Concepts: edge & settlement

Before composing a strategy it helps to be precise about three things: what a contract is, what an edge means, and why an edge does not last forever. None of this is a claim about profit — it is the vocabulary a strategy is built from.

Yes/no contracts

Parachute trades Kalshi crypto contracts on BTC and ETH. Each contract is a yes/no question with a fixed settlement time — for example, whether the price is above a strike at the top of the hour. At settlement the contract resolves to one side:

  • If the answer is yes, a YES contract pays 100¢ and a NO contract pays 0¢.
  • If the answer is no, it is the other way around.

Because the payout is always 0¢ or 100¢, the market price of a contract (in cents) is the market's implied probability of that outcome. A contract trading at 63¢ is the market saying "about a 63% chance." That single fact is what makes an edge measurable.

The hourly crypto series (the KXBTCD / KXETHD families) settle every hour, not once a day, despite the "daily"-sounding names — so a strategy on them is making many short-horizon decisions through the trading day.

Edge and expected value

A Model prices each contract to produce a fair value — the model's own estimate of the settle probability, in cents. The edge is the gap between that fair value and what the market is charging:

Edge = model fair value − market price

If the model thinks a contract is worth 70¢ and you can buy it at 64¢, the modeled edge is 6¢. Edge is the raw signal a strategy acts on. It is modeled edge, not realized profit: it is only as good as the model, and a model can be wrong on any given contract.

Expected value (EV) is the same idea carried to the payout. A position's expected value is the probability-weighted average of its outcomes minus its costs. This is why edge alone is never the whole story — fees and the bid-ask spread are real costs that eat into any modeled edge, and a strategy has to clear them before a trade is worth taking. The Net-Edge Filter exists precisely to check whether an edge survives those costs.

A subtle but important point: a larger modeled edge is not automatically better. An implausibly large edge is often a sign that the model is miscalibrated on that contract or that the quote is stale — not a bigger opportunity. The Overpriced-Edge Guard is the "too-good-to-be-true" check that refuses those.

How settlement works

A few settlement mechanics shape every strategy:

  • Contracts settle on a schedule, not on demand. A position held to expiry resolves at the settlement time; you don't choose when. A position can also be closed early on the market before settlement, but the headline P&L on your dashboard is settled results only — see Money safety and the dashboard docs for how early closes are shown separately.
  • P&L is net of fees. Realized profit and loss already accounts for the trading fees on both sides of a trade.
  • Kalshi does not net offsetting positions. If a strategy holds a position and a hedge on the opposite side of the same market, both legs settle independently — you are not automatically flattened. This is what makes the Tail Hedge and Hedge Ladder real, separately-settling positions rather than bookkeeping offsets, and it is why a hedge has its own cost.

Why an edge fades when it is crowded

An edge is a finite resource. It exists because a price is briefly out of line with what a model believes, and the market has not yet corrected it. As more capital trades the same dislocation, two things happen: the mispricing gets corrected faster, and the available size at a good price runs out. The edge compresses.

This is the intuition behind capacity: a strategy that works at a small size may not work at a large one, and a strategy that many people run at once competes with itself. Parachute's design anticipates per-strategy capacity limits for exactly this reason. The practical takeaways for a builder:

  • Do not assume a strategy scales indefinitely. More size is not always more profit.
  • A crowded, well-known idea has thinner edge than a selective one.
  • The durable edge in this venue tends to come from selectivity and risk control — trading only the best candidates and bounding the losses — rather than from a broad "always-on" bet. The archetypes page makes this concrete.

None of this guarantees any particular result. It is the reason the rest of this track spends so much time on filters (what to skip) and exits (how to bound a loss) rather than only on the entry signal.