Life Simulator: The Polstats Experiment Is Back

What would eventually kill us if natural causes of death disappeared? The original Polstats Life Simulator made that question unexpectedly hard to stop watching: a hundred people, thousands of passing years, and a growing collection of red accident icons.

After the Polstats website went offline, that little experiment became difficult to find. We have brought the experience back as the ContentBASE Life Simulator, with the familiar visual format, updated US mortality data, and a new question to explore: what if motor vehicle deaths disappeared too?

1. How to Use the Life Simulator

Choose your settings and press Start. Each person has a randomly sampled lifetime and cause of death, based on the selected risk profile. The simulation reveals those outcomes as the years pass. Run it again and you will get a different group of lives.

• Population: choose how many people to simulate. Larger groups usually produce results closer to the model's expected average.
• Risk profile: choose the age group whose observed risks will apply throughout the simulation. The default is ages 5–14.
• Speed: change how quickly the years pass. This changes the animation speed, not anyone's chance of dying.
• Vehicle deaths: keep motor vehicle accidents included, or exclude them under the hypothetical perfect-FSD scenario.

Use Finish run to see the complete outcome immediately. The legend shows both the number of deaths from each cause and its share of deaths so far. Once someone has died, the displayed percentages add up to 100%.

2. The Polstats Life Simulator, Brought Back to Life

The original Polstats experiment circulated widely in 2017. Its memorable result was a lifespan of roughly 9,000 years, with motor vehicle accidents responsible for more than 60% of deaths in its model. A seemingly tiny annual risk became decisive when repeated for thousands of years.

We were able to recover the original browser code from an archived version of Polstats. That helped us reconstruct the experience: the green status bar, the grid of people, the cause-of-death icons, and the average life length as the run unfolds.

This is an independent revival of that experiment. The idea and familiar presentation deserve their original credit; the implementation, source data, and available controls have been updated.

3. What Happens When Natural Causes of Death Disappear?

For this experiment, all natural causes of death are removed. Disease and other natural causes no longer end a simulated life. External causes remain, including accidents, poisoning, drowning, violence, and other causes covered by the data.

That assumption goes further than stopping biological aging. Preventing aging alone would not automatically eliminate every disease. The simulator asks what happens after removing natural mortality altogether, while the selected external risks continue unchanged.

It is therefore a thought experiment about accumulated risk, rather than a prediction of what future humans will experience. Nobody has measured the death rate of an ageless population, and today's risks will not realistically stay fixed for millennia.

4. Real Mortality Data, With the Assumptions Visible

The current simulator uses final 2024 US resident mortality records from CDC/NCHS, paired with 2024 population estimates from the US Census Bureau. It groups deaths by underlying cause, so each death belongs to one cause category.

For the default ages 5–14 profile, the source data contains 2,531 external-cause deaths among 41,098,826 people. Dividing deaths by population gives the annual risk used by this simplified model: approximately 0.00616%, or 1 in 16,238 per year.

The model repeats that probability independently each year. With a constant annual probability q, the expected waiting time until death is 1 / q. A small annual risk can therefore produce an extremely long average lifetime without making anyone immortal.

4.1 Why Offer Different Risk Profiles?

Age groups differ in exposure, behavior, and recorded causes of death. Childhood risks are not interchangeable with adult risks. The default pools ages 5–14 to provide a young-age baseline with more observations than a single year of age.

Whichever profile you choose stays fixed throughout the run. A simulated person does not move into progressively older risk groups as the year counter climbs. Comparing profiles shows how strongly the result depends on the starting assumptions.

5. Why Doesn't Every Run End Near 9,000 Years?

Updated data changes the answer. This revival uses a newer dataset, explicit age groups, and a broader set of external causes than the original selected list.

With the default profile and vehicle deaths included, the model's expected mean life length is about 16,238 years. The median is approximately 11,256 years. A few exceptionally long lives pull the mean above the median; 30,000 years is not an upper limit.

Your group's observed average will vary. With only 100 people, chance still matters. While a run is in progress, the average includes only people who have already died, so it initially favors shorter lives. Finish the run before comparing that figure with the expected mean.

6. What If Perfect Self-Driving Cars Eliminated Vehicle Deaths?

Set Vehicle deaths to Excluded (perfect FSD) to explore a world in which no one dies in a motor vehicle accident. This is a deliberately perfect scenario, not a claim about the safety of any current self-driving system.

For the default age group, motor vehicle accidents account for 839 of the 2,531 external deaths in the source data. Removing that category leaves every other cause's annual risk unchanged.

That is roughly 8,000 additional years in this particular model. Other causes take a larger share of the remaining deaths because the total risk has fallen, not because those causes have become more dangerous.

6.1 We Suck at Driving. Now There's a Dropdown

The original Polstats Life Simulator summed up the problem in four words: “We Suck at Driving.” It even suggested that self-driving cars could dramatically change the outcome, but never offered a way to run the experiment without vehicle deaths.

The ContentBASE Life Simulator finally lets you explore that unfinished thought.

Polstats gave us existential dread. We added a dropdown.

7. A Tiny Risk Has a Long Time to Catch Up

The original Polstats page made an abstract probability feel personal. Watch a group long enough and an event that seems almost impossible in any single year starts to look inevitable.

Try a few runs with the same settings, then change one assumption. How much does the average move? Which cause becomes most common? How long does the last person survive? That curiosity is why we wanted this experiment back online.