Positions and Payouts

A Signals v1 position is a claim on a contiguous price range for one market. If the settlement tick lands inside the range, the position pays its quantity; if it lands outside, the payout is zero. This simplicity is deliberate. It makes range positions easy to reason about and keeps pricing aligned with the CLMSR curve.

The position is not a probability; it is a contract on a discrete interval. The market’s continuous outcome is mapped to a tick, and the position pays based on that mapping. Positions are both precise and deterministic: there is a single rule that turns a real-world outcome into an on-chain payout.

What a position represents

A position bundles three ideas: the market, the range $[\text{lowerTick}, \text{upperTick})$ on the outcome grid, and the quantity paid if settlement lands inside that range. Positions are tokenized on-chain, but they can be read as transferable claims on a range payout.

That structure is what keeps continuous markets legible. The instrument is not a vague probability; it is a concrete payout rule tied to a specific interval on the grid.

Positions are composable. A wide belief can be expressed as a wide range. A more peaked belief can be expressed as multiple overlapping or adjacent ranges with different quantities. This is how continuous beliefs become structured exposure using standardized building blocks.

Positions are also grid-aligned. The bounds are expressed in ticks, not in raw observed units. Unit conversion happens at settlement time, when the oracle value is mapped to a settlement tick by the market's bounds and tick spacing.

Payout rule

Let $\tau$ be the settlement tick and $x$ be the quantity:

$$\text{Payout} = \begin{cases} x, & \tau \in [\text{lowerTick}, \text{upperTick}) \\ 0, & \text{otherwise} \end{cases}$$

Range positions are equivalent to a binary event over "in the range" versus "out of the range." That equivalence makes pricing and analysis tractable while still capturing continuous outcomes.

Because payout is binary over the range, the expected value of a position is directly linked to the probability mass the market assigns to that range. That connection is what allows price to be interpreted as a measure of belief.

If $p(R; q)$ is the current range price mass under the shared curve state $q$, then the instantaneous expected payout is $x \cdot p(R; q)$. The CLMSR cost function integrates that marginal price along the trade path as the state changes.

The same object also describes the immediate marginal price of adding a small amount of quantity to an existing position. Along a buy path, the marginal price of increasing quantity is the current range price mass. The executed average price is the path integral of that marginal price as the state moves.

Positions also have a natural "close value" before settlement. A sell is the reverse state update. If the current state is $q$, then selling quantity $x$ on range $R$ moves to $q' = q - x\mathbf{1}_R$, and the fee-free base proceeds are:

$$\text{baseProceeds}(x; q, R) = C(q) - C(q - x \mathbf{1}_R)$$

Convexity of $C$ implies that buying and then immediately selling the same range at the same state is not free: the state movement itself creates slippage in both directions, and fees are an additional overlay on top.

Smoother payoff shapes can be approximated by combining ranges with different quantities. For example, stacking narrower ranges in the center and wider ranges in the tails produces a bell-shaped exposure.

Half-open range convention

Signals uses half-open ranges $[\text{lowerTick}, \text{upperTick})$. The lower tick is included and the upper tick is excluded, which avoids double-counting adjacent ranges and keeps payouts unambiguous. It also ensures that every settlement tick maps to exactly one boundary interpretation.

This boundary rule is most visible when ranges are combined. Two adjacent ranges tile the grid without overlap, so their payouts never double-count a boundary tick. That keeps portfolio composition deterministic.

This convention also fixes the settlement boundary case. A settlement tick equal to an upper bound is treated as outside that range and inside the next range.

Example

If the outcome domain is $[50{,}000, 70{,}000)$ with tick spacing $100$, then the range $[60{,}000, 62{,}000)$ spans 20 ticks. With quantity $10$, a settlement price that maps to $60{,}900$ lands inside the range and pays out $10$. A settlement price that maps to $62{,}100$ lands outside and pays out $0$.

At the boundary, a settlement value exactly equal to $62{,}000$ maps to the tick at the boundary. Under the half-open rule, that tick is outside $[60{,}000, 62{,}000)$ and inside $[62{,}000, 64{,}000)$.

This example highlights the precision tradeoff: a narrow range is more specific but typically costs more per unit because it concentrates value in fewer ticks.

A narrow range can be read as a high-resolution claim. Missing the interval by one tick yields a zero payout. That is the economic cost of precision, and the curve prices that precision explicitly.

Composition shows up immediately. Consider two adjacent ranges with the same quantity $x$:

• $R_1 = [60{,}000, 62{,}000)$
• $R_2 = [62{,}000, 64{,}000)$

Exactly one of them pays out for any settlement tick in $[60{,}000, 64{,}000)$. Holding both positions produces a stepped payoff: $x$ on either side of the boundary, never $2x$ at the boundary.

Portfolio algebra

The payoff grammar is simple enough to do algebra directly at the level of ranges. Consider a set of adjacent ranges that tile a larger interval:

$$[a,b) = [a,t_1) \cup [t_1,t_2) \cup \dots \cup [t_{k-1},b)$$

Holding quantity $x$ on every tile produces a constant payoff $x$ on the entire interval $[a,b)$ (and still respects the half-open boundary rule). This is the discrete analogue of building a piecewise-constant function from indicator functions.

The same idea generalizes. A payout profile that assigns a different size to each tick interval can be expressed as a sum of disjoint ranges:

$$\sum_{i} x_i\,\mathbf{1}_{[\tau_i,\tau_{i+1})}$$

where $\tau_i$ are the grid ticks and $x_i$ are the desired payouts on each tick interval. Signals v1 does not implement arbitrary payout functions directly. It implements one range instrument, and the portfolio implements the function by holding multiple positions.

Positions are dynamic, not static

Positions can be increased, reduced, or closed before settlement. The cost of changing a position depends on the current pricing state, not on the path taken so far. That means a position is not a one-time bet; it is an exposure that can be managed over the trading day.

Execution strategy enters here. A trader can enter gradually to avoid large slippage, or rebalance as information arrives. Because cost is state-dependent, timing and sizing are part of the position’s economics, not afterthoughts.

Closing a position is symmetric: selling the range returns the current value of that claim, net of fees. Exposure can be adjusted during the trading day; it is not mechanically locked until settlement.

Edge cases

Range positions have simple surface rules, but edge cases come from grid alignment and settlement mapping:

• Zero-width ranges: $\text{lowerTick}=\text{upperTick}$ produces an empty interval and is not a meaningful position.
• Domain edges: the market's tick mapping clamps values below the lower bound into the first tick and values at or above the upper bound into the last tick.
• Boundary ticks: the half-open convention assigns boundary ticks to the right-hand range.

These conventions are part of the market contract that defines payout meaning before trading begins.

Related sections:

• Market Design
• CLMSR Pricing Curve
• Market Lifecycle