3 ms·

I'm suspicious of the "fundamental law of road congestion". I suspect what the law reflects is that freeways don't get widened until it's so overdue that the w

by tlb 9y ago

I'm suspicious of the "fundamental law of road congestion". I suspect what the law reflects is that freeways don't get widened until it's so overdue that the widened version is already over-capacity. If someone complained that their web server was overloaded, and they tried increasing capacity but it was still overloaded so there's no point, you might say "was this at Friendster?"

eesmith 9y ago
I believe your objection is that widening heavily congested roads doesn't help, and that only widening lightly congested or non-congested roads improves traffic conditions, yes? In that case, the linked-to report gives tepid support for that possibility, saying: > This (admittedly weak) trend downward suggests the conjecture that more roadway can lead to a more than proportional increase in traffic when roads are not congested. Alternatively, it may be that the most useful highway segments are developed earlier and receive more traffic. This second conjecture is consistent with John G. Fernald’s (1999) conclusion that the productivity effects of the US interstate system show a marked decline over time. Do note that the concerned cities are already congested, so your objection is neither here nor there. The economics model is not complicated. People are willing to spend a certain amount of money to commute. If the cost of travel goes down, they are more likely to travel longer distances because they can use the extra money on, say, a larger house. The question is, what is the shape of the demand curve? The linked-to paper tries to tease out those numbers, and show that increasing capacity doesn't meet the increased demand. I'll try to extend it to your web server analogy. People are willing to wait 0.5 seconds for information, and it's not linear. That is, people don't care if it takes 0.01 or 0.4 or 0.5 seconds, but they do care if it takes 1 second and really care if it takes over 10 seconds. (See https://www.nngroup.com/articles/response-times-3-important-limits/ https://www.nngroup.com/articles/response-times-3-important-... ) If you increase the capacity of the system which can respond in 0.5 seconds, then you can give a response in 0.3 seconds. Or, you can add more functionality and data in the same 0.5 seconds. Thus, it's possible that adding more capacity results in saturating the new capacity. (This isn't a great analogy. What makes it worse is that most people don't expect to pay for, say, Facebook. While the roads are paid for mostly by the users of the roads, through taxes.) In any case, what is your interpretation of the increased congestion of the 23-lane Katy Freeway after widening? "But no matter how wide it gets, added capacity just induces further flung development and more peak hour driving, with the result that the freeway is even slower today than it was when it was widened just a few years ago." I think you are suggesting it should be widened again? No other city needs such a wide freeway. BTW, when you write "I'm suspicious of ..." and present a conjecture about why a blog written by expert in the field, backed by a scholarly paper in the field, is incorrect, well, I can't help but wonder why you think your untrained intuition should carry any weight for other HN readers, or why you decided to not read the supporting documents.

by tlb 9y ago

eesmith 9y ago

I believe your objection is that widening heavily congested roads doesn't help, and that only widening lightly congested or non-congested roads improves traffic conditions, yes? In that case, the linked-to report gives tepid support for that possibility, saying: > This (admittedly weak) trend downward suggests the conjecture that more roadway can lead to a more than proportional increase in traffic when roads are not congested. Alternatively, it may be that the most useful highway segments are developed earlier and receive more traffic. This second conjecture is consistent with John G. Fernald’s (1999) conclusion that the productivity effects of the US interstate system show a marked decline over time. Do note that the concerned cities are already congested, so your objection is neither here nor there. The economics model is not complicated. People are willing to spend a certain amount of money to commute. If the cost of travel goes down, they are more likely to travel longer distances because they can use the extra money on, say, a larger house. The question is, what is the shape of the demand curve? The linked-to paper tries to tease out those numbers, and show that increasing capacity doesn't meet the increased demand. I'll try to extend it to your web server analogy. People are willing to wait 0.5 seconds for information, and it's not linear. That is, people don't care if it takes 0.01 or 0.4 or 0.5 seconds, but they do care if it takes 1 second and really care if it takes over 10 seconds. (See https://www.nngroup.com/articles/response-times-3-important-limits/ https://www.nngroup.com/articles/response-times-3-important-... ) If you increase the capacity of the system which can respond in 0.5 seconds, then you can give a response in 0.3 seconds. Or, you can add more functionality and data in the same 0.5 seconds. Thus, it's possible that adding more capacity results in saturating the new capacity. (This isn't a great analogy. What makes it worse is that most people don't expect to pay for, say, Facebook. While the roads are paid for mostly by the users of the roads, through taxes.) In any case, what is your interpretation of the increased congestion of the 23-lane Katy Freeway after widening? "But no matter how wide it gets, added capacity just induces further flung development and more peak hour driving, with the result that the freeway is even slower today than it was when it was widened just a few years ago." I think you are suggesting it should be widened again? No other city needs such a wide freeway. BTW, when you write "I'm suspicious of ..." and present a conjecture about why a blog written by expert in the field, backed by a scholarly paper in the field, is incorrect, well, I can't help but wonder why you think your untrained intuition should carry any weight for other HN readers, or why you decided to not read the supporting documents.