The Future Was Never Binary

There is something slightly peculiar about the way we have learned to ask questions about the future.

Will it happen or won’t it? Will Bitcoin cross the level? Will a candidate win? Will the team qualify? The structure is familiar enough that we rarely notice it anymore. A question is placed in front of us, two possible answers are attached to it, and a market is asked to decide which one deserves to be taken seriously.

It is a very efficient little machine.

It is also, perhaps, a rather poor description of the future.

The future almost never arrives in two pieces. A company does not simply succeed or fail; it misses, recovers, changes direction, gets acquired, gets lucky. An election produces margins and surprises and strange combinations of events that seemed unrelated until afterwards. In crypto, the interesting question is often not whether something happens but which of several competing things happens first, or which one becomes important enough to change everybody else’s expectations.

We know this instinctively. Prediction markets have simply been slower to reflect it.

Edmond Halley understood something about prediction that is easy to forget when the word is used today. The famous comet was not a prediction in the modern, casual sense of the word. Halley was working from observations, from records and from the uncomfortable fact that a useful prediction may outlive the person making it. He died in 1742. The comet returned in 1758, more or less on schedule.

There is something slightly strange about building a system around a future event when you know you may not be there to see the answer.

Perhaps that is part of what prediction is really for. Not certainty, but a way of making uncertainty legible to people who come later.

The modern prediction market inherited some of this ambition, although it eventually put it into a remarkably narrow container: yes or no.

That made sense. Binary contracts are easy to understand and relatively easy to settle. They are a good invention. The problem is that a useful invention has a tendency, once it becomes familiar, to start looking like a law of nature.

It isn’t.

Consider the question of which blockchain will lead DEX volume in a particular quarter. The usual instinct is to split it into a series of propositions: will Ethereum lead, will Solana lead, will BNB Chain lead? But this isn’t really how anyone is thinking about the question. The question is comparative. There are several futures sitting on the table, and the interesting thing is the distance between them.

A market with several possible outcomes can show that.

More importantly, it makes the person creating the market do something that is easy to overlook. They have to decide what the possible futures are.

That sounds administrative until you think about it for a moment. It isn’t. Choosing the outcomes is part of the forecast. You are saying, in effect, these are the branches of the future worth paying attention to; these others probably aren’t. The market begins before anyone places a trade.

This is one of the ideas behind Halley Market, an open prediction market network on BNB Smart Chain. A Halley market can contain between two and ten mutually exclusive outcomes, with the outcomes traded through an on-chain order book. Their prices move as participants trade and can be read as the crowd’s changing estimate of the probability of each result.

That sounds like a technical distinction. It is one, but only in the same way that a road map is a technical distinction when what you really care about is where you can go.

The more outcomes a market can contain, the less the market has to pretend that uncertainty is a coin toss.

And this is where the Halley name becomes slightly more appropriate than it first appears. Halley was interested in recurring things, things that looked chaotic if you watched them for a moment but became less mysterious when you had enough observations and enough patience to put them beside one another, and perhaps that is what a prediction market is trying to do in miniature, although the market has the peculiar disadvantage that it has to ask its questions before the evidence exists, which is also why the people who are most certain are not necessarily the people who are most useful.

That is probably too much to ask of a comet.

It may not be too much to ask of a market.

The catalogue problem

There is another assumption built into the traditional prediction-market model which is less visible, because it sits outside the contract itself.

Someone has to decide what gets listed.

This is perfectly reasonable. A prediction market needs resolution criteria. It needs some confidence that a question has an answer and that the answer can be established. A platform also has to think about regulation, reputation, liquidity and whether anyone will bother trading a particular market.

But once a company becomes the gatekeeper for all of those decisions, the catalogue inevitably starts to reflect the company’s own boundaries.

Some questions are too obscure. Others are difficult to resolve. Some attract attention the operator would rather not have. Some aren’t commercially interesting enough to justify the work.

One decision at a time, the list gets shaped.

This is not necessarily a failure of the people doing the shaping. It is simply what happens when an institution sits in the middle of a market.

An open network makes a different bet.

If someone thinks a question deserves to exist, they can create it. If other people care, they trade it. If nobody cares, it doesn’t magically become important because somebody at the company put it on a homepage.

That changes the potential catalogue quite dramatically.

A niche community can have its own market. A creator can put a prediction to work. A question that would never pass through the machinery of a large platform can exist because somebody, somewhere, actually wants the answer.

This is familiar elsewhere on the internet. Television networks decided what viewers could watch. YouTube didn’t need to know in advance which videos deserved an audience. Retailers decided what deserved shelf space; Shopify let the merchant make the decision.

Prediction markets may be arriving at a similar question rather later than those industries did.

The interesting part is that the person creating the market may also be the person who already has the audience.

The audience is already arguing

Think about a creator who has spent years talking about one subject.

A crypto analyst makes a call. A sports commentator predicts a result. A newsletter writer says that one protocol will overtake another. The audience replies. Half of them agree. Half don’t. Someone writes a long response explaining why everyone else is an idiot.

Then the feed moves on.

There is an odd economic structure hiding inside this very ordinary internet ritual. The argument is engagement, the engagement is valuable, the platform captures most of that value, and the person who started the argument gets another increment to the follower count.

An open prediction market gives the argument a different destination.

The creator defines the question, chooses the outcomes, specifies how the answer will be established and points the audience toward it. The people who disagree can do more than disagree in the comments. They can take the other side.

The market becomes a more permanent version of the argument.

Halley’s economics are designed around this relationship. A creator can open a market without putting up liquidity and receives 80 percent of the resolution fee when it settles.

The important idea isn’t really the 80 percent. That is the incentive mechanism. The more interesting part is who is being rewarded: the person who supplied the question and brought the people interested in answering it.

Of course, this doesn’t make the cold-start problem disappear. It would be nice if technology worked that way, but it doesn’t. A market without traders is still a market without traders. Established venues have a substantial advantage because they have spent years concentrating liquidity and attention.

Halley Market’s answer is simply to move part of that responsibility outward.

If the person creating the market already has an audience, the demand does not have to be found after the market exists. It can arrive with it.

Whether that is enough to build deep markets at scale is an empirical question. But it is at least an interesting answer to a problem that every open financial network has to face.

The categories were always ours

There is a particularly strange kind of market that becomes possible once you stop insisting that every question belong to a category.

What happens first this Saturday: Real Madrid scores, Bitcoin crosses $100,000, or Taylor Swift posts on Instagram?

It sounds frivolous. It is also a perfectly coherent event market. There is a time limit, there are three mutually exclusive outcomes and, assuming the criteria are written properly, there will eventually be a public answer.

What makes it strange is not the question.

It is that our platforms have taught us that these things belong in separate rooms.

Sports. Crypto. Culture.

People don’t necessarily live that way.

The person watching Real Madrid may also be watching Bitcoin. The person watching Bitcoin may also be following Taylor Swift. The internet has spent twenty years making these collisions increasingly ordinary, while the institutions built to organise information have generally continued to sort everything into boxes.

Halley’s mash-up markets are an almost comically simple response to this. The protocol can treat the outcomes as possibilities in the same market rather than asking which department they belong to. A football event, a crypto event and a cultural event can therefore become three outcomes of one question, with one order book and one settlement.

It is a small example, but it points toward something larger.

When the infrastructure stops caring what category an outcome belongs to, the people creating markets can start making stranger questions.

Some of those questions will be terrible.

Most things on the internet are.

A few will probably be surprisingly good.

That is the trade an open system makes.

What a prediction market is for

There is a tendency to judge prediction markets by the answer they eventually produce.

Who won? Was the market right? How accurate was it?

Those things matter, obviously. But the market is also interesting while nobody knows the answer.

That is when the prices are doing their most useful work. A five-outcome market can show not only which future is favoured but how far behind the alternatives are. It records the changes in belief as information arrives.

And, eventually, it records what people believed before they knew.

That record is valuable in a way that an ordinary prediction is not.

A post disappears into a feed. A market leaves behind a question, a set of alternatives, a series of prices and, eventually, an outcome.

Halley Market is trying to build that record in a more open form: markets created by the people who want to ask the question, rather than only by the people who operate the venue. Its contracts are designed to run on-chain, with outcome shares held in users’ wallets and settlement governed by the protocol’s rules.

Whether this becomes the dominant model is, of course, another prediction.

And perhaps that is the appropriate place to leave it.

Halley the astronomer spent years looking at old observations to understand something that had not yet happened. He died before the experiment was complete. The sky, rather inconveniently, did not care.

The comet came back anyway.

Prediction has always been a slightly strange human activity for that reason. We make claims about things that have not happened, using incomplete information, knowing that somebody else may eventually get to compare our confidence with reality.

The machinery around prediction is changing now.

Maybe the important change is not that we are becoming better at predicting.

Maybe we are finally getting better at describing what we mean when we say we don’t know.

And that, unlike a yes-or-no question, leaves rather more room for the future.