Readings: Innovation, Robots, Inflammation, Inbreeding, Queuing, etc.

I think a lot about lines, and most of life is lines (queues, if you’re from the UK). They are everywhere: merging onto a freeway, on-phone hold times, why a laptop is so hot, chatbots at online retailers, weekend dim sum, a ski lift will open on a powder morning, and so on. These are all lines, with many similar and important properties.

It is less well understood than it should be, but one of the most important developments of the last few decades is that of queuing theory, a body of research that allows us to engineer systems to deal with stuff showing up, stuff waiting for service, and then said stuff leaving. The “stuff”, of course, can be human, can be cars, can be packets on a network — it can be many things — but they’re all often in queues. And like so many things at which we can throw algorithms, the cost of engineering a queue has collapsed in recent decades: we know more, and we can manage them less expensively. To a first approximation, queues are cheap and everywhere.

An example will help. Most queues can be characterized via two parameters: arrival times and service times; the rate at which things show up in the queue, and the speed with the queue is processed. Both of these parameters have distributions. The simplest version comes when people (let’s use that example) show up a constant, predictable rate, and when service rates are also fixed. Most real-life scenarios aren’t like that, of course, and both arrival rates and service rates vary wildly, but can be assigned distributions, like the exponential, that make the problem tractable.

So, here is a simple example. Both arrival and service times are exponentials, and I’ve set them equal. What’s interesting inthis simple example is how queues build and dissipate, even though the arrival and service rates are equal. All it takes is a few complex cases, or a few extra people showing up — both of which are predictable given the underlying distribution -= and suddenly people are waiting longer than expected.

Things get much messier with no underlying parameter changes. In the following case, lines ballooned, even though the arrival rate and the service rate are still the same.

What’s interesting about queuing theory is how useful it is, and how easily you can model real-world situations in ways that make important problems kinda go away. But the problems haven’t really gone away. All it takes is a temporary change in the parameters to make everything go bananas — an accident on a freeway, a stuck process on a computer, a powder day, a market crash, etc. — and the queues run wild. There is a kind of hidden wildness in systems that disappears, until it doesn’t.

Look around you for queues. You will find them everywhere, and we are obsessed with making them more efficient. This has consequences, and only subtle changes will throw us into entirely new regimes.


A few papers worth thinking about:

The Geography of Unconventional Innovation

I have long argued that the primary benefit of doing startups in dense cities (to a point) is that it increases the likelihood of “collisions”, of people, by chance, running into other people. That can lead to an exchange of ideas, which provokes new ways of thinking, and sometimes turns into innovations. I am pleased to see this being born out in a new study, that shows that overall innovation is flatter than is often modeled, but “atypical innovations” are more closely associated with dense urban areas where collision foment them.

Psychological reactions to human versus robotic job replacement

People are funny about how they feel about robots and jobs. For example, they prefer that other people’s jobs aren’t replaced by robots, but they would prefer their own job, if it gets eliminated, be replaced by robots than by humans. Why? The authors argue that “being replaced by machines, robots or software (versus other humans) is associated with reduced self-threat”. This is intriguing, and not at all what we usually think happens.

Association of Blood Marker of Inflammation in Late Adolescence With Premature Mortality

We are becoming increasingly aware of health problems tied to inflammation, and the problems apparently start even earlier than previously known.. According to a new paper using erythrocyte sedimentation factors (how quickly red blood cells fall to the bottom of a test tube as a proxy for inflammation) in adolescence are highly predictive of death due to cancer and cardiovascular disease.


A provocative, rich, and fascinating talk by Danny Hillis on how complexity emerges from simplicity, whether we are talking about life, technologies, or almost any other common system.

Other reading:

Live Through This: Courtney Love at 55
• Financial analysis of transfer values in the top 5 football leagues
Extreme inbreeding in a European ancestry sample from the contemporary UK population

Readings: Cancer, statins, and CFOs

“Once for all, he accepts the stock of commonplaces, prejudices, fag-ends of ideas or simply empty words which chance has piled up within his mind, and with a boldness only explicable by his ingenuousness, is prepared to impose them everywhere”
― José Ortega y Gasset, The Revolt of the Masses

We live in strange times. As I write this, the UK is debating yet another Brexit-related motion, meanwhile, in a kind of streaming seppuku, one of its own members crosses the floor and makes the current UK government largely a non-government. All the while, UK politicians continue to genuflect in the general direction of The Voter, promising that they are doing  what The Voter wants, but no-one has any idea what The Voter wants, unless it’s what said politician thought in the first place.  Sometimes we hear about Ordinary Constituents who send Letters too, but it’s not clear whether Ordinary Constituents and The Voter are fungible on a 1:1 basis. It is a remarkable moment, combining sophistry, arrogance, confusion, and populism.

Of course, it’s not just Brexit. There is a sense of fragility in much of current events, from politics, to economics, to the environment. Society feels as if it’s going through a Galloping Gertie phase, where under the pressure of sustained pressure systems are finding harmonics they didn’t previously realize existed, and oscillating in increasingly destructive amplitudes. 


A few papers and articles worth reading: 

Variations in common diseases, hospital admissions, and deaths in middle-aged adults in 21 countries from five continents (PURE): a prospective cohort study

While cardiovascular disease retains its top spot worldwide with respect to killing humans, it is important to note (as the above study does) that cancer is no longer the top cause of human death in higher-income countries. No, that position has been taken by cancer, which recently passed cardiovascular disease in those countries. This is an artifact of an aging society as much as anything else, but it is noteworthy.

Seeing is believing? Executives’ facial trustworthiness, auditor tenure, and audit fees

Some truly fascinating and bizarre research is coming out of the financial community lately as it exploits new sources of data. This is that, as researchers show that the firms of CFOs with more “trustworthy” faces — as judged by a machine learning algorithm — are charged 5.6% lower audit fees. Perhaps unsurprisingly, facial trustworthiness is shown to have no association with either financial reporting quality or litigation risk.

Cross-national evidence of a negativity bias in psychophysiological reactions to news

I am generally of the view that you should ignore news, perhaps checking in once a year or so, and even then maybe only quickly scan news from twelve months before that. Too much recent news is irrelevant, inflammatory, unimportant, etc. This new study reminds us of that, showing how skewed it is to negative news — despite there being a large and mostly untapped audience for positive news.

Do statins really work? Who benefits? Who has the power to cover up the side effects?

There is a punchy and spot-on critique of statins, their role in health, and their unanticipated consequences. It seems increasingly clear that such drugs are overprescribed, and that our cholesterol-lowering fixation is killing people, whether directly or indirectly.

Fully automated snow depth measurements from time-lapse images applying a convolutional neural network

I’ve long thought that computer vision is a much more important technology than it’s usually given credit for being. It is an example of a general-purpose technology, one that can absorb many other modes via which we learn about and understand the world in which we live. Combine computer vision, machine learning and, okay. snow, and you’ve really got my attention.

Readings: Four papers, back from hiatus, etc.

Summer’s ending, or at least the kids-out-of-school part is, so this newsletter will get going again. Sorry about that — either the hiatus or the recommencement: your call — but things will recommence next week.

To give you a taste, here are a few papers I enjoyed reading this weekend . Next edition will go back back to the usual preamble followed by a few interesting papers.

Hope you had a great summer.


While a sheet of pluripotent stem cells engineered to become corneal cells hasn’t return this Japanese woman’s vision to normal, it has arrested decline. Only a month post-surgery, but this is fascinating and important work.

Fear-mongering about food has moved onto pet food, with expensive consequences. Pets,to a first approximate, are scavengers — this is how they survived and formed human attachments. Pretending otherwise is expensive silliness.

A seminal paper on tissue properties from a pioneer of biomedical engineering, YC Fung, who turned 100 this week. A lovely man, with whom I have had the good fortune of spending a little time.

Feedback from skin on the foot is highly specific to the region of the foot, turning feet and limbs into sensors, with important implications for walking and running.

A new algorithm seems able to extract tradable sentiment data from news feeds. While this is not new, most prior such models either found weak signals, no signal, or were so temporally unstable as to be worthless.

Readings: Cheating CEOs, Inflammation, Seedy hedge funds, etc.


Here are a few graphs that caught my eye this week.





The Best Books on Modern German History


The Bob Emergency: a study of athletes named Bob


Readings: Bullshitters, Foreign-body ingestion, & Market manipulation

Catching up after some time away. A few things worth reading, some graphs, some moving pictures, and crampons.





Graph: H&R Block lobbying spending
Graph: Worker-to-consumer ratio peaking


Deadwood: The Movie (2019). Official Trailer.

Peter Thiel: Technology should treat death as the enemy


Gear: Black Diamond Neve Pro crampons
Book: Routes of Power
Shoes: Nike Terra Kiger 5

Readings: Sabotage, Crap econ model, PRP, and Mass fetishization


While it might seem a good idea to relocate to a small town with lovely views, cheap real estate, air service, a decent hospital, and Amazon Prime delivery, you’re missing at least one thing. Towns without said thing are a risky destination for you, even if you are crazy enough to build a spreadsheet and spend fifteen years looking.

That thing? A high likelihood of people  — especially young people — getting hurt in disproportionate numbers. The presence of that cohort — in a time of aging populations — brings in expensive doctors, which helps make hospitals viable, which helps make towns viable, which helps make relocatees … viable.

Viable hospitals matter for health, but they also matter for a local economy. After all, small-town hospitals are often the largest employers in their region, and their disappearance cuts into the economy, as well as scaring off would-be relocators. A double whammy.

This is why small places like Mammoth and Jackson are, perhaps surprisingly, good relocation bets, while, say, Celina, Tennessee, which recently lost its hospital, is much less so. What Celina needs, to my way of thinking, is a few bike parks, a wingsuit club, possibly a thriving surf scene, and maybe an indoor ski hill or a top skate park.

While I’m mostly kidding, I have a genuine point. Towns that want to remain vital and viable in the upcoming Dual Age of Senescence and Remote Work are going to have to move closer to a societal risk frontier, one that they may not like.






Hedge fund investor, scientist, and philanthropist Dr. Jim Simons on the life and career of Dr. Jim Simons.

Readings: Hospital distress, Rogue finance academics, and the Cult of Peloton


Fixed cost businesses often do surprising things, like go bankrupt just when you least expect it. We have seen that recently in the profusion of European discount airlines that, despite being crowded with passengers, have recently gone under, in a few cases they went under so speedily that people were left on the wrong end of a trip with no way to come back again. This might seem obvious, but people too often miss that point when noticing how busy said businesses are. We constantly conflate the busyness of a business with its profitability. They’re not the same. At all.

Putting this in accounting terms, a big problem with a fixed cost business is, of course, that their costs are, to a large degree, fixed. This is also a huge advantage. You can’t put too fine a point on this, the idea that some businesses have costs that mostly change with the number of customer san orders, and some businesses’ costs don’t do that. For example, a family-owned furniture maker might be an example of the former, where its costs more or less run coincident with the number of pieces of furniture they make. Granted, as long as they keep their costs less than revenues, things work out, but there is no sudden inflection in their business when BUCKETS OF MONEY POUR DOWN FROM THE SKY.

Buckets of money can sometimes pour down from the sky in fixed cost businesses, however. Like software, or airlines, or, today’s example, hospitals. If, say, you run a 500-bed hospital, and your fixed costs are about 80% of revenues, then once you have those fixed costs covered, and you continue filling beds with paying patients, you, to a first approximation, get BUCKETS OF MONEY POURING DOWN THE SKY. Sure, there are some per-patient costs, like laundry and bed-pan-emptying, but those pale against the fixed costs, things like HVAC, doctor and nurse salaries, 24-hour staffed emergency roos, gigantic machines that go PING, new construction wings with huge letters on them, insurance, and so on.

So, this raises an interesting question: If hospitals are, like most fixed cost businesses, and, unlike most fixed costs business, hospitals seemingly have a limitless supply of people who want their services, why do hospitals go bankrupt? And, more specifically, why are so many hospitals currently in financial distress? That is the question I asked me recently when I saw yet another headline about a bankrupt hospital, and saw that distressed filings for hospitals were up to almost 13% of total filings, from 1% a decade ago, as you can see in the graphic below. Healthcare service distress is soaring.

What is going on here? In short, this is what happens when fixed cost businesses’ costs go up and their customers go down. Hospitals are consolidating across the country, and no-one wants to own hospitals in rural areas where populations are emptying out, as you might expect given the high fixed costs of operating a hospital. Even the self-styled entrepreneurs buying up distressed hospitals in rural areas are mostly themselves becoming distressed. At the same time, small hospitals have no pricing power, and are forced to rely almost entire on Medicare/Medicaid payments, which makes them even more reliant on aging locals whose numbers are declining. It’s a vicious circle in a precarious fixed-cost business.

You can see some of the consequences in this simple two-way sensitivity analysis I did of a 500-bed hospital. Given some fairly standard assumptions about revenue per patent, patients per year, and the distribution of fixed versus variable costs, the profits and losses produced can be prodigious.

You can expect more of this sort of thing. Rural areas aren’t going to stop emptying out, and hospital costs aren’t getting any less fixed. The US is, as a result, wildly over-hospitaled, in much the same way that Europe has been wildly over-airlined. The result is financial distress, and a lot of it.





The Accidental Homo Sapiens: Genetics, Behavior, and Free Will. What we inherit, and what it has meant for us.

Readings: Enron, the Fermi Paradox, Discount airlines, & Parasites


Major stadiums have names that get said a lot, and so over the last few decades most of those names have been purchased, usually by large companies who think it would be good to have people say their names a lot. This sometimes creates issues, given what might euphemistically be called temporal and objectives mismatches between companies and stadiums.

First, companies tend not to last as long as stadiums. This might seem obvious, but the difference can be starker than you might think. For example, the Houston Astros got a sweet, 30-year naming deal in 1999 for its stadium. Sadly, however, the sponsor was a company called Enron, and only a couple of years later said company no longer existed, for practical purposes. This sort of thing happens fairly regularly, with once-highflying companies no longer in a position to exist, so their naming rights revert or lapse.

Second, fans and even teams are sometimes not such big fans of the companies that bought their stadium’s name. A good example is that of the Philadelphia 76ers, whose owners have long refused to call the place where they play the Wells Fargo Arena, and instead use euphemisms like “the arena where we play”. While impressively passive-aggressive, this does get in the way of Wells Fargo extracting maximum value for its stadium sponsorship. Some companies try to get around local truculence by purchasing naming rights to multiple stadiums, even in the same league, like the way American Airlines has done with rights to both Miami and Dallas.

The most recent example of name game fun is that of United Airlines and its $69-million deal for naming rights to the Los Angeles Memorial Colosseum. Under the deal with USC — part of a $270-million renovation — the stadium was to be renamed the United Airlines Memorial Colosseum. This is a strange name, to be fair, sounding more than slightly like it’s a memorial to United Airlines, like United has gone away, which it hasn’t, at least not yet.

Local politicians and veterans aren’t happy, however, arguing that the new name gives short shrift to the key “Memorial” part of the name. The enterprising development people at USC have proposed changing the name, to something more like the United Airlines Field at Memorial Colosseum, which is a) a handful, and b) likely to, in practice, make the United Airlines Field part ignored, which already has United wondering why it should pay $69-million anyway. Perhaps unsurprisingly, United has offered to back out of the whole deal, which is causing more scrambling.

Given all of this, why do companies persist in purchasing naming rights for stadiums? The direct economic rationale is dodgy, given that it’s not because having naming rights adds persistently to company profitability. Instead, it seems to be a combination of corporate ego, stadium owner greed, and small numbers of exclusive goods for sale.  All of this suggests that nothing is going to change, and companies will keep purchasing naming rights, which will continue annoying fans, team, and politicians, but producing entertaining Wikipedia articles.




Readings: MBAs, Sewage, Cheating, and Fear Factor


How much finance is too much finance in an economy? Trick question: It’s enough finance when you get a finance job that pays you lots of money to do very little (or when your do-nothing company goes public ). But it’s too much finance, of course, when that idiot you once knew in college turns out to be working at Goldman and making bank.

Finance employment remains a main target for young people, especially recent MBA graduates, especially recent Harvard MBA graduates. If you combine Financial Services and Consulting, which, c’mon, are both just finance, in real people terms, then more than half of the Harvard graduating class has been going into finance for decades. And for good reason, of course, given that Harvard is expensive and finance pays a lot, with new hires often able to earn twice as much as there as in other fields. This is, of course, ridiculous, but it’s also kinda hard to ignore. See also: Student debt.

Many people (including me in this paper) have argued that the increasing financialization of the economy isn’t such a great thing. It diverts people from more productive things, like building products and services people need; it reduces the number of entrepreneurs; it leads to asset inflation; and so on. To be fair, it also means that most truly annoying finance people end up living in Manhattan and away from the rest of us, so it’s got that going for it, but that may not be enough.

So, how much finance is too much? I tried to answer that at the outset, but that my-college-idiot-friend-at-Goldman answer is … not entirely satisfactory. A recent paper takes another tack, arguing that increased financialization leads to more liquidity — more trading activity — and that is where you can draw a kind of bright-ish line. The authors show that, across 136 countries studied over the period 1961-2015, when liquidity surpassed a little more than 100% of GDP it started having a negative effect on GDP growth. And given that more financialization breeds more financialization, which brings more liquidity, countries, all else equal, are pulled inexorably toward that cliff of Too Much Trading Of Everything.

This suggests that there is a role for Tobin taxes and that sort of thing, but I won’t jump off that particular policy cliff today. Instead, let’s bring it back to that idiot college friend of yours. If he goes to Goldman, worry; but if he starts a new trading firm, panic. QED.




This is an instant classic, with medical historian LIndsey Fitzharris talking to Joe Rogan about the history of surgery. It’s like Fear Factor, but way, way too real.

Readings: Avalanches, Pay TV, Exercise, Robots, & Greeting Cards


Being caught in a small slab avalanche while skiing is like sliding across a steep room as someone pulls the carpet out, and then said carpet fractures and accelerates with you on bits of it, all going very fast in a direction you weren’t planning on going. (Being caught in a large slab avalanche is like having the surface of the earth pulled out from under you, then fracture into blocks, then accelerate to 60 mi/h with many trees and rocks over cliffs & through more trees and rocks, before burying you and then freezing solid.)

Neither is recommended. My perspective, which is that of sociologist Diane Vaughan, is to be aware of base rates (how statistically likely bad outcomes are), and never expand the risk envelope, no matter how many times you get good outcomes. Vaughan called the tendency to do the opposite — to respond to lucky non-failures by expanding the launch conditions — as “normalization of deviance”. She used an ethnographic study of the launch of the space shuttle Challenger to show how it happens, with “successes” causing steady and inexorable expansions of launch conditions, until there was a catastrophica failure.

Until your own risk-taking catches up with you, it is easy to convince yourself of two things: one, that you’re smart enough to operate near the risk frontier while knowing you’re there; and two, that you could remedy things if the worst happens.

Both of these are bad ideas. First, the only way to control backcountry risk is to know, as best you can, what they are, and then back way, way off. This is one of the reasons why small groups are safer than large ones: there is less risk-taking pressure. Second, having even a small slab fracture under you is instant mayhem, like being a small dish when a drunk, after-dinner magician pull out the tablecloth. People who think they can predictably ski off a small slab fracture watch too many YouTube avalanche videos.

Happily, however, you don’t need many experiences — ideally: none — with slab avalanches before you decide to never put yourself in that position again. Both of my small avalanche experiences were years ago, which, in terms of risk, is a very good thing. In any case where outcomes tend to be binary and too often terminal, taking a Kelly criterion approach to risk — let alone expanding the risk window — can be a very bad idea indeed.


  • Eccentric exercise has long been the gold standard when treating Achilles tendinopathy, so it’s surprising to see that pressure massage delivered similar results in a recent study. Granted, a small study size, but it does speak, at the very least, to our continuing profound ignorance about tendon disorders.
  • Surgeon promotes fraudulent research that kills people; his employer, a leading hospital, defends him and attacks whistleblowers. Business as usual.