Haun Ventures Leads River Markets Seed Round

On the morning of April 26, 1973, a small group of traders filed into a converted smoking lounge on the fourth floor of the Chicago Board of Trade building. They could now buy exchange-listed calls on just sixteen stocks, and by the closing bell that day only 911 options contracts had been traded. By the end of the first month just under 35,000 contracts had exchanged hands, and by the end of the first year that figure was 1.5 million monthly contracts. The Chicago Board Options Exchange had created a new kind of machine which took a single stock and splintered it into dozens, sometimes hundreds of different instruments - across strike prices, expiration dates, and directions - each of which could be used to express a more specific view of the company’s future than simple equities allowed. Importantly however, this precision came at a cost: the fragmentation of liquidity. Instead of just a single orderbook for a given asset, suddenly there were many. The complexity of executing trades successfully with options meant that it was no longer sufficient for traders to be experts in predicting the future of the underlying asset they were trading – they suddenly needed to be experts at executing trades in order to monetize their insights.

Fast forward to today, we now witness a similar pattern playing out across the prediction markets ecosystem. We believe event contracts have many of the same issues that options contracts first did, alongside a number of new ones that necessitate a new kind of execution layer for trading on prediction markets. We’re excited to be backing River Markets as they build a prime brokerage focused on this developing market.

How deep is the book?

For a given exchange to compete in the market, it must offer users a liquid orderbook so that users can arrive at the venue and execute trades in the size they’re looking for. Today it’s cheap and easy to create markets around just about anything you can think of; the real issue is attracting enough liquidity to those markets that they can be useful in surfacing a coincidence of wants between counterparties. 

Take for example the simple BTC/USD pair that’s traded on all large crypto exchanges. Each pair on each exchange has its own network effects - a snowball of resting orders and constantly moving trades which makes it even easier to onboard marginal liquidity and new traders executing larger and larger orders. Each trading pair also benefits from the exchange’s network effects, when traders bring their winnings from trading one pair to other pairs on that venue (or use cross margin to execute their trades more efficiently). Liquidity begets liquidity. The snowball rolls downhill. 

When options were first introduced, liquidity fragmentation was a real problem, but eventually it was resolved with scale. Even though options all had their own expiration dates, the traders themselves (and the market viewpoints they held) didn’t suddenly evaporate when an option expired. Traders would easily roll over their position into the next expiry date, which offered more or less an extension of the same viewpoint they were already trying to trade on. This resulted in a similar snowball effect of deepening liquidity, with small drop-offs when a given contract expires. When we look at prediction markets, we see a very different pattern emerging. 

Event contracts are discrete, with a natural inability to roll over a trade into a new version of the same contract. The vast majority of event contracts that are traded today are not recurring or continuous - they are isolated and unique. Perhaps most importantly, the alpha that a trader has about a particular event is rarely applicable to other events being traded soon thereafter. The result is that prediction markets deliver a fundamentally different market structure, whereby liquidity needs to be bootstrapped from scratch on a continuous basis, vs accruing via the same network effects that deepening pairs benefit from. While some network effects do exist because users return to venues they trust, each contract still has to start with effectively $0 of resting liquidity.

This issue is exacerbated by the fact that a given event might have dozens of different frames of reference being introduced, or different variables being isolated for resolution. Take for example a given election event - this would produce a myriad of different markets, such as: did a particular candidate win? Did a candidate receive more than 50% of the vote? Will their party control the legislature? Will they win a particular state? Will the margin exceed a specific threshold?

These liquidity bootstrapping headwinds are the very reason why the prediction markets ecosystem has seen such large shifts in volume market share across venues over time. While traditional asset markets trend towards consolidation (from the aforementioned liquidity network effects), prediction markets have revealed a persistent opportunity for new entrants to supplant the existing high-volume venues that are live today. Both new venues and old ones must start from similar footing in attracting liquidity to every single new event contract they list. While it’s unclear what the terminal state of the market structure will be, standing DCM/DCO applications paint a clear picture: this market is likely to remain fragmented for a number of years as traders decide which platform tradeoffs they prefer and venue operators compete via liquidity incentives. 

Can you help me source some size in this?

When options first introduced liquidity fragmentation, they gave rise to an entire industry dedicated to solving the problem of how to execute a given trade efficiently. One could determine an option was mispriced, but struggle to buy a meaningful amount without moving the market. 

To solve this problem, a whole host of new financial infrastructure was built. Institutional traders increasingly turned to brokers who could scour venues, find counterparties, and work large orders over time. Eventually they began to route orders electronically, and what was originally a human task was quickly taken over by software. Smart order routers would search across venues to find the deepest liquidity. Execution algorithms would then slice large orders into smaller ones, and dynamically adjust how aggressively they traded. Prime brokers gave funds a single interface through which positions, financing, clearing, and execution could be managed across a growing number of trades. 

Now that we are witnessing increased institutional demand to trade on prediction markets in size, the time has come for a new layer of the market to be built, which prioritizes the needs of clients trading large volumes across multiple venues concurrently.

When we first met Oscar and Antonin from River Markets, we were immediately impressed by their depth of knowledge of both traditional prime brokerage execution (from their time spent as quants at Blackrock and Valkyrie), as well as the specific issues that traders faced today when trying to monetize their alpha. They had earned these insights the hard way: through boots-on-the-ground iteration over several years, developing and testing the algorithms to power their own trades. As we spoke to more and more traders in the space, a clear picture emerged: these guys aren’t just extremely capable quants; they’re the kind of humans you want by your side to successfully trade in these constantly evolving markets. 

We are delighted to lead the Seed round of River Markets, and we’re excited to witness them ushering in a new era of institutional prediction markets trading.