My family unit is looking for a new apartment. Out of three “professional” viewings we’ve experienced, all three have involved the agent slyly telling us that the place we inquired about was taken or occupied or otherwise unavailable — but would we be interested in looking at another, slightly more expensive place? We grit our teeth, feeling slightly exploited, but agree anyway — not that we’ve actually taken any of the up-sold apartments.
I have the sense that this is a well known real-estate “trick of the trade” — but why? Are we sure this actually benefits the agents in any way, or are they just doing it out of some now outdated instinct — the posthumous twitch of a dead commercial appendage? Surely the idea is that we’ll be so won over by the nicer apartment that the one we had intended to see will become unacceptable by comparison — but are people really so dumb and gullible that they forget about the price difference? And isn’t it insulting to be told that the home you initially wanted to see is actually a disgusting hellhole, but one that’s only $25 more per month is quite lovely? How could the market be so inefficient? Is the agent planning to rip us off on the nice apartment, or was she planning to rip us off on the cheap apartment? Either way, she’s an asshole.
In general, aren’t these petty machinations out of place in a world where real-estate agents (and used-car salesmen etc.) are ex ante thought to be loathsome and untrustworthy? I’m already suspicious, already prepared to cut things short at the first whiff of bullshit — and surely I’m not the only one! So the bait-and-switch treatment is at least as likely to make me call it quits as to grin and bear it. Not a good risk-reward for the scamster!
Real estate is terrible.
“Good” books these days don’t tend to include plot puzzles; your reward for reading is the “finely observed” writing itself. By contrast, it seems to me (and I can’t say this with a high level of confidence since I don’t actually consume much culture), “good” movies often do have plot puzzles, and the whole thing becomes an exercise in visual memory — did you notice how Scene C alludes to that mysterious thing you saw in Scene A, allowing you to infer Explanation X?! As a person with a pretty weak visual memory, I am irritated by (what I take to be) this years/decades-old trend. What’s the deal, culture people?
(I think the problem I have with plot puzzles in general is similar to the problem I have with my sense of direction when I’m with other people. On my own, I’m moderately competent, but when following someone else I just never remember to pay attention. Likewise, I notice a lot of anomalies in stories, but I tend to assume they’re things I’m misunderstanding, so I make a mental note to revisit them later but never do.)
“perfectly” combines the main interests of all self-respecting teenage boys:
a) Nah, and this sort-of gets to a comment I left out to collect more notes: (in a world where prices are primarily determined by machines) capitalism, or at least the capitalist, is fucking moot. Traders ought to be jamming this software with their sabots. (Maybe you have front-line experience of such resentment?)
I’ve detected minimal front-line resentment, I think in part for ideological reasons (it would seem very unsporting for a human stock-jobber to quibble with the market’s verdict that he or she was obsolete), in part because huge swaths of finance (especially bond markets) are still barely electronic (“Communism is power of the soviets plus electrification…”), and in part because human operators are still used to run and monitor the algorithms, many of which are just very finely honed implementations of human rules of thumb (e.g. if you have a big buy order, don’t try to execute it all at once because you might drive up the price) used as helpmeets in the course of making some big dumb bet that, for better or worse, emerged from very old-fashioned wetware. In a way this is a more discouraging outcome than your vision of a mechanical black-box central planner, because in reality there’s still very, very little money (in the scheme of the tens of trillion that are at stake) invested “systematically.” As far I can tell, the robots split into 1) a faction that basically commits automated plagiarism (“Ah ha! Some rube is trying to ‘invest’ based on an ‘idea’ — allow me to grab a third of a penny from him (relative to an unobserved counterfactual universe in which I didn’t exist) by becoming his doppleganger”) and 2) a faction that just boringly, diligently, consistently follows time-tested investing rules without overlaying any particular “narrative.” Perhaps surprisingly, the administrators of the world’s great fortunes don’t seem tremendously interested in either 1 or 2 — they’re just not that fun! So Skynet specializes in “microstructure,” while human luminaries scrutinize the credibility of Taco Bell’s counterattack on Chipotle.
It’s similar to an entity like McDonald’s — so much technology and organizational know-how and skillful social engineering marshaled toward a tragically sub-trans-human end. The more optimization power there is out there, the more embarrassed we (do? should?) feel about what we want to use it for.
P.S.: the “liboors” are funny and all, but it’s another case of nothing really mattering. All the sushi-based quid pro quo was in service of maybe but probably not actually nudging a set of interest rates a couple of hundredths of a percentage point in one direction or another. (“Probably not actually” because there was already a system in place for filtering out outliers, which the boors may not even have understood.) (Approximately) no one noticed or cared about this in real time because it was so damn picayune.
As much as 70% of price moves in commodity markets are generated by trading programs reacting to one another,
a) Is there a principled distinction between human cognitive algorithms reacting to one another and “trading programs” (designed by humans, often based on abstractions of human traders’ behavior) doing so?
b) If my 30 seconds of Google-ing don’t deceive me, this is a pretty misleading summary of the underlying study. The article’s phrasing made me think that it was basing this claim on some kind of data about who was trading with whom, but I don’t think that actually exists. Sure enough, some (smart) Swiss people just took high-frequency price data and fit it to “the Hawkes self-excited conditional Poisson model” and then used that to estimate that 60-70% of commodity price changes are “endogenous.” This doesn’t actually have anything to do with “programs” as such. The “moves” could be endogenous even before electronic trading existed. Just look at what Washington Irving said about John Law!