This is sort of like estimating the weight of a cow by assuming it's perfect sphere with a radius.
But this isn't entirely surprising. It's a common sentiment on Hacker News, where highly complex markets are reduced to incredibly simple single-variable, Econ-101-stye supply/demand curves. The poster usually then goes on to believe they've solved some major problem and wonders how everyone else can be so blind as to see such a simple solution.
But far be it to leave at just snarkiness. There are a whole slew of reasons why pricing as a lever cannot be used to fine-tune demand. Simply off the top of my head:
- Desirability of restaurants is heavily determined by their pricing context. That hole in the wall Thai place with the IKEA chairs is a great place to get $5 Pad Thai, and is always crowded. If they charged $15 the expectations for service, food, decor, location, etc, would be entirely different. Cannot simply pull the price lever.
- Desirability of restaurants is heavily determined by perceived authenticity. That Chinese restaurant where all the Chinese immigrants eat is popular with everyone because it communicates authenticity. It loses that authenticity if it prices itself out of reach of the people who lend it authenticity. Funnily enough, authenticity is frequently derived from the working-class everyman. Cannot simply pull the price lever.
- Restaurants are divided into price ranges, with not a lot of intermediate prices. This limits the "resolution" to which you can tune your pricing before you jump into the next range (which comes with expectations, authenticity, and other issues as above). There are demand cliffs moving from one range to the next - in other words restaurant pricing is not a continuous function and cannot be modeled as such. It's not even a "sorta kinda linear" curve.
- As price increases demand does not drop off linearly, or close to linearly. If you've got a line out the door for $5 burritos, a $6 burrito may cause only a small dropoff in customers, while a $7 burrito may cause your business to largely dry up. The holy grail of "that price where everyone who wants a burrito can get one without waiting very long" may not be achievable by simply pulling the price lever.
Pricing food is complex. Until the day where all food becomes standardized, commoditized, and we can treat our restaurants like we treat our nuts-and-bolts supplier, it will remain like this.
That's a strange phenomenon I've also noticed, since posters here usually have the opposite instinct in any other field. Most systems end up complex and difficult to model by simple rules, unless you restrict your model to a specific range of parameters or behavioral regime. Do you want to know how increasing or decreasing pressure affects fluid flow? Most knowledgeable people would immediately say that it depends: does the pressure change impact the flow regime? If it stays within a flow regime, there is one answer. But if it changes from laminar to turbulent flow or vice-versa, the answer may differ. What initially seemed like a simple relationship has a confounding factor of macro-scale regime-switching.
But in economics, HNers feel somehow confident just giving a blanket answer without analyzing the situation. Price and demand of restaurants are linearly related; raising minimum wages increases unemployment; etc. Blanket statements they'd never make in an unqualified form in fields they actually know about, somehow become fine in a field they haven't studied. Nobody seems to feel the need to inquire, well, is it a simple relationship, or can changes in a parameter change the operating regime? Very strange.
And after this, the minute someone opens a restaurant where food is not standardized, commoditized and that can't be treated like a nuts-and-bolts supplier... a market is (re)born.
I have to disagree with your point 2: The blue-collar working man doesn't want to queue with a bunch of students for half an hour for his food, however authentic it may be.