You start with $ in the bank.
You start with cofounders.
We will simulate years.
We will run this simulation times.
Selected plan:
Modify tasks:
Behold, the Ghost of Christmas Future.
Well, of my future anyway. This is a simulator. It uses the Monte Carlo method to forecast the outcome of a startup.
The arrow buttons below will walk you through a series of simulations. The results will be displayed in the graph to the left. Each simulation is run many times, and each line in the graph represents a single run. (If you only see a couple of lines, that means a couple of breakout successes have completely dwarfed the other runs.)
You can click the graph to rerun the simulation. Use the buttons below to continue this explanation.
In this first example, I simply continue to work as I have in the past. That is, I build stuff. I work on whatever seems reasonable at the time, while occasionally making half-hearted efforts at finding a cofounder.
As you can see, the expected outcome is not so good. This graph tends to show a survival rate of zero to five percent, which means I'm probably going to run out of money.
What can I do to improve my chances?
This is what I would call the Standard Plan, recommended by startup experts. It basically goes: 1. Find a cofounder, 2. Build a prototype, 3. Raise money, 4. Make something people want.
As you can see, this graph shows a much greater survival rate. But one thing that's missing is the exponential curve the previous graph had. Why is it missing?
You can click the plus button to the left of each plan to see the code it uses. For each simulated day, the simulator calls select_task on the current plan and the returned task is what is worked on for that day.
As you can see, in the HistoricalPlan, I would sometimes spend a day improving my productivity. I never do that in the StandardPlan. What happens if I do?
There's the best of both worlds.
Now, this is where people tend to start looking at me funny. Can we really expect to get exponential returns from semi-daily productivity improvement over a ten year period? Honestly, I'm not sure. But maybe.
I think the proposition is actually pretty modest. If you look at the code for the ImproveProductivity task you can see the claim is that if I spend an entire day working on improving my productivity I have a 50% chance of improving it by 1%.
By the way, you can edit the code in any of those widgets if you want to experiment with different parameters. Let me know if you produce any interesting results.