Model changelog

This is the dated record of every review and release of the engine behind your number. Releases change the model. Reviews check it against new research and sometimes change nothing, and a review that changed nothing is still listed, because "we looked and it held" is part of the record.

The engine's design, every source it draws on and the reasoning behind each release live on the methodology page. This page is the timeline.

Live model: engine 2.3.0 · calibration 2026.11.1 · last reviewed 1 November 2026. Every saved estimate records the engine and calibration version that produced it. When the model changes, your history is re-scored from your original answers under the new version, so an old measurement and a new one are always compared on the same math.

How to read this page

2026

17 November 2026 · The smoking question asks about cigarettes. No model change. The question read "How often do you smoke or vape?" while every study behind our smoking figures measured cigarettes, so a person who only vapes was read as a smoker with no source behind the figure. It now reads "How often do you smoke cigarettes?" and says that vaping is not scored. We found no study that measures how vaping alone changes life expectancy. Unmeasured is not the same as safe. If you smoke and vape, answer for your cigarettes. If you answered the old question for vaping alone, re-measure and your estimate will follow the new wording. In the same change, the smoking lines in the report were reworded to what each study says, and the lines we could not source were cut.

14 November 2026 · Engine 2.3.0 and calibration 2026.11.1 released. Smoking at its published size (erratum). Until this release smoking shared one combined ceiling with four other habits, so a daily smoker of 40 read about two and a half years below someone who never smoked. The studies say more than three times that. The error was ours. We had read one study's combined figure for five habits (Li et al., 2018) as the best answers against the worst and shared it out evenly, when it compares all five low-risk habits with none, and anyone who ever smoked counts as not low-risk. Smoking now uses published figures by age, from a 2024 study of American smokers (Le, Mendez and Warner, American Journal of Preventive Medicine), the same for women and men. A daily smoker of 40 now reads 8.7 years below a never-smoker, and 5.1 at 70. The years shown for quitting now fall with age, as every study finds, 6.8 at 40 and 1.2 at 70, where before they stayed near 2.3 to the age of 50. A current smoker's potential now counts quitting today at their age, so it is lower than it was, by about two years for the median daily smoker. Someone who quit within the last five years no longer sees a smoking gain to collect, because their estimate already counts staying quit. In a run of 20,000 simulated profiles a never-smoker moves up by about 0.4 of a year and never more than 0.6, a daily smoker down by 4.2 years at the median and at most 6.3, a some-days smoker by 2.6 and a recent quitter by 1.5. Biological age moves with them, up by about five and a half years for the median daily smoker. The same release corrects a second error of ours. Each habit's years were rounded partway through the sum, so the same answer on one habit could show a different figure depending on your other answers, and a small gain could print as zero. The sums are now carried exactly and rounded once, when a figure is shown. On its own that moves about a fifth of estimates by a tenth of a year and none by more, and about a third of the years shown for a single habit by a tenth. The limits, which the methodology lists in full, are that the study is the lowest of four and we have read its summary only, that we do not ask how many a day or the age at quitting, and that vaping is counted with cigarettes although no study we read measures it.

1 November 2026 · Engine 2.2.4 released. Three errata in how the estimate is read. No factor's adjustment changed. First, the survivorship path is now the only path: your estimate is the CDC's figure for someone who has already reached your age, plus your habit adjustments at 85% of their size (75% from 75). The engine used to compute the at-birth figure plus your adjustments at full size as well, and show the larger. With a conditional figure at every age from 18 since October, that second path could only win for the healthiest young profiles, and for them it produced an estimate that fell with each birthday (a man of 18 with the best answer on everything and a BMI of 22 read 84.6, 84.5 at 20, 84.4 at 22 and 84.3 at 25; he now reads 84.1, 84.1, 84.2 and 84.3). Only people under about 27 with well-above-average habits move, down by up to 0.7 of a year on the estimate and on the potential. Second, two intermediate values were rounded to a tenth before the final rounding, which made the average for an age and sex zigzag between neighbouring ages (men of 18 to 22 read 80.0, 79.9, 80.0, 79.9, 80.0, and now read 80.0 at each); both are carried unrounded, and about a third of estimates move by a tenth, in either direction, none by more. Third, biological age interpolates between the two points of the life-table curve either side of your value instead of taking the nearer one; the old grid drifted by a hair and broke every tie downward, so about a third of biological ages read a tenth low. About one in five moves by a tenth, in both directions. The at-birth figure the engine keeps for an age outside its table is the CDC's 76.5 for men and 81.4 for women at one decimal (77 and 81 before); it reaches no estimate between 18 and 100. Design notes in the version history.

1 November 2026 · Calibration 2026.11.0 released. The two combination penalties removed. Until this release two combinations carried an extra penalty on top of the factors' own adjustments. Daily smoking with a BMI of 30 or more took one more year, and two from a BMI of 35. Smoking with more than about five drinks a week took up to a year and a half more. Both were our own modelling choice, not figures from a study, and the November review went looking for an all-cause mortality study that supports either. It found none. The studies that measure smoking together with obesity find the two risks combine about additively or less, never more (Mehta and Preston, 2016; Banack and Kaufman; Luijckx et al., 2019); the one cohort that tested smoking together with heavy drinking found no interaction (Hart et al., 2010); and the only source behind the alcohol penalty was a review of four cancers of the mouth and throat. Each habit's own adjustment already carries its full effect, so an extra penalty on top needs evidence of its own. Nobody's estimate falls. Only people who smoke daily or on some days and either carry a BMI of 30 or more or drink more than about five drinks a week move, all upward, by a third of a year to three years depending on age, sex and how far past each threshold they sit; their biological age falls by about half a year to five and a half years. The potential ceiling does not move, so for these readers the years still to gain from smoking are smaller by the amount their estimate rose. Everyone else is unchanged. Design notes in the version history.

1 November 2026 · Monthly science review (November). No engine change, and the one calibration change the review led to is the release listed above. The three sources the baseline itself is read from were checked on their own pages and are still the current editions (the CDC's 2024 life tables, the 2024 mortality brief, the 2026 Trustees Report's mortality assumption), and every habit source was checked for newer work, with nothing that moves a number. Five corrections, each checked against its paper. The body-weight source now reads as the paper reports it (3.6 million adults, 1.97 million of them never-smokers); the sense-of-purpose figures name what each model adjusts for (about 15% lower mortality after health behaviors and conditions, about 9% in the cohorts that also accounted for depressive symptoms); the short-sleep line in the biological-age article no longer attributes a faster DunedinPACE pace of aging to short sleep, a finding the study made for insomnia; the stair-climbing action describes the trial's three flights of 60 steps instead of "about a minute"; and the smoke-free rationale no longer calls a former smoker's cancer risk "optimal", since a heavy smoking history keeps lung cancer risk above a never-smoker's for decades. The family-history question's subtitle drops a percentage that traced to no paper, and the methodology page now states the three heritability figures on their three definitions. In the habit library, every proposal the review made was checked against its paper the same night: seven actions were re-sourced or reworded (alcohol near bedtime to Gardiner 2025; Lp(a) and the coronary calcium scan to the 2026 ACC/AHA guideline; handgrip holds to Lin 2026; the eating-earlier action now says a two-week trial found no insulin-sensitivity gain from timing alone; the cancer-screening action drops its figures; the exercise rationales no longer promise a "maximum benefit"), nine ungraded actions were softened where they asserted an effect with no source, and one action was added, the shingles vaccine from 50 (the CDC's recommendation, with the early dementia finding from Eyting 2025 named as a signal not yet proven, graded emerging). The library stands at 100 of 168 actions carrying a graded primary source.

6 October 2026 · Calibration 2026.10.1 released. Every age in the table. The life table inside the engine now holds the CDC's 2024 figure for every single age from 18 to 100, read from the CDC's spreadsheets by a script and kept at the files' own precision. It used to hold every fifth age and draw a straight line between them, which ran up to a tenth of a year high from 75 up. Estimates move by at most 0.2 of a year, and biological age by at most one year at the oldest ages. The average used for the other person on a compare card moves by one year at a few ages.

6 October 2026 · Calibration 2026.10.0 released. The 2024 life tables. The baseline moved from the CDC's United States life tables for 2023 to the 2024 edition (National Vital Statistics Reports volume 75, number 5). Remaining life expectancy is higher at every age in the new tables, by about 0.7 of a year for men and 0.4 for women at 18, narrowing to a tenth of a year or less by 80, and estimates rise by about the same amounts. No factor's adjustment changed. The 2024 edition was published on 25 August 2026 and we moved to it six weeks later, after one monthly review had passed without noticing it. That review covered the studies for each of the nine habit factors and did not check the life table's own edition. It does now, as its first step.

6 October 2026 · Engine 2.2.3 released. An erratum. The CDC's remaining-life-expectancy figure for your age was being rounded to a whole year before your habits were applied, and adults under 30 were measured against the at-birth figure alone because our copy of the table began at 30. The average for an age and sex therefore dipped slightly with each birthday and jumped a full year at about twenty of them, and a person with below-average habits gained about three years on turning 30, more at a very high weight. The figure is now carried unrounded into the adjustments, and the table starts at 18 (ages 18, 20 and 25 added from the same CDC spreadsheets). No factor's adjustment changed. From 30 up, estimated lifespans move by about a third of a year at the median, in either direction, and never more than one year. Under 30 they rise, typically by half a year to a year, and by about three years for the least healthy profiles, more at a very high weight. Under 30 the years each habit has left to give are smaller by about 15%, because the same discount on habit effects that applied from 30 now applies from 18; the order of the levers does not change. Biological age is read off the same curve and moves by about 0.2 of a year at the median, more at the oldest ages; an exactly average profile now reads exactly its own age. The average used for the other person on a compare card moves by one year at a few ages. Design notes in the version history.

3 October 2026 · Monthly science review (October). No engine or calibration change. Every number touched was checked against its primary paper. Four corrections across the articles and the report. The sense-of-purpose figure now reads as the paper reports it, about 24% lower mortality (HR 0.76), on every surface including the paid report; the step-count bands by age in the exercise article were the wrong way round and now follow Paluch 2022 (8,000 to 10,000 steps under 60, 6,000 to 8,000 at 60 and older); the Mediterranean-diet figure is 8% (Sofi 2010) everywhere; the alcohol article's short answer now matches its body.

In the habit library, "Cut caffeine after noon" was promoted to established evidence on Gardiner 2023, a new action was added (swapping table salt for a potassium-enriched salt substitute, from the SSaSS trial, Neal 2021, with a kidney and medication safety note), and the secondhand-smoke action gained its source (He 1999) and a safety note. The library stands at 99 of 167 actions carrying a graded primary source.

2 October 2026 · Engine 2.2.2 released. The potential ceiling and each habit's "years available" now use the best answer a person can still reach. A current or former smoker can never become a never-smoker, so the ceiling for a smoker is now "quit long ago", not "never smoked". Body weight gained its own years-available figure through the same transform as every other habit, so it ranks in the one order with them. No factor's adjustment, endpoint or current estimate changed. Design notes in the version history.

9 September 2026 · Calibration 2026.09.0 released. The body-weight adjustment above BMI 40 keeps rising instead of flattening, and below 18.5 it rises as BMI falls. One value had been standing in for every weight in each tail, so BMI 40 and BMI 54 had scored the same. Nothing between BMI 18.5 and 40 changed, and no estimate became more favourable. The alcohol slider's starting position moved from two drinks a week to zero, because about a third of adults do not drink.

5 September 2026 · Engine 2.2.1 released. A consistency fix. The potential ceiling and every per-habit figure now pass through the same transform as the headline estimate (the same survivorship discount, cohort adjustment and floor). Before this, one habit's "years available" could exceed the whole gap between your estimate and your potential. No adjustment, endpoint or estimate changed.

5 September 2026 · Monthly science review (September). The first review run on the monthly routine, with 26 findings and 15 recommended edits, all applied. The diet figure in the paid report was corrected from 9% to 8% (the 9% belongs to Sofi 2008, the model cites Sofi 2010). Three Blue Zone rationale sentences in the library were replaced with cohort evidence (Wang 2023, Sutin 2026). The ultra-processed-food share of U.S. adult calories was corrected from 57% to 53% (NCHS 2021 to 2023). The 75-minutes-a-week exercise gain was corrected to 1.8 years (Moore 2012). The smoking article's quit-attempt claim was corrected ("quadruples attempts" rather than "triples success"). Citations were advanced to their newest editions (Lindson 2026, Livingston 2024, Kaminski 2026 for NordICC's 13-year follow-up). No model change.

16 August 2026 · Erratum, the life table. An accuracy audit found that the conditional life-expectancy table inside the engine had drifted from the CDC source it cites (United States Life Tables 2023, NVSR volume 74 number 6). Men aged 40 to 60 were overstated by up to 1.0 year and women at 30 and 35 were understated by about 0.5 years, the result of a partial update from an older vintage. The table was corrected cell by cell to the CDC spreadsheet, and a build check now pins it to that source so the drift cannot recur. The same audit corrected fourteen articles, among them a misattributed colonoscopy citation, the alcohol years-lost table (rebuilt on Wood 2018) and a smoking sample size (202,248, not 216,917).

20 July 2026 · Erratum, "Rather not say". The "Rather not say" answer for sex was given the average of the two baselines, as intended, but the women's range for habit effects, which made every habit adjustment about 7% too large. The habit range for that answer is now the average of the two as well. Found in a review of the assessment.

1 July 2026 · Calibration 2026.07.0 released. The quarterly citation refresh, on engine 2.2.0, which dates from 8 June and brought the sliders for alcohol, activity and sleep. Two evidence-driven corrections. Light drinking (one to three drinks a week) moved from a small benefit to neutral, on Zhao and Naimi 2023 (4.8 million people, bias-corrected) and two Mendelian-randomization cohorts, which show the protective J-curve was a sick-quitter artifact. The sense-of-purpose adjustment was shrunk because reverse-causation analyses show the raw association is inflated. Sources for smoking, activity, social connection, stress, preventive care, sleep and diet were advanced to larger and newer cohorts, and the research base behind the model passed 40 million participants.

Before July 2026. The model's foundations, from the centering rebuild on, are described in the version history on the methodology page.

What a review checks

Each month the research behind all ten factors is read again, along with the source behind every health claim in the habit library. The check is the same every time. Does the cited paper still say what we say it says, with the effect size we quote. Has a larger or newer study superseded it. Has anything been retracted or corrected. Does every surface that quotes the number (the results screen, the report, the library, the articles, this site) quote the same one. Findings become edits, edits become an entry here, and a change to the model becomes a release with a version number you can see on your own saved result.

See your number

Your estimated lifespan and how it compares against the average, from the engine above. Free, in two minutes.

Take the Assessment →