How It Works

Noise markets trade continuously. Their prices come from the order book. Funding creates incentives around that market price, while isolated margin keeps risk contained to each market position.

​

The two prices

• Market price is a quantity-weighted price calculated from the best bid and best ask. It appears on market charts and in live trading references.
• Relevance is a five-minute exponential moving average of market price, shown with the ° mark. It appears across trend, discovery, position, funding, and risk surfaces.

Inside the system, market price is the mid-price and relevance is the mark price. These docs use only the product terms from here on.

​

Market price

Market price uses the best bid, best ask, and the quantity available at each price. Each price is weighted by the quantity resting on the opposite side of the book. Let B be the best bid, A the best ask, Q_B the quantity at the best bid, and Q_A the quantity at the best ask: $$P_{\text{market}} = \frac{Q_A B + Q_B A}{Q_B + Q_A}$$ This is commonly called a microprice. A larger bid quantity moves market price toward the ask, reflecting stronger buying pressure. A larger ask quantity moves it toward the bid. When both quantities are equal, market price is the plain midpoint. Because the calculation needs both sides, Noise cannot produce a current market price when either the best bid or best ask is unavailable.

​

Relevance

Relevance smooths market price with a 300-second exponential moving average: $$R_t = \operatorname{EMA}_{300\text{s}}\!\left(P_{\text{market},t}\right)$$ The five-minute window reduces the effect of brief order-book changes while preserving sustained moves. Relevance remains entirely order-book-derived. A value of 581.50 is displayed as 581.50°. The EMA form and window are market-structure parameters. Noise can tune them as liquidity and risk requirements evolve.

​

Funding

Funding periodically transfers value between long and short positions. It responds to market conditions around relevance and helps limit persistent imbalances between the two sides of a market. A positive funding rate means longs pay shorts. A negative rate means shorts pay longs. Noise is moving the funding calculation from a top-of-book microprice input toward an impact price. Impact price reflects the order-book depth needed to execute a representative trade, making the funding signal less dependent on a single best bid or ask. The exact price input, component weights, caps, and settlement interval are market-structure parameters. They can change as liquidity develops. The rate and next settlement time shown in the trading interface are the current values for that market.

​

Liquidation and position risk

Liquidation is a forced risk action taken by the matching engine when a position no longer has enough equity for its margin requirements. Noise uses isolated margin, so the action applies to that market position without drawing margin from another position. The matching engine tracks this risk through the margin ratio. It compares margin already in use, plus margin reserved for resting orders, with the position’s total equity. Let M_used be used margin, M_orders be open-order margin, and E_total be total equity: $$\rho = \frac{M_{\text{used}} + M_{\text{orders}}}{E_{\text{total}}}$$ Total equity accounts for deposited collateral, realized and unrealized profit or loss, trading costs, funding, and withdrawals. A rising margin ratio means the position has less equity available to absorb further losses.

• Below 80%: safe. No risk action.
• 80% to below 90%: margin call. Noise warns the trader to add margin or reduce exposure. The position has not entered forced liquidation.
• 90% to below 100%: preventative liquidation. The matching engine places a small forced order to reduce the position and its risk.
• 100% or higher: regular liquidation. The matching engine forcibly reduces the position at available order-book prices. If the book cannot absorb the liquidation order, Noise can use auto-deleveraging to complete the risk reduction.

Liquidation can realize losses, and its execution price can differ from relevance because the order executes against available liquidity. Traders can lower their margin ratio before forced action by adding collateral where supported, reducing or closing the position, or cancelling resting orders that reserve margin.