Expected Value Is the Only Number That Matters in Prediction Markets

The market is pricing a Fed rate cut at 28 cents. You think the real probability is 38%. That gap is the entire game. Everything else — the news flow, the talking heads, the chart patterns — is noise until you run it through expected value.

Here is the math. If you pay 28 cents for a contract that you believe is worth 38 cents, your edge is 35.7%. That means for every dollar you put into this trade, you expect to get $1.357 back over the long run. That is a monster edge in any market. Bond desks would mortgage the building for that kind of spread.

If you are not calculating this number before you trade, you are not investing. You are guessing.

What Expected Value Actually Means

Expected value is the probability-weighted average of all possible outcomes. For a binary prediction market contract, the math is clean:

EV = (your probability × payout) − (1 − your probability) × cost

A YES contract on Kalshi pays $1.00 if it resolves YES and $0 if it resolves NO. If you buy it at 28 cents and your true probability is 38%:

EV = (0.38 × $1.00) − (0.62 × $0.28) = $0.380 − $0.174 = +$0.206 per share

That is 73.6 cents of expected profit for every dollar risked. In a market this size, that is not a close call. That is a trade.

The Three Signals: BUY, SELL, SKIP

Every market price sits in one of three zones relative to your probability estimate.

BUY: Your probability is significantly above the market price. The market is underpricing the event. You buy YES.

SELL: Your probability is significantly below the market price. The market is overpricing the event. You buy NO (or short YES where available).

SKIP: The gap is too small to justify the risk. The market is roughly right. You walk away.

The threshold that matters is your minimum edge. Mine is 5 percentage points. A 28-cent market where I think the true probability is 30%? Skip. A 28-cent market where I think the true probability is 38%? That is a BUY I size aggressively.

Anything under 5pp of edge and the bid-ask spread plus execution slippage wipes out most of your theoretical advantage. Discipline here is what separates people who compound from people who chase.

Where Your Probability Estimate Comes From

This is where most traders fall apart. They calculate EV correctly but garbage-in-garbage-out their probability estimate. Your estimate is only as good as the work behind it.

Four inputs I use:

Base rates. What is the historical frequency of this type of event? The Fed has held rates in 18% of meetings where CPI was above 3.5%. If the market is pricing a cut at 35%, that is a 17-point gap from the base rate. That demands a specific narrative for why this time is different.

Recent evidence. Bayesian updating. Start with the base rate as your prior. Add the evidence you have. Revise. The CPI print came in hot? That prior needs to move down. NFP missed badly? It moves up. The revision should be proportional to how much the evidence actually discriminates.

Market structure. Who is on the other side? On a Kalshi market about a regulatory outcome, if the regulated industry is known to hedge its exposure on prediction markets, their buying could push prices above fair value. That is a short opportunity, not confirmation that the market is right.

Resolution criteria. Read the contract. Carefully. I have seen markets where the resolution date does not match the event date, where "GDP growth" uses a different data release than traders assumed, where the contract resolves on a specific revision that changes the math. Every misread resolution criterion is money left on the table — or worse, money you gave away.

Edge % Is Not Position Size

Here is the trap. You find a 35% edge and want to put everything you have into it. Do not.

Edge percentage tells you whether to trade. It does not tell you how much to trade. That is Kelly Criterion — a separate calculation that accounts for your edge and your bankroll simultaneously. A 35% edge on a 50-cent contract calls for a specific fraction of your bankroll, not 100%.

Use the Kelly Criterion calculator to get the right size once you have your edge.

Overbetting a positive-EV trade is as destructive as betting negative EV. The Kelly framework exists precisely because edge and optimal size are two separate problems.

The Discipline Habit

Before every Kalshi trade, I run this sequence:

1. What is my probability estimate, and what specifically supports it?

2. What is the market price?

3. What is my edge in percentage points?

4. Is the edge above my minimum threshold?

5. What is my Kelly-adjusted position size?

Steps 1 through 3 take ninety seconds with the EV calculator. Steps 4 and 5 add another minute. Every trade I make that skips this sequence is a trade I regret.

The market does not care about your conviction. It cares about whether you are right more than the price implies. Expected value is how you measure that.

Run your next trade through the EV Calculator before you click.