4 ms·

Your last comment is not actually a retort, but I'll play along for now... Some hard numbers. We can talk about "millions" of dollars in train costs as if that

Your last comment is not actually a retort, but I'll play along for now... Some hard numbers. We can talk about "millions" of dollars in train costs as if that's some bad thing, actually some back of the envelope math reveals this is a pittance. Average cost of a vehicle (generous) 30000 (actually much higher after repairs, maintenance), lets say 35000, our drivers all experience great luck and need very few repairs over the life of their vehicles. Average number of vehicles, per person, lets say, 5. So about 17500$. Lets maybe round that up for easy math, the consumer needs to also pay for insurance (150 USD a month for 12 months and about 40 years of driver time, another 72000 USD). Already, 5 consumers then cost about 1 million in car costs. Average lifetime of a train, 40 years. Average cost of a (traditional) train track about 1 million per mile. So the number here is already pretty low per mile (of track, we are assuming for simplicity the mile of track is paid for for 40 years - actual life time is usually 50-60 years, so we are about safe here). Lets say we cost fit, lets say, 20 villages along a route. Each one causes average 2 miles of track divergence (about 2 million cost per village), and because of politics/regulation, each 40k$ plot of land is closer to 100k$. Roughly 42$ million cost of fitting the track along these routes between two cities. Using our numbers from earlier, we need about 5*42 or ~210 customers over 40 years to just break even with the automobile cost(s). If we assume we just want to break even, no profit, over those 40 years, that's about 5k$ per consumer per year, or assuming weekly trips, about 96$ per trip. For 20 villages at ~1k population, I'm gonna bet a lot more than 210 consumers are going to want to use these tracks. Lets assume best case, all 20 villages of 1k average populations are utilizing these trains now, we have 20000k users now, or 95x times the room for growth. That 96$ ticket? It can be a dollar and still break even. Charge 10$ and you've covered the cost. And we haven't even gotten into amortizing the costs due to regular city traffic, or talked about real cost of ownership, energy inefficiencies, maintenance costs (hint, much much lower). The benefits are absurdly high, even after a quick back-of-the-envelope run of the numbers. You have to have some absurdly high costs or do something really dumb, like install T junctions that lead to nowhere, for a rail system to become unfeasible.

ImPostingOnHN 3y ago
Your numbers seem to assume a lot that isn't supported by the data or by the use case we're discussing. For example: > Lets say we cost fit, lets say, 20 villages along a route. Each one causes average 2 miles of track divergence Let's say, instead, they aren't along the route, because they likely aren't, and let's say it's closer to 10-20 miles of track divergence each. Even at the super, super low estimate of $1m/mile (the real numbers from recent commuter rail expansions tend to be at least an order of magnitude or two larger), that adds up to $28-38m for the track and station. Remember, the use case is a village whose taxi needs are satisfied by 1 car per village, so I don't know where you got 5 cars per person or five people per car. That means the self-driving option costs 1 self-driving car per village. Maybe 2. Additionally, this is a village whose train needs add up to fewer than 4 passengers per hour. But then you mention, > 20 villages of 1k average populations are utilizing these trains now, we have 20000k users now Here you're conflating train riders with literally everybody who lives in the village, which is a huge red flag. The demand is fewer than 4 rides per hour per village, so 20 villages adds up to fewer than 2,000, an order of magnitude smaller. Plus, now that it's 20 villages, the cost is 20*(millions to billions) instead of just (millions to billions). Those numbers just don't break even, especially when you add in the cost of maintaining those 20 stations and the additional track. Going back to the example, at an optimistic cost of $28-38m per additional station, and a ridership of 4/hr, and let's say a ticket price of $10. The math says each additional station would take 80 years to pay off, and that's if the station and rails don't cost anything to maintain. But they do, so it'd likely take hundreds of years.
- smaudet 3y ago
  > Remember, the use case is a village whose taxi needs are satisfied by 1 car per village Eh? No, absurd. I'm not sure where you got that I was limiting myself to villages that only need 1 taxi per village. I think you are thinking of some 3 person 'towns' not a standard village. Yeah sure you can dial down your prospective customers to some ridiculously low amount, but I already said, I was never talking about rail systems in obscure regions without cities. If that's your beef, you're wasting your time. There might be different reasons trains would be more economical in some settings but for someone in a hut in Siberia no of course trains are not a sane idea. > Your numbers seem to assume a lot that isn't supported by the data What data? My numbers are facts easily and quickly verified on the web, cite sources, please. Saying 'the data' is tantamount to nonsense, here. > Those numbers just don't break even Yes so for your 10 huts in the prairie your self driving car is a superior solution, yay, sarcasm. Meanwhile, my rail line is servicing hundreds of thousands of people, in cities and towns, below cost. One is far more economical than the other.
  - seanmcdirmid 3y ago
    How about Bethel Alaska? Most taxis per person, most people travel by taxi there.
  - ImPostingOnHN 3y ago
    > > Remember, the use case is a village whose taxi needs are satisfied by 1 car per village > Eh? No, absurd. I'm not sure where you got that Scroll to the top of the thread you've been discussing: > The square mile around me has fewer than 4 journeys an hour to anywhere[0] > a taxi service would be brilliant, my son could get to scouts and back under his own steam. Automated or manual doesn’t matter, but we can’t sustain a taxi service. - there’s one in a town about 6 miles away. One taxi.[1] So, not only not "absurd", but explicitly specified. ---- > I already said, I was never talking about rail systems in obscure regions This thread concerns villages, not suburbs of cities. Specifically villages with 0-1 taxis serving them, and a total trip demand of fewer than 4 people per hour. This thread is about running self-driving cars vs. train lines for that city and cities like that. You joined this thread. I think we can be forgiven for thinking you were talking about the topic of the thread you joined, and not something else :) Everybody makes mistakes, right? [0]: https://news.ycombinator.com/item?id=39333564 https://news.ycombinator.com/item?id=39333564 [1]: https://news.ycombinator.com/item?id=39333882 https://news.ycombinator.com/item?id=39333882

ImPostingOnHN 3y ago

Your numbers seem to assume a lot that isn't supported by the data or by the use case we're discussing. For example: > Lets say we cost fit, lets say, 20 villages along a route. Each one causes average 2 miles of track divergence Let's say, instead, they aren't along the route, because they likely aren't, and let's say it's closer to 10-20 miles of track divergence each. Even at the super, super low estimate of $1m/mile (the real numbers from recent commuter rail expansions tend to be at least an order of magnitude or two larger), that adds up to $28-38m for the track and station. Remember, the use case is a village whose taxi needs are satisfied by 1 car per village, so I don't know where you got 5 cars per person or five people per car. That means the self-driving option costs 1 self-driving car per village. Maybe 2. Additionally, this is a village whose train needs add up to fewer than 4 passengers per hour. But then you mention, > 20 villages of 1k average populations are utilizing these trains now, we have 20000k users now Here you're conflating train riders with literally everybody who lives in the village, which is a huge red flag. The demand is fewer than 4 rides per hour per village, so 20 villages adds up to fewer than 2,000, an order of magnitude smaller. Plus, now that it's 20 villages, the cost is 20*(millions to billions) instead of just (millions to billions). Those numbers just don't break even, especially when you add in the cost of maintaining those 20 stations and the additional track. Going back to the example, at an optimistic cost of $28-38m per additional station, and a ridership of 4/hr, and let's say a ticket price of $10. The math says each additional station would take 80 years to pay off, and that's if the station and rails don't cost anything to maintain. But they do, so it'd likely take hundreds of years.

smaudet 3y ago
> Remember, the use case is a village whose taxi needs are satisfied by 1 car per village Eh? No, absurd. I'm not sure where you got that I was limiting myself to villages that only need 1 taxi per village. I think you are thinking of some 3 person 'towns' not a standard village. Yeah sure you can dial down your prospective customers to some ridiculously low amount, but I already said, I was never talking about rail systems in obscure regions without cities. If that's your beef, you're wasting your time. There might be different reasons trains would be more economical in some settings but for someone in a hut in Siberia no of course trains are not a sane idea. > Your numbers seem to assume a lot that isn't supported by the data What data? My numbers are facts easily and quickly verified on the web, cite sources, please. Saying 'the data' is tantamount to nonsense, here. > Those numbers just don't break even Yes so for your 10 huts in the prairie your self driving car is a superior solution, yay, sarcasm. Meanwhile, my rail line is servicing hundreds of thousands of people, in cities and towns, below cost. One is far more economical than the other.
- seanmcdirmid 3y ago
  How about Bethel Alaska? Most taxis per person, most people travel by taxi there.
- ImPostingOnHN 3y ago
  > > Remember, the use case is a village whose taxi needs are satisfied by 1 car per village > Eh? No, absurd. I'm not sure where you got that Scroll to the top of the thread you've been discussing: > The square mile around me has fewer than 4 journeys an hour to anywhere[0] > a taxi service would be brilliant, my son could get to scouts and back under his own steam. Automated or manual doesn’t matter, but we can’t sustain a taxi service. - there’s one in a town about 6 miles away. One taxi.[1] So, not only not "absurd", but explicitly specified. ---- > I already said, I was never talking about rail systems in obscure regions This thread concerns villages, not suburbs of cities. Specifically villages with 0-1 taxis serving them, and a total trip demand of fewer than 4 people per hour. This thread is about running self-driving cars vs. train lines for that city and cities like that. You joined this thread. I think we can be forgiven for thinking you were talking about the topic of the thread you joined, and not something else :) Everybody makes mistakes, right? [0]: https://news.ycombinator.com/item?id=39333564 https://news.ycombinator.com/item?id=39333564 [1]: https://news.ycombinator.com/item?id=39333882 https://news.ycombinator.com/item?id=39333882