰ — administrative offense for drivers without impairment signs (§ 24a StVG).
- 0.3 ‰ — potential criminal offense if impairment signs are present.
- 0.0 ‰ — for novice drivers during probation, drivers under 21, and commercial drivers.
The European Commission recommends 0.5 g/L or lower for the general population and 0 or near 0 for novice drivers. BarCount displays these values purely as information — never as permission to drive.
11. What BarCount deliberately does not do
- BarCount does not tell you whether you may drive.
- BarCount does not measure — it only estimates based on a model.
- BarCount is not a medical device and does not replace professional advice.
- BarCount stores nothing in the cloud and shares no data with third parties.