Self-generated Seasonal Cycles

Why is Black Friday the biggest shopping day of the year? Back in 1961, Jay Forrester identified an endogenous cause in Appendix N of Industrial Dynamics, Self-generated Seasonal Cycles:

Industrial policies adopted in recognition of seasonal sales patterns may often accentuate the very seasonality from which they arise. A seasonal forecast can lead to action that may cause fulfillment of the forecast. In closed-loop systems this is a likely possibility. … any effort toward statistical isolation of a seasonal sales component will find some seasonality in the random disturbances. Should the seasonality so located lead to decisions that create actual seasonality, the process can become self-regenerative.

I think there are actually quite a few reinforcing feedback mechanisms, some of which cross consumer-business stovepipes and therefore are difficult to address.

Before heading to the mall, it’s a good day to think about stuff.

Update: another interesting take.

The Insidious Dynamics of Driving to School

When I passed by my old high school a few years ago, I was astonished to see that they’d paved over a nice grass field to make room for a vast parking lot, which must be for students. There’s really no excuse for driving to school in Palo Alto, CA – the weather is great, it’s flat, and no one lives more than a couple miles away.
Most of the responsibility falls to this nest of positive feedback loops:

I’ll start with a perception: parents worried about the safety of their kids start driving them to school (or, in Palo Alto, buy them a BMW so they can drive themselves). All that extra driving adds to traffic density, reinforcing the perceived danger on the roads. Over the long haul, all that traffic demands more lane space, so bike lanes and sidewalks get crowded out. And who wants to bike next to a bunch of hot, smelly tailpipes?

The more students drive, the less fit they get, which diminishes the fun of riding. They also become less tolerant of weather – in spite of Gore Tex, a lot of people react to a little water falling from the sky like the Wicked Witch of the West.

The result of all this is a kind of phase transition – at some point, conditions are right for all these positive loops to kick in, and everyone shifts from bike-dominated transport to driving.

This transition should not be irreversible, if one is patient. One can move the point at which the phase transition occurs, to encourage bicycling. I think there are two leverage points. First, a society that can afford cars for kids can afford to provide Dutch- or Danish-style traffic separation, breaking the safety loop and decreasing the attractiveness of driving by removing traffic lanes, which causes congestion until people go back to bikes. Second, make the cars pay for the infrastructure they use and the environmental and safety externalities they cause. Once people are back on bikes, they’ll get fitter and healthier, and the positive loops will help lock in a more sustainable mode.

Inspired by a comment in Bert de Vries’ talk this morning at the 30th Balaton Group meeting.

Thinking about stuff

A while back I decided to never buy another garden plant unless I’d first dug the hole for it. In a single stroke, this simple rule eliminated impulse shopping at the nursery, improved the survival rate of new plants, and increased overall garden productivity.

This got me thinking about the insidious dynamics of stuff, by which tools come to rule their masters. I’ve distilled most of my thinking into this picture:


Click to enlarge.

This is mainly a visual post, but here’s a quick guide to some of the loops:

Black: stuff is the accumulation of shopping, less outflows from discarding and liquidation.

Red: Shopping adjusts the stock of stuff to a goal. The goal is set by income (a positive feedback, to the extent that stuff makes you more productive, so you can afford more stuff) and by the utility of stuff at the margin, which falls as you have less and less time to use each item of stuff, or acquire increasingly useless items.

So far, Economics 101 would tell a nice story of smooth adjustment of the shopping process to an equilibrium at the optimal stuff level. That’s defeated by the complexity of all of the other dynamics, which create a variety of possible vicious cycles and misperceptions of feedback that result in suboptimal stuffing.

Orange: You need stuff to go with the stuff. The iPad needs a dock, etc. Even if the stuff is truly simple, you need somewhere to put it.

Green: Society reinforces the need for stuff, via keep-up-with-the-Joneses and neglect of shared stuff. When you have too much stuff, C.H.A.O.S. ensues – “can’t have anyone over syndrome” – which reinforces the desire for stuff to hide the chaos or facilitate fun without social contact.

Blue: Stuff takes time, in a variety of ways. The more stuff  you have, the less time you actually have for using stuff for fun. This can actually increase your desire for stuff, due to the desire to have fun more efficiently in the limited time available.

Brown: Pressure for time and more stuff triggers a bunch of loops involving quality of stuff. One response is to buy low-quality stuff, which soon increases the stock of broken stuff lying about, worsening time pressure. One response is the descent into disposability, which saves the time, at the expense of a high throughput (shopping->discarding) relative to the stock of stuff. Once you’re fully stocked with low-quality stuff, why bother fixing it when it breaks? Fixing one thing often results in collateral damage to another (computers are notorious for this).

I’m far from a successful minimalist yet, but here’s what’s working for me to various degrees:

  • The old advice, “Use it up, wear it out, make it do or do without” works.
  • Don’t buy stuff when you can rent it. Unfortunately rental markets aren’t very liquid so this can be tough.
  • Allocate time to liquidating stuff. This eats up free time in the short run, but it’s a worse-before-better dynamic, so there’s a payoff in the long run. Fortunately liquidating stuff has a learning curve – it gets easier.
  • Make underutilized and broken stuff salient, by keeping lists and eliminating concealing storage.
  • Change your shopping policy to forbid acquisition of new stuff until existing stuff has been dealt with.
  • Buy higher quality than you think you’ll need.
  • Learn low-stuff skills.
  • Require steady state stuff: no shopping for new things until something old goes to make way for it.
  • Do things, even when you don’t have the perfect gear.
  • Explicitly prioritize stuff acquisition.
  • Tax yourself, or at least mentally double the price of any proposed acquisition, to account for all the side effects that you’ll discover later.
  • Get relatives to give $ to your favorite nonprofit rather than giving you something you won’t use.

There are also some policies that address the social dimensions of stuff:

  • Underdress and underequip. Occasionally this results in your own discomfort, but reverses the social arms race.
  • Don’t reward other peoples’ shopping by drooling over their stuff. Pity them.
  • Use and promote shared stuff, like parks.

This system has a lot of positive feedback, so once you get the loops running the right way, improvement really takes off.

Distilling Free-Form Natural Laws from Experimental Data

An interesting paper of that name came out in Science two years ago. There’s a neat video:

For centuries, scientists have attempted to identify and document analytical laws that underlie physical phenomena in nature. Despite the prevalence of computing power, the process of finding natural laws and their corresponding equations has resisted automation. A key challenge to finding analytic relations automatically is defining algorithmically what makes a correlation in observed data important and insightful. We propose a principle for the identification of nontriviality. We demonstrated this approach by automatically searching motion-tracking data captured from various physical systems, ranging from simple harmonic oscillators to chaotic double-pendula. Without any prior knowledge about physics, kinematics, or geometry, the algorithm discovered Hamiltonians, Lagrangians, and other laws of geometric and momentum conservation. The discovery rate accelerated as laws found for simpler systems were used to bootstrap explanations for more complex systems, gradually uncovering the “alphabet” used to describe those systems.

The Eureqa application used to mine data for relationships has been released at the authors’ Cornell site.

I think an interesting question is, will this approach work on noisy or ill-defined systems like climate or organizations? My guess is that it will have the same limitations as human-produced science. There’s a reason that a lot of physical laws were nailed down centuries ago, but our models of biological, economic and social phenomena are still pretty limited.

Drunker than intended and overinvested

Erling Moxnes on the dangers of forecasting without structural insight and the generic structure behind getting too drunk and underestimating delays when investing in a market, with the common outcome of  instability.

More on drinking dynamics here, implemented as a game on Forio (haven’t tried it yet – curious about your experience if you do).

Androids rule the earth

Android activations are apparently growing 4.4% per week, on a basis of around 100 million sales per year.

By the rule of 72 for exponential growth, that means sales are doubling every 16 weeks, or about three times per year.

If sales are growing exponentially, the installed base is also growing exponentially (because the integral of e^x is e^x). Half of the accumulated sales occur in the most recent doubling (because the series sum 1+2+4+8+…+n = 2*n-1), so the integrated unit sales are roughly one doubling (16 weeks) ahead of the interval sales.

Extrapolating, there’s an Android for everyone on the planet in two years (6 doublings, or a factor of 64 increase).

Extrapolating a little further, sales equal the mass of the planet by about 2030 (ln(10^25/10^8)/ln(2)/3 = 19 years).

Limits? What limits?

Cool videos of dynamics

I just discovered the Harvard Natural Sciences Lecture Demonstrations – a catalog of ways to learn and play with science. It’s all fun, but a few of the videos provide nice demonstrations of dynamic phenomena.

Here’s a pretty array of pendulums of different lengths and therefore different natural frequencies:

This is a nice demonstration of how structure (length) causes behavior (period of oscillation). You can also see a variety of interesting behavior patterns, like beats, as the oscillations move in and out of phase with one another.

Synchronized metronomes:

These metronomes move in and out of sync as they’re coupled and uncoupled. This is interesting because it’s a fundamentally nonlinear process. Sync provides a nice account of such things, and there’s a nifty interactive coupled pendulum demo here.

Mousetrap fission:

This is a physical analog of an infection model or the Bass diffusion model. It illustrates shifting loop dominance – initially, positive feedback dominates due to the chain reaction of balls tripping new traps, ejecting more balls. After a while, negative feedback takes over as the number of live traps is depleted, and the reaction slows.