How to Research a Prediction Market Before You Trade

Prediction market prices move on information. If you trade without researching, you're betting that everyone else on the other side did less homework than you. This guide is the process we use. It works for any market on Kalshi or Polymarket, whether it's a Fed decision, an election or a box office number.

The short version:

1. Read the resolution rules
2. Turn the price into a probability
3. Find the base rate
4. Gather the news that matters
5. Map the catalysts
6. Make your own estimate and compare it to the price
7. Check liquidity before you size
8. Write it down

1. Read the resolution rules first

Every market has a rules section. It spells out exactly what has to happen for the market to resolve "Yes", which source decides it, and by when. Read it before anything else, because the market pays out on the rules, not on what you think the question means.

Look for:

• The resolution source. Is it an official agency release, a specific news outlet, a data provider? A market about inflation might resolve on the first published CPI number, not the revised one.
• The deadline and time zone. "By December 31" can mean 11:59 p.m. Eastern or midnight UTC. That gap matters for late-breaking events.
• Edge cases. What happens if the event is postponed, a candidate withdraws, or the data release is delayed? Good rules say so. If they don't, that ambiguity is a risk you're pricing in.
• How disputes work. Kalshi resolves markets itself under its published rules. Polymarket uses an outside oracle process in which proposed outcomes can be challenged. Know which system you're in before the outcome is contested.

2. Turn the price into a probability

A prediction market contract pays $1 if the outcome happens and $0 if it doesn't. So the price is the market's implied probability: a "Yes" contract trading at 62¢ means the market collectively puts the odds at roughly 62%.

Two adjustments make that number more honest:

• Use the midpoint, not the last trade. If the best bid is 60¢ and the best ask is 64¢, the market is saying "about 62%". The last trade could have hit either side.
• Account for costs. Fees and the spread are what you pay to get in and out. If you buy at 64¢ and your fair value is 66%, your edge is thinner than it looks once you count those costs.

3. Find the base rate

Before you read a single headline, ask how often this kind of thing happens. This is the step most traders skip, and it's the one that protects you from getting carried away by a story.

• How often does the Fed move rates at a meeting when futures markets are pricing a hold?
• How often does a bill that has passed one chamber become law within the same session?
• How often does the favorite win this kind of award or match?

The base rate is your starting estimate. News should move you away from it, and you should be able to say why and by how much.

4. Gather the news that matters

Now look at what's happening. The goal isn't to read everything. It's to find the information that changes the probability.

Prioritize primary sources. An agency's own release, a court docket, a company filing or a candidate's official statement beats a summary of it. Secondary coverage is useful for context, but check the original when a claim is decisive.

Check recency. Prediction markets react within minutes. A story that is two days old is almost certainly already in the price. What you want is information the market hasn't fully absorbed yet, or context that changes how a known fact should be read.

Separate signal from noise. For each item, ask: if this is true, how much should the probability change? A poll with a small sample, an anonymous single-source report, or a viral post usually deserves a small update, if any.

Look for the other side. Actively search for the best case against your view. If you can't find one, you probably haven't looked hard enough, or the market already agrees with you.

5. Map the catalysts

Write down the scheduled events between now and the resolution date that could move the price:

• Data releases (jobs reports, CPI, earnings)
• Votes, hearings, court dates and filing deadlines
• Debates, games, award ceremonies
• The resolution date itself

Catalysts tell you when risk arrives. A market that resolves in three weeks but has its key catalyst tomorrow behaves like a one-day trade. Also note what an unscheduled shock could look like, such as a resignation, a ruling or an injury, and whether the rules handle it.

6. Make your own estimate and compare it to the price

Start from the base rate, adjust for the evidence you found, and write down a range, not a single number. For example: "I think this is 55–65% likely."

Then compare:

When your view differs sharply from the market, take that seriously in both directions. Sometimes you've found an edge. Often the market knows something you don't. Before trading, ask: what would someone need to believe for the current price to be right?

7. Check liquidity before you size

A thin market can have a price that looks wrong simply because nobody has traded it recently. Before you size a position:

• Look at the order book depth near the current price.
• Check recent volume. Can you exit if you change your mind?
• Size so that being wrong is survivable. Even a well-researched 65% call loses about one time in three.

8. Write it down

Record your estimate, the price you traded at, the two or three reasons behind your view, and what would change your mind. This is the single best habit for getting better over time, because it lets you tell skill from luck once markets resolve.

A simple log is enough.

A quick checklist

Before you place any prediction market trade:

• I've read the rules and know the resolution source and deadline
• I know the implied probability at the midpoint, after fees
• I have a base rate
• I've checked recent primary sources and the strongest counter-argument
• I know the upcoming catalysts
• My estimate range sits outside the market price by more than costs
• I can enter and exit at a reasonable size
• I've written down my reasoning