On the surface, newborn babies seem rather simple. Feed ’em, burp ’em, change ’em, cuddle ’em, rock ’em, sing to ’em, put ’em down for a nap. The more we learn about infants, however, the more we realize how capable they are of responding to nuances. Luckily, most of the beneficial things to do with tiny ones come naturally. A little experimenting goes a long way and a graduate degree in child psychology is neither necessary nor probably of any help.
Compare, however, trying to get it right with a two-year-old. They’re ambulatory, strong, strong-minded, and not much more reason-able than an infant. You’ll do a whole lot more experimenting to get considerably less success. Still, coping with a two-year-old is nothing compared to dealing with the most complex system on Earth, a creature more unfathomable than black holes, quarks, and the continued interest in Donald Trump’s personal life: the teenager.
Any confidence you gained because some interaction you had with your 14-year-old turned out well will haunt you within a half-hour; their moods change, your moods change, the stars have shifted. Is the situation hopeless? Absolutely, we’re discussing teenagers. Still, you might be able to increase your chances. We’re going to create a simulated teenager so you’ll know how to make yours happy, which I’m sure is at the top of your to–do list.
Start your Mac and load the spreadsheet. Name the file SimTeen. Set up columns for interactions between you and Johnny and assign values to them. For example, asking Johnny to take out the garbage gets a negative five. Now set up the necessary spreadsheet formulas.
You won’t need a psychoanalyst to understand this teen. A quick glance at the spreadsheet below will tell you all there is to know about Johnny. The key formula adds the numbers in the checked columns to arrive at the total and displays sullen if the score falls below 17 and happy if it falls at 17 or above (see figure 1).
 
Now suppose we make it a little more complex. In our present sample, asking Johnny to take out the garbage and reminding him to pick up his towel adds minus three to minus five. Let’s say that two negative interactions strung together, as they might be in real life, multiplies the negatives rather than adds them. (For our purpose, multiplying two negative numbers will give us a negative result.) Also, let’s design it so that if a positive interaction is four times or more the size of the negative one, the negative one is simply thrown out of the equation. Telling Johnny to take out the garbage before he leaves with your little red sports car and credit card is not likely to turn his mood foul. And last, since in real life a similar interaction can be positive or negative, depending on the circumstances, we should do the same in our sim. In the sample below, bestowing a compliment on Johnny’s appearance is a negative, as it may well be regarded coming from a parent who consistently lobbies for the “wholesome” look. Nevertheless, any praise is likely to be well received when it breaks a string of negative interactions.
Let’s take stock. Our SimTeen is still sufficiently simple that it’s easily understood with a glance at the formulas. Suppose, however, I add hundreds of interactions and link them in every which way. While each link might be made with simple cause-effect logic, the resulting web could create a mind-boggling complexity. At this point our teenager—and we’re now talking about a teenager and the very big world in which he lives—can no longer be understood from the inside any more than you could do an IQ test via a brain scan. Our simulation is going to be sufficiently complex that we’d better find a different avenue to learn about Johnny. Great! That’s why we created SimTeen: to explore through interaction what makes Johnny tick. First, however, we must finish our sim.
Since Johnny is now a complex guy, we expect a wider range of expression than “happy” or “sullen,” and a wider range of behaviors than taking out the garbage and the like. As with any modern sim, we’ll use graphics as a means of communication between the formulas and the users of SimTeen. Our Johnny will be programmed to display body language and facial expressions to match his moods.
SimTeen is now complete. Johnny has—God help us—the traits of a teenager, so we can learn about teens and increase our ability to cope with them.
I created SimTeen to trim some of the mystery from simulations. While my sample is facetious, and commercial sims are far more complex and coded in sophisticated programming languages, at heart, most of them are still glorified spreadsheets—formulas and their graphical face.
At their best, simulations are both immensely entertaining and instructive. They are even more instructive (but perhaps less entertaining) if you are aware of the assumptions behind the formulas. By assigning values, SimTeen makes assumptions about the impact of specific interactions. Commercial simulations such as SimCity 2000 do likewise. SimCity 2000 (SC2K) the upgrade of the first successful, commercial computer simulation, SimCity, has added many features to make it more “true-to-life.” For instance, under Ordinances, you can choose to spend money on items such as an anti-drug education program. In SC2K the anti-drug program reduces crime. Sounds good. Meanwhile, there’s little hard evidence that their real life counterparts have accomplished more than the production of slogans.
I don’t mean to single out SC2K; any simulation that assumes logical human behavior in its calculations is crossing its fingers behind its back. That life is more complex than forty–dollar software isn’t going to surprise anyone. Still, the kiddie versions aside, these sims reflect the authentic world of business and economics far more than say, the issues on which our politicians run for office. The view of the universe that fits within a politician’s sound bite doesn’t translate into worldly ideas. Business, which includes economic and social factors, has begun to reflect the complexity of the natural world, a domain—if you look beyond a few years—that defies “management.” We should be moving towards an experimental rather than a management model.
Such a world suggests a different sort of preparation from what aspiring managers sought in the past. The idea that all the significant variables in commerce can be known and understood, any better, that is, than we understand teenagers, must be abandoned. Managing becomes a matter of working with models that you know are incomplete and inaccurate. So what can you do? Throw something at a system and observe what happens. Evaluate. Abandon failure, proceed with success. And wait for success to become failure. Start again. Successful football coaches, research biologists, and parents operate this way. While you may never be able to claim that you understand the system, you should be able to improve at predicting near-term outcomes. That’s as good as it gets.
To cram this all into a nutshell, canned expertise is out, analysis and response is in. Kids (and grown-ups) who get some seasoning on computer models are likely to be wiser when it counts.
