Winner-takes-most

Most markets in the economy can support a crowd of competitors. There are dozens of viable airlines, hundreds of coffee brands, thousands of restaurants. Exchanges are different. They drift, almost mechanically, toward a single dominant venue — one order book that ends up holding most of the trading in its category, with everyone else fighting over the remainder. It isn't an accident of branding or a quirk of any one company. It's the math of liquidity, and once you see the loop it's hard to unsee.

This is the structural fact behind the business of prediction markets, and it's the reason the crypto-exchange playbook from the last piece is worth so much. Winning a license is the entry ticket. Winning the liquidity is the prize — and the prize, unlike the ticket, compounds.

The liquidity flywheel

Start with the thing a trader actually cares about: can I get in and out at a fair price, right now, without moving the market against myself? That property is liquidity, and it has a self-reinforcing quality that ordinary products don't.¹ We pulled the mechanism apart in the market-maker problem; here's the loop in four steps.

More traders on a venue means more resting orders — a deeper book. A deeper book means a smaller gap between the best bid and the best ask: a tighter spread, and less slippage when you trade size. Tighter spreads and better execution mean a better price for everyone who shows up — which is exactly what pulls in the next trader. Round and round. Each new participant makes the venue marginally better for the next one, and that improvement is free to the operator: the users build the moat by showing up.

You can write the punchline as a one-line dependency. Volume on a venue is roughly self-feeding — the flow you have attracts the next increment of flow:

When the next period's activity is an increasing function of this period's, you have positive feedback, and positive feedback doesn't settle into a tidy split. It tips. Two venues that start nearly even will, after enough rounds, separate — the slightly deeper one offers a slightly better price, wins the next trade, gets deeper still. Small early leads turn into large permanent ones.

Why it's most, not all

The honest version of this is winner-take-most, not winner-take-all — and the distinction is the whole reason a challenger ever has a shot. The feedback loop is strong, but it isn't a law of nature. It can be contested, and it regularly is, by three things: a meaningfully better experience that's worth the worse price; maker incentives — paying liquidity providers to seed the book until it's deep enough to stand on its own;² and the fragmentation that's native to this category, where the same outcome can trade on an on-chain venue and an off-chain one at the same time. Liquidity that's split across rails is liquidity the leader doesn't fully own.

So the realistic shape isn't a monopoly. It's a dominant venue that holds the lion's share, a credible second place, and a long tail — concentrated, but live. The leader's lead is durable, not permanent. That's the right amount of humility to carry into the business case, and it's also why the next part of this story is the cleanest evidence I know that the concentration is real.

The Korea experiment

Economists rarely get a controlled experiment. Markets usually move for ten reasons at once. But occasionally a policy change acts like one, holding most things fixed and flipping a single variable — a natural experiment. Korea ran one for us, by accident, in 2017.

Through that year, crypto trading in Korea exploded, and a large share of it ran through offshore exchanges outside any domestic rulebook. The government moved to bring the activity onshore: it required real-name, bank-verified accounts, leaned hard on the unregulated venues, and effectively pushed trading toward licensed domestic operators.³ The variable that flipped was simply where it's legal and easy to trade. And when the dust settled, one onshore venue — Upbit, the licensed domestic leader, launched that October — ended up holding the overwhelming majority of national volume, on the order of 80%+ of Korean trading.³

That's the experiment. Hold the asset constant, hold the appetite constant, change only the legal surface — and the market doesn't spread itself across the field. It collapses onto the first licensed venue with real liquidity, and that venue runs away. The crowd wasn't told to converge. It converged because every individual trader, choosing the deepest legal book, was choosing the same one — and each choice made that book deeper still.

The Schelling point

There's a name for the thing the crowd lands on. Thomas Schelling pointed out that when people need to coordinate without communicating, they gravitate to the obvious focal choice — the option that's salient simply because everyone expects everyone else to pick it.⁴ An exchange's liquidity is the purest Schelling point I can think of. I want to trade where the others are, because that's where the price is fair and the fill is instant; they want the same; so we all want the venue we all expect to win. The expectation is self-fulfilling, and the prize for being the venue everyone expects is that you become the venue everyone uses.

Layer that on top of a license and the moat gets deep. Onshore, the field of legal venues is small to begin with. The first one to combine a license with real liquidity becomes the focal point inside the rules — the obvious place to trade — and dislodging a focal point is far harder than beating a normal competitor, because you're not fighting a product, you're fighting an expectation that's busy confirming itself. This is the same conclusion the broader series keeps arriving at: the mechanism is copyable, the license is granted one market at a time, and the liquidity, once it pools, stays pooled.⁵

What this means for the category

Put the pieces together and the strategy writes itself. Be early, be licensed, be liquid — in that order, because each enables the next. Early gets you to the license while the field is still forming; the license clears the legal field; liquidity, seeded deliberately at first, then takes over and compounds on its own. The flywheel does the rest. That's the bet I work on at Seeker, market by market — and to be exact about status, that license is the goal, the thing you earn by showing up first and building the surveillance and settlement in from the first trade. It is not a thing already in hand.

The honest footnote is the one from earlier: this is winner-take-most. Dominance here is strong, durable, and — under enough pressure from better UX, smarter maker incentives, or liquidity that lives across several rails at once — still contestable. A focal point can be moved; it's just expensive, and the meter runs against the challenger every round. Most of the volume in these markets today is sports,⁶ and that's worth saying plainly, because the network effect doesn't care what people are trading — it concentrates the venue all the same. The interesting question isn't whether one venue per market will dominate. The experiment already answered that. The question is who gets there first.