Rikiddo Scoring Rule

The Rikiddo scoring rule is a specialized automated market maker (AMM) mechanism for prediction markets, designed to improve on classic models like LMSR by making pricing:

  • more capital efficient
  • more adaptive to liquidity
  • more stable under large trades

It’s most closely associated with Gnosis / GnosisDAO research.

First: What is a “scoring rule” market maker?

A scoring rule is a mathematical function that:

rewards traders based on how accurate their probability estimates are.

In markets, this becomes:

  • You “move” probabilities
  • You pay a cost to do so
  • If you’re right → profit

Canonical baseline: LMSR

The classic model is the Logarithmic Market Scoring Rule (LMSR):

C(q) = b * log( sum_i exp(q_i / b) )

Where:

  • q_i = shares for outcome i
  • b = liquidity parameter

Problem with LMSR

  • Fixed liquidity parameter b
  • Either:
    • too expensive → no trading
    • too cheap → market maker risk explodes

Enter Rikiddo

Rikiddo is a dynamic scoring rule that:

adapts liquidity based on market conditions (volume, uncertainty, participation)

Core idea

Instead of fixed b, Rikiddo:

  • adjusts pricing sensitivity dynamically
  • reacts to:
    • trading volume
    • market entropy
    • participation levels

Intuition

  • Early market (low liquidity) → gentle price movement
  • Active market → tighter spreads, more confidence
  • Large trades → controlled slippage

Conceptual model

Traditional LMSR:

  • fixed curvature

Rikiddo:

  • adaptive curvature (breathes with the market)

What it improves

1. Capital efficiency

  • Requires less locked capital than LMSR

2. Price stability

  • Avoids extreme swings from small liquidity

3. Better early-stage markets

  • Bootstrap liquidity without huge subsidies

Comparison

FeatureLMSRRikiddo
Liquidity parameterFixedDynamic
Capital efficiencyLowHigher
Early market behaviorFragileStable
AdaptivityNoneHigh
ComplexitySimpleMore complex

Deep insight

Rikiddo is essentially:

A feedback-controlled AMM

Where:

  • the pricing function adapts based on system state

This is very different from:

  • AMMs → static curves
  • vAMMs → static formula + funding

Rikiddo introduces:

  • control theory into market design

Connection to systems thinking

1. Like a dynamic aggregation structure

  • LMSR = static
  • Rikiddo = adaptive weighting over time

2. Like a liquidity system with feedback

  • liquidity becomes a function of participation

3. Similar to sketches

  • not exact
  • optimized for:
    • responsiveness
    • bounded risk
    • efficiency

Trade-offs

Pros

  • More efficient
  • Better UX for traders
  • Lower capital requirements

Cons

  • Harder to reason about
  • Requires tuning
  • Less battle-tested than LMSR

Where it fits

Ideal for:

  • prediction markets
  • low-liquidity environments
  • long-tail event markets

Less ideal for:

  • high-frequency trading
  • deep, liquid markets

Bottom line

Rikiddo is an adaptive scoring-rule market maker that dynamically adjusts liquidity to improve efficiency and stability in prediction markets.