We’re building a tool to let anyone create models like this one — models that incorporate uncertainty in their variables and outputs. This is the kinda stuff people do in Excel right now, with much difficulty.
We’re working on a couple more demo models — let us know if you have any suggestions. And if this kind of tool resonates with you, then we’d love to chat — hi@causal.app. Thanks!
I think the examples you've chosen are good but not great examples of where you can add significant value. A house being rented where there are a few variables and a good/better/best outcome is not especially gnarly Excel for the Hacker News audience. I think marketing and sales forecasts and stock option pricing and DCF might be more impactful, if more esoteric, in showing the power of what you're building.
BTW, it'd be cool to be able to anchor a variable to a historical base or average. So maybe if you have seven or eight forecasts of CAGR, Casual could accept 'citations' and in its output show a user of the model that this isn't totally speculative.
Love what you're doing and have signed up. Well done again.
You’re totally right that these aren’t the best examples to show the real value of the approach. We wanted to keep things super simple to start getting feedback, but are working on a more involved example now. DCF valuations and options pricing sound like great ideas – will look into them!
What do you mean re: citations? Validating Causal models using forecasts from other means?
To me, one of the advantages of representing the story in a probabilistic way, similar to how you have shown it, is that I am able to sell an idea to my audience.
If you are looking for use cases to build for a demo, in addition to DCF, as someone suggested before, you might want to consider currency movements and also IRR calculations.
One feature request, if I may, is the ability to manipulate the end result (time horizon in the demo of rent vs buy) and see how it impacts one of the features (rent, market return, etc).
All the best and glad to discuss.
Signed up, curious to see how the sausage is made.
Example scenario: should I pay off my mortgage, which is in a foreign currency?
The bottom line isn't clear because there's too much information. I'm expecting a single number in 60-point font like "You'll save X amount of money by renting instead of buying". Some information can be hidden behind progressive disclosure.
Also, it's US-specific. Let me choose a currency, and format numbers accordingly, like "₹2 crore", not "$20m". In India, we use lacs and crores, not millions.
I also like that you're using Monte Carlo simulations. Though it would be good to augment that with real-world data from the user's country. If I'm in India, which Indian city has had the worst-case real-estate investment performance? The worst-case is useful information to consider when making an investment.