Don't watch it grow — bet against it.
A reinforcing loop is the simplest engine in the world: the more there is, the faster it grows. Interest earns interest. A rumour reaches people who repeat it. A population of rabbits makes more rabbits who make more rabbits. The output of the loop is the input to the next turn — so growth doesn't add, it multiplies. And multiplication does something to our event-brain that it cannot handle: it crawls along looking like almost nothing for a long time, and then, with no change in the rule, it runs away.
So don't take my word for how far it runs — commit a guess and let it correct you. Pick a growth per step that feels trivial, a few percent. You're shown only the first stretch — the part that looks flat. Now drag your guess for where it ends up after all the steps, lock it in, and press reveal. Almost no one guesses high enough; the gap between your number and the truth is the whole lesson, and it is far bigger than your intuition will let you believe.
This is why “it's only a few percent” is the most underestimated sentence in finance, biology, and epidemics alike. A constant growth rate has no flat part — it only has a part you can't see yet. The honest reading of a reinforcing loop is not the level it is at today. It is the rate, and the patience to multiply it out.
“A reinforcing feedback loop enhances whatever direction of change is imposed on it.”
Donella Meadows · Thinking in Systems
And that last phrase — whatever direction — hides the whole danger of the loop. The same machine that compounds you up will, the instant the sign flips, compound you down just as fast. That's Game Two.
The same loop is a virtuous circle — and a death spiral.
A reinforcing loop has no preferred direction. It amplifies whatever it is handed. Hand it a nudge up and it builds a virtuous circle: savings that fund more savings, a reputation that wins the work that grows the reputation, a healthy bank whose deposits prove its strength and attract more deposits. Hand the very same loop a nudge down and it builds a vicious circle from identical wiring: debt whose interest deepens the debt, a panic that empties the bank that justifies the panic. One structure, two destinies — the sign of the first nudge decides which one you live.
Below is a single reinforcing loop balanced on a knife-edge. Drag the nudge away from zero and the loop seizes it and runs — up if you pushed up, down if you pushed down, by exactly the same amount. The two curves are not two systems. They are the same system handed opposite first steps. Watch how a push too small to matter at the start decides everything by the end.
This symmetry is why reinforcing loops are the most dangerous part of any system to ignore. A virtuous circle you are enjoying and a death spiral you are dreading are the same object seen from two sides, and the distance between them can be a single bad quarter, a single rumour, a single missed payment. The move is not heroics at the bottom of the spiral — by then the loop is doing the work against you. The move is to defend the sign while it is still cheap to defend.
“Systems with similar feedback structures produce similar dynamic behaviors, even when the outward appearance of these systems is completely dissimilar.”
Donella Meadows · Thinking in Systems
Put two reinforcing loops in the same arena, competing for the same scarce thing, and the amplification turns on the gap between them. The leader's lead becomes the reason it leads. That's Game Three.
A small early lead becomes a permanent one.
Now two reinforcing loops share one pie. Each grows in proportion to the share it already holds — the bigger player wins more of the new growth, which makes it bigger, which wins it more of the next. Meadows called this archetype success to the successful: customers flock to the busier restaurant, links point at the page that already has links, capital flows to the fund that already won. The advantage is not a fixed quality of the leader. It is manufactured by the loop, fresh, every single round.
Start the two almost level — a lead so small it looks like noise — and turn up the feedback strength: how sharply being ahead converts into winning the next round. Below a certain sharpness the two coexist. Above it, the tiny initial gap is seized, widened, locked, and the laggard is driven toward zero no matter how good it was. This is the engine under network effects, monopolies, runaway inequality, and first-mover advantage — and the unsettling part is how little a head start it needs.
Success to the successful is why the fair-sounding rule “let the best one win” so often produces an outcome no one would call fair: the loop rewards being ahead, not being good, and being ahead came first. If you want a different ending you cannot wait and referee the finish — by then the gap is locked. You change the wiring early: give the laggard a floor, cap the leader's compounding, or reset the round before the loop runs away. The structure decided the winner long before the contest looked decided.
“The more an actor wins, the more resources that actor is able to bring to bear to win in the future. This is a reinforcing loop, rapidly dividing players into winners who keep winning and losers who keep losing.”
Donella Meadows · Thinking in Systems
Defend the sign, not the level.
A loop that feeds on itself: invisible, then unstoppable; the same wiring for the boom and the bust; and, set against a rival, a maker of monopolies out of margins. None of it lives in how big the stock is today — it lives in the rate and the direction the loop is amplifying. That is the discipline a reinforcing loop demands. Don't wait to act until the level is alarming; by then the loop has the momentum. Catch the rate while it is small, protect the sign while it is cheap, and never assume a virtuous circle can't run backward — it is built to. The engine of your best outcome and your worst is the same machine; the only question it ever asks is which way you pointed it. Part of the same family as the rest — a system that can ruin you is one whose loops you didn't read.
“Remember, always, that everything you know, and everything everyone knows, is only a model. Get your model out there where it can be viewed. Invite others to challenge your assumptions and add their own.”
Donella Meadows · Thinking in Systems
Donella H. Meadows (1941–2001) was an environmental scientist and lead author of The Limits to Growth (1972); her primer Thinking in Systems (2008, ed. Diana Wright) is the source of every quotation here. The reinforcing loop, the virtuous/vicious symmetry, and the “success to the successful” archetype are hers. The three games are deliberately simple toy models — a stock multiplying by a constant factor, a reinforcing loop around an unstable equilibrium handed mirror-image nudges, and a two-player share that grows in proportion to a power of the lead. They are directionally honest illustrations of the archetype, not a forecast of any particular market, population, or bank. The staging, the prose, and the thread are my own. This is the reinforcing-loop chapter of a systems-thinking thread that begins with It's Just a Loop and continues through It's Just a Thermostat (balancing loops), It's Just Late (delays), and It's Just Noise (randomness in loops). See all concepts →