Expected Value (EV) for Event Contract Traders

What Is Expected Value in Event Contract Trading?

Expected value (EV) is the average result of a trade if you made it many times under the same conditions, weighted by probability. It tells you whether a position has an edge. That edge shows up across a consistent strategy applied to many markets. On any single contract, you either win or lose.

A share pays all or nothing, even when the market has more than two outcomes. A flight market has three outcomes: on time, delayed past the threshold, or cancelled. A Delayed share settles at 1.00 USDT if Delayed is the outcome that happens, and at 0.00 USDT in both other cases. This changes your maths. The probability you need to beat is Delayed against everything else, cancellation included, and on time is only part of that. The price you pay is the chance the market currently gives your outcome. Your job is to decide whether your own estimate differs from that number, and whether you have good reason to trust it.

The core formula:

EV = (P × W) − ((1 − P) × L)

Where:

P = your estimated probability of the outcome happening
W = your gain per contract if you’re right (settlement value minus entry cost, in USDT)
L = your loss per contract if you’re wrong (your entry cost, in USDT)

Run every position on Gaduin through this equation before you open it. Positive EV means you have an edge over the long run. Negative EV means the current price doesn’t favour you. The Wikipedia entry on expected value has the formal derivation if you want it.

The EV Formula: Breaking Down Each Component

Each part of the formula matches something you can see in the market.

Say Delayed (at the market’s threshold) on a specific flight trades at 0.40 USDT a share. It settles at 1.00 USDT if the flight arrives later than the threshold, and 0.00 USDT if it arrives on time or is cancelled.

Entry cost / L: 0.40 USDT, the amount at risk per contract.
W: 1.00 − 0.40 = 0.60 USDT, your gain per contract if the outcome goes your way.
P: 0.60, your estimate from the data that the flight will be delayed.
q = 1 − P = 0.40, your estimated probability of anything else happening: on time, or cancelled.

Plugging into the formula:

EV = (0.60 × 0.60) − (0.40 × 0.40) = 0.36 − 0.16 = +0.20 USDT per contract

An EV of +0.20 USDT means the trade is worth taking, as long as your P comes from data you can check and isn’t a hunch. If EV is negative, the price doesn’t favour you. Skip it and look at other markets.

All numbers here are examples. Your real EV depends on how good your probability estimates are.

Reading Contract Prices: Implied Probability

On Gaduin, a price is a probability. A share pays 1.00 USDT if its outcome happens, so the price is the chance the market gives that outcome. This is called implied probability.

For a share that settles at 1.00 USDT, there’s nothing to convert:

Implied P = contract price in USDT

A share at 0.35 USDT means the market gives that outcome a 35% chance. At 0.70 USDT, it’s 70%. You can read the probability straight off the price.

Your edge is the difference between the implied probability and your own estimate:

Edge = Your P − Implied P

If the market implies 35% and your research supports 50%, your edge is +15 percentage points and EV is positive. If your estimate matches the market’s, EV is about zero and you have no reason to trade.

Take the market price seriously. It’s where the price settled after everyone who looked at this flight traded on what they knew, and some of them looked harder than you. Use it as your starting point, and look for specific reasons you can check for why your estimate differs, before the price moves to meet it. For more, see how prices move.

Where Your Edge Comes From: Private Information vs. Market Price

Edge comes from public information the price doesn’t fully reflect yet, or from a better model applied to the same data.

For flight delay event contracts, the main sources of edge are:

Where the aircraft is. If the plane due to operate a departure is already late on its previous flight, the price may not reflect that yet. Live flight tracking is public.
Airport congestion and slot limits. Airport slot controls and hub congestion cause delays that repeat on predictable days and at predictable hours. Traders who model ATFM (air traffic flow management) restrictions can find markets where the price is too low for congestion delays.
Historical delay rates by route, airline and hub. The U.S. Bureau of Transportation Statistics (BTS On-Time Statistics) publishes detailed monthly history for US domestic flights, and Europe’s air navigation performance statistics break down delay by cause. Both give you a figure you can check for P, instead of one you remember.
Knock-on delays. One early delay spreads through the fleet as the day goes on. If you understand how cascade flight delays compound through a schedule, you can find markets that haven’t priced in those later effects yet.

Edge doesn’t last. As information spreads, prices adjust. Traders who act quickly on solid research do better than those who react late.

EV Calculation Example: A Flight Delay Contract

This example uses made-up numbers to show the EV steps on one contract.

Contract: “Delayed” (arriving more than 15 minutes late) on an LHR→FRA afternoon departure.

Research: On this route, about 38% of flights arrive more than 15 minutes late. Today the inbound aircraft was 90 minutes late on its previous flight, which makes a delay clearly more likely than usual.

Probability estimate: P = 0.55 (a 55% chance of Delayed, weighing the inbound delay against the historical rate).

Market price: 0.40 USDT a share, a 40% chance. Your edge = 55% − 40% = +15 percentage points.

EV calculation:

Entry cost = L = 0.40 USDT
W = 1.00 − 0.40 = 0.60 USDT

EV = (0.55 × 0.60) − (0.45 × 0.40) = 0.33 − 0.18 = +0.15 USDT per contract

An EV of +0.15 USDT with a 15-point edge is enough to justify a disciplined entry. The Gaduin Frankfurt Connection case study shows how layovers and aircraft rotations add up to delay risk.

All figures above are examples. Real estimates depend on the flight data available when you trade.

Position Sizing: Kelly Criterion for Event Contracts

Finding a positive EV trade is only the first step. How much you put in decides whether your edge grows your balance over time or exposes you to losses you didn’t need to take.

The Kelly Criterion gives the fraction of your capital that maximises long-run growth:

f* = (b × p − q) / b

Where:

b = gain-to-loss ratio: W / L (gain per contract if right / loss per contract if wrong)
p = your probability estimate
q = 1 − p

Using the example above:

b = 0.60 / 0.40 = 1.5
p = 0.55, q = 0.45

f* = (1.5 × 0.55 − 0.45) / 1.5 = (0.825 − 0.45) / 1.5 = 0.375 / 1.5 = 0.25

Full Kelly says to put 25% of your allocated capital into this contract. In practice that’s too aggressive, because small errors in P make results swing a lot. Many traders use Half Kelly (12.5%) or Quarter Kelly (6.25%) to allow for mistakes in their estimates.

Sizing rules for event contracts:

Apply Kelly separately to each position that isn’t related to the others.
For related contracts, such as several departures from the same congested hub in the same time window, treat the total as one Kelly allocation.
Never put more into one contract than you can afford to lose entirely.

The Wikipedia article on the Kelly Criterion has the full derivation. For applying it across many positions, see how prices move.

Common EV Mistakes Event Contract Traders Make

Most traders understand EV. Their mistakes come from not applying it every time.

Survivorship bias when estimating P. People remember winning trades more clearly than losing ones, so their edge seems bigger over time than it is. Base P on complete historical data and your own trade records, and don’t rely on the wins you remember.

Ignoring what size costs. Gaduin charges no trading commission, so no fee comes out of W. Your own order size does: a large order fills at a rising average price, so use the average you expect to pay as L, and not the first price you saw. Moving USDT in and out also costs money. A deposit pays the standard Ethereum network fee, which hurts most on small amounts, and a withdrawal carries a Gaduin commission. Recalculate EV with those costs included. A trade with positive EV at the screen price can be negative at your fill.

Treating related positions as independent. Contracts on several flights from the same hub at peak times share the same cause of delay. Sizing each one with Kelly as if it were independent makes your total exposure look safer than it is.

Estimating the wrong outcome. Each market defines its own outcomes and threshold. Check that your P matches the exact outcome in the contract you’re trading before you calculate EV. This matters most on multi-leg trips, where intermodal flight-to-train delay risk makes estimates harder across different kinds of transport.

Changing P after you enter because you want to be right. Once a position is open, don’t update P just because the price moved. Price moves after you enter may come from other people’s trades and may say nothing new about the flight. Doing the research before you enter is the fix.

Applying EV Analysis on Gaduin

Gaduin is a prediction market on transport delays. Prices move as people trade, and settlement is automatic under terms fixed when the market opened. For a flight, that’s the flight’s recorded arrival compared with its scheduled arrival. Each share of the outcome that happens pays 1.00 USDT.

The EV steps for each trade:

Find the contract. Browse by route or departure time. Note the current price and the exact outcome definition (threshold, what’s measured, settlement rules).
Read the implied probability. The price in USDT is the implied P.
Make your own estimate of P. Use BTS historical data for US routes and the published European delay-cause statistics for European ones, plus same-day NOTAMs and public flight tracking for the inbound aircraft. Remember that q includes cancellation as well as an on-time arrival.
Calculate EV and edge. Apply the formula. Only open the position if EV is above zero and your edge clears a minimum, for example 5 percentage points, to allow for errors in your model.
Size with Kelly. Use Half or Quarter Kelly to set the size against your allocated capital.
Record every trade. Write down the price you paid, your P, the market’s implied chance, the EV you calculated and the outcome that settled. Keeping records is the only way to check whether your model has the edge you think it has.

Situations worth researching include tight connections at congested hubs, late-day departures on aircraft with known inbound delays, and routes where tarmac delay rules come into play, which makes delays more likely than the base rate.

Risks, Limitations, and Responsible Trading

EV is a long-run average. Positive EV doesn’t guarantee you’ll win any single contract. Results vary, and even a strategy with a real edge will lose on some positions. That’s normal and doesn’t by itself mean your model is wrong.

Model risk. Your estimates depend on your data and your assumptions. Systematic errors, even small ones, can make your real EV negative when your calculated EV is positive. No estimate can be proven right before settlement.

Other traders. Other people may get public data faster or model it better. An edge based on information everyone has and analyses shrinks quickly as prices adjust.

Thin markets. When few people are trading, the price may be out of date and tell you little, and your own order moves it further as it fills. Both make the price less reliable as a starting point.

USDT means real money at risk. Losses are in USDT and grow with position size. Spreading positions reduces swings but doesn’t remove the risk of loss.

Every position on Gaduin can lose money. Size your positions with that in mind.

Event contracts on Gaduin can lose you money. Past performance doesn’t predict future results. This article is for education only and is not financial or investment advice. Gaduin event contracts are not available to U.S. persons.