[AnimalMuppet]

All true. But if you're one of the folks who regularly buys a new car, you kind of know that. If you're one who keeps a car as long as you can, you know that too. So you can make, not totally accurate calculations, but calculations that are enough within the ballpark to show you whether transit is an economically interesting alternative.

[dataflow]

How accurate is "enough within the ballpark" though? This isn't something where an order-of-magnitude estimate or even a factor of (say) 2 error is negligible. Can you really predict the future to within like 10% accuracy? How accurate do you think you need to be?

Let's do something resembling reality. Say your subway transportation costs around $13/weekday. (I just looked up a random round-trip BART ticket for the Bay Area, in case that makes it easier for people here to relate.) Probably more like $20/weekday if you add bus and parking.

Now say the drive for that is around 30mi. At $4/gal gas (pre-2022) and 30 mi/gal the driving cost might cost you $4/day just for gas (probably optimistic). Say you buy a $35k car and plan to keep it 8 years, then sell it for $5k. That'd cost you roughly $14.50/weekday in depreciation. Add that to the gas price and it's around $19.50/day.

I haven't even accounted for maintenance yet (on the car side), nor for non-work rides (on the public transportation side), but so now you have ~$20/day vs. ~$20/day. How in the world do you decide?

Now consider gas prices shot up > 50% in a year. And we have a recession coming and you might lose your job (less gas usage, I guess?) and so decide to keep your car longer. And we have inflation. And your car might get totaled in between... or not. And you might have random stuff come up that increase costs on each side... or not. Exactly how do you do the math here to figure out which one is more beneficial?!

I'm not saying you can't do the math, but as I see it, every factor adds in such a huge margin of error into both the numerator and the denominator that the result of the computation easily becomes pretty useless. And this is before even accounting for convenience/QoL improvements that you can't put into numbers... what's the value of being able to get your kids their favorite meal, or show your friends around town, or...

[markvdb]

> I'm not saying you can't do the math, but as I see it, every factor adds in such a huge margin of error into both the numerator and the denominator that the result of the computation easily becomes pretty useless.

You just described TCO, unpredictability in TCO, capital intensivity and several kinds of risk. All super useful heuristics to guide a decison. Just don't try to squeeze them into one filter...

Take us deciding on buying a car versus car sharing. We wanted low TCO, low cash flow risk, and low environmental impact. The signal from the low TCO filter was slightly ambiguous, but both other filters were super clear. Car sharing it was.

[_carbyau_]

The problem is economically interesting is only part of the equation for getting people to make the jump. The sheer convenience of a car is priceless to many people.