When will it be done? Promise a date and watch 1,000 futures
The client wants a date. You have the throughput of the last 12 weeks and a backlog. Drag the flag, run a Monte Carlo simulation and find out in how many futures your promise holds.
Steady team
The client wants to know when the 36 items of the new portal will be done. This team ships about the same amount every week.
Drag the flag to the week you would promise the client.
You can also tap a week or use the buttons.
How it works
Each scenario shows how many items the team delivered in each of the last 12 weeks, the backlog pile and the next 16 weeks. You do one thing: drag the promise flag to the week you would tell the client. You can also tap the week or use the arrow keys.
Then the game draws 1,000 futures. Each simulated week repeats a real week from the history, picked at random, until the backlog runs out. The futures drop into a histogram over the timeline, with P50, P85 and P95 marked. Finally, one of those futures becomes reality and the backlog burns down week by week until delivery.
What you learn in the three scenarios
- Scenario 1: steady teamEven with little variation, the date based on the average holds in only about half of the futures.
- Scenario 2: same average, more swingAverage throughput is the same, but the range of possible dates gets much wider. The average hides the risk.
- Scenario 3: scope growsThe backlog gains about 20% midway, as in the team’s previous projects. The forecast has to include that growth.
Each scenario’s score rates the decision, not luck: it is 100 when the flag sits between P80 and P90. A promise that is too early breaks trust; one that is too late loses the deal.
Who it is for
For people who hear "when will it be done?" every week and still answer with a single date, for teams that already track throughput and want to swap estimates for forecasts, and for managers who need to explain to a client why a range of dates is more honest than a guess.
Using it in a class or workshop
In 20 minutes: everyone plays the three scenarios with no guidance and writes down the weeks they promised. Then the group compares who promised on the average and who promised near P85, and discusses the closing question with a real date from their own team.
What the game simplifies
The future looks like the last 12 weeks, every item counts the same and there are no vacations or changes in team size. The simplifications are listed at the end of the game.
Keep practicing
- Forecast when it will be done →Monte Carlo using your team throughput.
- Measure what the system delivers →Throughput: the basis of every delivery forecast.
- See how WIP changes lead time →The gap between the time the customer feels and the working time.
- Run the kitchen at Kanban Pizzaria →Six nights at a pizzeria to feel WIP limits, bottlenecks and classes of service. No sign-up.
Frequently asked questions
What is Monte Carlo forecasting?
It means drawing many possible futures from the team’s real data. Here, each simulated week repeats the throughput of one week from the history. With 1,000 futures you see the odds of finishing by each week, instead of a single number.
Why promise on P85 and not on the average?
The date based on the average holds in about half of the futures: a coin flip. P85 is the week by which 85% of the futures have finished. You accept a small, known risk and still offer a competitive date.
What is throughput?
It is the number of items delivered per period, for example per week. It feeds the forecast because it shows what the system actually delivers, without relying on effort estimates.
Do I need an account to play?
No. Your best result is saved in your browser. With a free account, the game also counts toward My Progress.
How do I forecast with my team’s data?
Use the forecaster at /simular: enter the throughput of the last weeks and the backlog size, and it returns P50, P85 and P95 with the same method as the game.
Now with your team’s data
Enter the throughput of the last weeks and see P50, P85 and P95 for your backlog.
