Two Wrongs That Make a Right

You're offered two coin-flipping games. You play Game A a thousand times: you lose money. You play Game B a thousand times: you also lose money. Now someone says, "alternate between them." You do. And you win. This isn't a magic trick. It's a theorem.

The Worst Advice That Actually Works

Imagine I walk up to you at a party and say: "I've got two investments. They both lose money on average. But if you switch back and forth between them, you'll make money." You would, reasonably, call me either a fraud or an idiot. Possibly both.

And yet, in 1996, a Spanish physicist named Juan Parrondo proved that exactly this can happen — not as a conjecture, not as a heuristic, but as a mathematical certainty.1 Two games, each provably unfavorable to the player, can be combined into a game that is provably favorable. The losers combine into a winner. The whole is better than its parts — not by a little, but by a lot.

If this bothers you, good. It should. It means some of your most basic intuitions about winning and losing are wrong.

The Two Games

Let's make this concrete. You start with some amount of money — say $100 — and you play coin-flipping games where you either gain $1 or lose $1.

Game A is simple. You flip a biased coin that lands heads 49.5% of the time and tails 50.5% of the time. Heads, you win a dollar. Tails, you lose a dollar. This is a losing game. Not by much — it's almost fair — but over thousands of flips, you'll steadily bleed money. The house edge is tiny but relentless, like a casino's.

Game B is trickier. It uses two coins, and which one you flip depends on how much money you have right now:

Game B seems like it should be a winner — after all, two-thirds of the time you're flipping a coin that wins 74.5% of its flips. But here's the rub: the game's internal dynamics conspire against you. When you win a couple of rounds with the good coin, your capital drifts toward a multiple of 3, which triggers the terrible coin. The bad coin knocks you back down, and the cycle repeats. It's like climbing a sand dune where every two steps up causes a small avalanche that sends you one step back — and then some.2

Play Game A alone? You lose. Play Game B alone? You lose. The expected value of each game, played repeatedly in isolation, is negative. This is provable, checkable, and not in dispute.

Now alternate: play A, then B, then A, then B. Or if you prefer, switch between them randomly, choosing each with 50% probability on every round.

You win.

Why It Works: The Ratchet and the Reset

The key to Parrondo's paradox is that Game B's outcome depends on your current state — specifically, whether your capital is divisible by 3. Game A, by contrast, doesn't care about your state at all. It just docks you a tiny fraction every round.

And that tiny, seemingly harmful nudge from Game A is exactly what saves you.

Here's the mechanism. When you play Game B alone, you naturally drift toward multiples of 3 (because the good coin keeps pushing you there), and then you get clobbered by the bad coin. It's a trap. The game creates its own headwinds.

But when you interleave Game A, something remarkable happens. Game A's small random perturbation knocks your capital off the multiples of 3. It breaks the pattern. It rescues you from the bad coin and sends you back to the good coin. Game A is like a controlled stumble that keeps you from walking into a pit.

The physicist's analogy is to a Brownian ratchet — a hypothetical device from thermodynamics where random thermal fluctuations, combined with an asymmetric ratchet mechanism, can produce directed motion.3 In physics, this seems to violate the second law of thermodynamics (it doesn't — you need an energy input somewhere). In Parrondo's games, the "ratchet" is the modular arithmetic of Game B, and the "fluctuations" are the random noise from Game A.

Parrondo himself drew the connection explicitly. His paradox was inspired by the work of Richard Feynman on ratchet-and-pawl devices, and by the flashing Brownian ratchet model of Ajdari and Prost.4 The mathematics of the coin-flipping games is, in a deep sense, a discrete version of the physics. Same structure, different costume.

Try It Yourself

Don't take my word for it. Below is a simulator that lets you play all three strategies — Game A alone, Game B alone, and the alternating combination — and watch the capital trajectories diverge in real time. Run it a few times. Watch the green line climb while the red and blue lines sink. Then try to convince yourself it isn't happening.

It's Not Just a Parlor Trick

When Parrondo first presented his paradox, some people dismissed it as a mathematical curiosity — a clever construction with no real-world relevance. They were wrong. It turns out the paradox shows up everywhere, once you know what to look for.

Evolution and Population Genetics

Consider a population of organisms in a fluctuating environment. In Environment 1, a certain genotype is slightly disadvantageous. In Environment 2, the same genotype is also slightly disadvantageous. But if the environment flips back and forth between these two states — which real environments constantly do — the genotype can actually increase in frequency.5

This isn't hypothetical. Reed and colleagues showed in 2007 that Parrondo-like dynamics can explain how genetic diversity is maintained in fluctuating environments — a longstanding puzzle in evolutionary biology. The environment itself acts as the "switching mechanism" between the two losing games.

Finance and Portfolio Strategy

Here's one that should make every investor sit up straight. Suppose you have two assets. Each one, individually, has a negative expected return (maybe they're both slightly losing bets after transaction costs). But rebalancing between them at fixed intervals can produce a positive expected return.6

This is related to the well-known "volatility pumping" or "rebalancing bonus" effect in portfolio theory. The mechanism is the same as Parrondo's paradox: switching between two state-dependent processes disrupts their individual unfavorable equilibria. It's not free money — you need the right kind of correlation structure and mean-reversion — but it's a genuine mathematical phenomenon that portfolio managers exploit.

As the economist Andrew Lo pointed out, this connects to a broader insight: in nonlinear, state-dependent systems, the order in which things happen matters at least as much as what happens.7

Game Theory and Evolutionary Strategies

In evolutionary game theory, Parrondo effects appear in competing populations where no single strategy dominates. A population that switches between two individually losing strategies can outcompete populations that stubbornly stick with one. This has implications for understanding biodiversity: it suggests that "suboptimal" strategies persist in nature not despite being suboptimal, but because switching between suboptimal strategies can be optimal.8

The Fine Print (Because There's Always Fine Print)

Before you run off and start alternating between two bad stocks, a few caveats.

First, Parrondo's paradox requires state-dependent dynamics. At least one of the games must have outcomes that depend on your current position. If both games are simple independent coin flips with negative expected value, no amount of alternation will save you. You can verify this yourself: two biased-against-you coins, flipped in any order, will still lose. The paradox only works because Game B's behavior changes based on your capital modulo 3.

Second, the paradox is not a violation of any law of probability. It feels like you're getting something for nothing, but you're not. The alternation changes the distribution of states you visit, and since Game B pays off differently in different states, you end up spending more time in the states where Game B is generous. You're not creating value from nothing — you're navigating a landscape more cleverly.

Third, in practice, you need to know the structure of the games to exploit the paradox. In Parrondo's original setup, you need to know that Game B depends on capital mod 3. In finance, you need to know the correlation structure of your assets. In biology, evolution "knows" this through natural selection — the organisms that happen to switch at the right frequency survive, and those that don't, don't.

What Parrondo Teaches Us

Ellenberg's How Not to Be Wrong is, at its core, a book about the dangers of linear thinking — about how our instinct to extrapolate, average, and simplify leads us astray. Parrondo's paradox is a perfect case study.

We are wired to think that if something is bad, more of it is bad, and mixing it with another bad thing is worse. This is what Ellenberg calls the "linearity fallacy" — the assumption that effects add up simply. They don't, not when the system has state-dependence, feedback, and nonlinearity.

Parrondo's paradox says: stop evaluating strategies in isolation. A strategy that looks terrible on its own might be exactly what you need as a complement to another strategy. The value of a tool depends on what other tools are in the box.

This has a humbling implication for decision-making. We like to evaluate options one at a time — this job vs. that job, this investment vs. that investment, this policy vs. that policy. But Parrondo tells us that the interaction between options can be more important than the options themselves. A diverse portfolio of mediocre strategies can beat a concentrated bet on the "best" strategy. A career that zigzags between two seemingly unproductive paths might outperform a straight line.

There's a reason that evolution — the greatest optimizer we know — doesn't pick one strategy and stick with it. It maintains diversity. It switches. It mutates. It does exactly what Parrondo's paradox says you should do: it plays two losing games in alternation, and it wins.

The lesson isn't that two wrongs make a right. The lesson is that in a complex, state-dependent world, "wrong" and "right" are not properties of individual actions — they're properties of sequences of actions. And the sequence can have a logic that the individual steps do not.

Which, if you think about it, is a pretty good description of how life works.