- Brake a car and watch how a phantom traffic jam grows or fades on a one-kilometre ring road.
- Change density and driver settings, then compare matched runs with hypothetical coordinated vehicles.
- Inspect the time–space diagram, replay the same inputs, and share a scenario link or result card.
- Make Your First Phantom Traffic Jam
- The Cars Go Forward. The Jam Can Go Backward
- Change One Road Setting at a Time
- A Following Gap Is Not a Reaction-Time Slider
- Can Coordinated Cars Calm the Wave?
- What the Traffic Statistics Mean
- Share a Repeatable Experiment
- Traffic Jam Simulator Questions
- Keep Experimenting
You brake for a moment. The car behind you brakes a little harder. A few drivers later, someone has to stop. Eventually the queue disappears, and there is nothing at the front of it: no accident, no roadworks, no obvious explanation.
That is the kind of puzzle ContentBASE Traffic Jam Lab lets you explore. This interactive traffic jam simulator puts cars on a closed ring road, gives them simple car-following rules, and lets you disturb the flow. One small brake input can turn into a surprisingly persistent problem.
Traffic Jam Lab
Brake one car on a ring road and explore how a phantom traffic jam forms. Adjust density, following gaps and hypothetical coordinated vehicles.
Loading the traffic experiment… Enable JavaScript to explore it.
An independent numerical experiment based on the documented Intelligent-Driver Model.
Start with free Canva bundles
Browse the freebies page to claim ready-to-use Canva bundles, then get 25% off your first premium bundle after you sign up.
Free to claim. Canva-ready. Instant access.
1. Make Your First Phantom Traffic Jam
Start with Classic Phantom Jam. It uses 36 cars on a one-kilometre ring, with a brake input scheduled for car 1 after ten simulated seconds. Watch the green and amber cars move clockwise, then follow the slower colours as the disturbance reaches the drivers behind.
You can also tap a car directly to brake it, or use the car selector and Brake selected car button. Each input asks that driver to decelerate for 2.5 seconds. Nobody has to crash or leave the road for a queue to emerge.
Try Room to Recover next. The road is the same length, but there are fewer cars. The difference is the experiment: a disturbance that grows at one density can fade at another.

2. The Cars Go Forward. The Jam Can Go Backward
A traffic wave is not a particular group of cars parked in one place. Vehicles enter the slow patch, lose speed, and eventually leave it. The disturbance can move upstream through the stream even while every driver continues travelling forward.
The time–space diagram makes that easier to spot. Read position from left to right around the ring, and time from top to bottom. A slanted band of slow colours reveals a moving disturbance. The diagram stores the latest three simulated minutes, with one row per second.
Its colours use the nearest vehicle’s speed. They do not represent a camera image, and an empty stretch of road is not literally filled with coloured traffic. The diagram is a way to inspect the pattern in the model.
3. Change One Road Setting at a Time
Traffic Jam Lab gives you several useful experiments:
- Cars on the road: change how many drivers share the same one-kilometre loop. Density is a powerful part of whether a disturbance grows.
- Desired speed: set the nominal free-road speed. Actual speed depends on the space ahead and what other drivers are doing.
- Following-time gap: change the spacing drivers aim to maintain at speed.
- Acceleration response: change how strongly drivers accelerate back toward their desired speed. The lower settings are useful for making persistent waves visible.
- Coordinated vehicles: introduce a hypothetical rule that passes some of the leading car’s braking response backward earlier.
- Scenario seed: keep the small driver differences repeatable, or generate a new arrangement.
Road-setting changes start a fresh experiment. Playback speed changes how quickly you watch the existing run. Replay inputs starts again with the same recorded brakes; Reset traffic clears those inputs while keeping the road settings.
4. A Following Gap Is Not a Reaction-Time Slider
The model is based on the Intelligent-Driver Model, or IDM. Each driver’s acceleration depends on speed, the gap to the leader, and the difference between their speeds. These local decisions combine into the behaviour of the whole ring.
The following-time setting is an IDM spacing parameter. It is not a literal delay before a person notices brake lights. Calling it reaction time would make the control sound more familiar while changing what it actually means.
Small seeded variations keep the drivers from behaving like identical copies. The model uses fixed simultaneous integration steps of 0.05 seconds, preserving the same result when an experiment is replayed at a different viewing speed.
5. Can Coordinated Cars Calm the Wave?
Choose Coordinated Comparison, or start a comparison with all-human traffic. The tool runs two matched versions: the selected coordination share and a zero-coordination baseline. The seed, density, driver traits and brake inputs stay the same.
Compare average speed over the run and mean time nearly stopped per car at the same simulated moment. These are more informative than comparing one run at two minutes with another at five minutes.
The coordination rule is deliberately explicit: a coordinated car adds part of its leader’s negative acceleration to its own response, without shortening the chosen following gap. This is an illustrative controller, not a prediction about Tesla FSD or any particular autonomous vehicle.
Real ring-road experiments have investigated how controlled vehicles can damp stop-and-go waves. That research provides useful context, but its vehicles and controllers differ from the simple rule used here. Our numerical results should not be presented as a replication or validation of those field results.
6. What the Traffic Statistics Mean
Average speed is the instantaneous mean across all cars. Cars nearly stopped counts vehicles below 2 km/h. Speed spread is the standard deviation of the cars’ speeds: a large spread can reveal a road where some drivers are moving quickly while others are crawling.
Flow estimate is vehicle density multiplied by the instantaneous mean speed, expressed in vehicles per hour. It is a model-derived quantity, not a detector counting cars past a physical point.
The matched comparison also reports a run-average speed, calculated from total distance travelled per car divided by elapsed simulated time. Mean stopped time uses stopped car-seconds divided by the number of cars.
7. Share a Repeatable Experiment
Copy replay link preserves the seed, road settings and recorded brake inputs at their exact integration steps. A friend gets a fresh run of the same experiment. The link does not connect them to a live road or secretly change the model’s randomness.
You can preview and save a 3:2 result card with the road, settings, speed statistics and model context. Include the replay link when sharing the image so someone else can investigate the cause of your impressive little traffic disaster.
8. Traffic Jam Simulator Questions
8.1 Is This a Real Traffic Forecast?
No. It is a documented numerical experiment. The ring has no lane changes, junctions, merging, crashes, exits, weather or road gradients. A minimum-gap numerical safeguard prevents cars overlapping; that does not turn the simulator into a crash-risk model.
8.2 Does It Keep Running When I Switch Tabs?
Yes. Simulated time continues, and the physics catches up when the browser resumes execution. An explicit pause stays paused when you return. Each experiment stops after ten simulated minutes.
8.3 Why Can the Same Settings Produce the Same Jam?
The seed controls the small driver differences, and brake inputs are recorded at fixed physical steps. Keeping those inputs makes the run repeatable. A new seed or a different braking moment gives you a different experiment.
8.4 Can I Use It Without Clicking Moving Cars?
Yes. Choose a car from the selector and use the brake button. Presets also schedule an initial brake input. With reduced motion enabled on your device, the road stays still while the model and statistics continue.
9. Keep Experimenting
If traffic is your favourite recurring inconvenience, try the ContentBASE Life Simulator and explore what motor vehicle deaths mean in a very different thought experiment. For a wider perspective, visit Your Life on Earth or Breathing Earth. The collection lives in ContentBASE Tools.