Skip to main content
The Mark Price is the internal reference price used for all critical trading operations on Ascend. It determines your unrealized PnL, margin requirements, and liquidation thresholds. While the Index Price reflects external oracle data, the Mark Price applies smoothing and stabilization to protect traders from short-term noise and sudden spikes. This separation is fundamental to operating leveraged prediction perpetuals safely.

​
Why Not Use the Index Price Directly?

Prediction markets are sensitive to:
  • Thin liquidity causing price jumps
  • Sudden probability updates from breaking news
  • Oracle update granularity and timing gaps
Example scenario: Breaking news drops. A whale buys $500k on Polymarket, spiking the price from 55% to 72% for 2 minutes. The news turns out to be a rumor. Price returns to 57%. If Ascend used the Index Price directly:
  • Traders with SHORT positions could get liquidated at 72%
  • Their positions would be closed at a loss
  • Two minutes later, they would have been fine
Using the raw Index Price for liquidations and PnL would expose traders to:
  • Forced liquidations from short-lived probability spikes
  • PnL volatility unrelated to actual market conditions
  • Manipulation through brief oracle distortion
The Mark Price ensures that risk is assessed on a stable signal, not on momentary noise.

​
How the Mark Price Is Constructed

The Mark Price is derived from the Index Price through three layers:
1

Smoothing (EMA Filter)

An Exponential Moving Average dampens short-term fluctuations in the Index Price.
Mark_t = Mark_t-1 + λ × (Index_t − Mark_t-1)
Where:
  • Mark_t-1 is the previous Mark Price
  • Index_t is the current Index Price
  • λ is the smoothing factor (between 0 and 1)
What λ means in practice:
λ ValueBehaviorUse Case
0.1Very slow, very stableLow-volatility event markets
0.5BalancedMost markets
0.9Fast, less smoothingHigh-frequency markets
Lower λ means more stability and slower reaction. Higher λ means faster response to Index changes. Each market defines its own λ based on expected volatility.
2

Price Clamping

The Mark Price is constrained to stay within a defined range of the Index Price.
If |Mark − Index| > Clamp Limit:
    Mark = Index ± Clamp Limit
This prevents the Mark Price from drifting too far from reality while still providing stability. Clamp limits are market-specific and defined in percentage terms (e.g. ±1%).Why clamping matters:Without clamping, a very low λ could cause the Mark Price to lag far behind during a sustained move. A trader could enter at 50%, watch the Index move to 70%, but see their Mark Price stuck at 55%. The clamp ensures Mark stays within a reasonable distance of truth.
3

Precision Rounding

The final Mark Price is rounded according to market precision rules (e.g. 2 decimal places).This ensures consistent PnL calculation, deterministic settlement, and alignment between the UI and trading engine.

​
Example: No Clamp Triggered

Assume:
  • Index Price = 62.0%
  • Previous Mark Price = 60.5%
  • λ = 0.5
  • Clamp Limit = ±1.0%
Step 1: Apply EMA 
Mark_raw = 60.5 + 0.5 × (62.0 − 60.5)
         = 60.5 + 0.75
         = 61.25
Step 2: Check Clamp 
|61.25 − 62.0| = 0.75
0.75 < 1.0 → No clamp needed
Final Mark Price = 61.25% The Mark Price moved toward the Index but did not fully reach it due to smoothing.

​
Example: Clamp Triggered

Now assume a sudden Index spike:
  • Index Price = 72.0% (sudden jump)
  • Previous Mark Price = 60.5%
  • λ = 0.5
  • Clamp Limit = ±1.0%
Step 1: Apply EMA 
Mark_raw = 60.5 + 0.5 × (72.0 − 60.5)
         = 60.5 + 5.75
         = 66.25
Step 2: Check Clamp 
|66.25 − 72.0| = 5.75
5.75 > 1.0 → Clamp triggered
The Mark Price would be 5.75% away from the Index, which exceeds the 1% clamp limit. Apply Clamp: 
Mark = Index − Clamp Limit
     = 72.0 − 1.0
     = 71.0
Final Mark Price = 71.0% The clamp pulled the Mark Price closer to the Index, ensuring it does not lag too far behind during a large move.

​
How the Mark Price Is Used

The Mark Price drives all internal trading operations:

Unrealized PnL

Your position value is calculated using the Mark Price, not the Index Price. This ensures PnL reflects sustained probability movement rather than transient spikes.

Margin Requirements

Initial and maintenance margin are evaluated against the Mark Price. This prevents sudden oracle jumps from instantly changing your margin status.

Liquidation Checks

Liquidation thresholds reference the Mark Price. Temporary probability spikes cannot trigger liquidation if the Mark Price remains stable.

Funding Calculations

Funding rates are computed from the difference between Mark Price and Index Price, incentivizing convergence without destabilizing positions.

​
Mark Price Behavior in Different Conditions

The Mark Price tracks the Index Price closely with slight smoothing. Traders experience stable PnL that reflects genuine market movement.Example: Index moves from 60% to 62% over 5 minutes. Mark Price follows, reaching 61.8% in the same period.
The EMA filter slows the Mark Price response. If the spike is temporary, the Mark Price may never fully reach the spike level before the Index returns to normal.Example: Index spikes from 60% to 75% for 30 seconds, then returns to 62%. Mark Price rises to maybe 65%, then settles back to 62%. No liquidations triggered by the spike.
The Mark Price converges toward the new Index level over time. The clamp ensures it cannot lag too far behind during sustained moves.Example: Index moves from 60% to 80% over 10 minutes and stays there. Mark Price follows, staying within 1% of Index throughout the move.
If oracle data becomes stale or erratic, the Mark Price maintains stability based on recent valid data. The clamp prevents extreme divergence.Example: One oracle goes offline. Index Price recalculates with remaining sources. Mark Price continues smoothly without sudden jumps.

​
Parameters by Market

Each market defines its own Mark Price parameters:
ParameterDescriptionExample
λ (smoothing factor)How quickly Mark responds to Index changes0.3 to 0.7
Clamp LimitMaximum allowed divergence from Index±0.5% to ±2%
PrecisionDecimal places for final Mark Price2 decimals
Update FrequencyHow often Mark Price recalculatesPer block or per second
These parameters are set at market creation and visible to all traders. Higher volatility markets may use higher λ and wider clamps. Stable markets may use lower λ and tighter clamps.

​
Why This Matters

The Mark Price system allows Ascend to:
  • Anchor markets to external truth through the Index Price
  • Protect traders from noise and manipulation through smoothing
  • Maintain stable, continuous trading even during volatile probability updates
  • Apply perpetual-style risk management to probability markets
Without this separation, leveraged prediction markets would be vulnerable to the same issues that plague thin liquidity venues: flash crashes, manipulation, and unfair liquidations. The Mark Price makes leveraged prediction perpetuals viable.

Next: Margin Model

Learn how margin requirements and liquidation logic protect positions