Domain-Weighted 2/3 Majority Tie-Breaking Mechanism Explained

The domain-weighted 2/3 majority tie-breaking mechanism assigns a 1.5× voting weight to a pre-selected domain expert while requiring an option to garner at least two-thirds of the total weighted votes to achieve consensus, ensuring specialized knowledge influences outcomes without permitting unilateral control.

The Council of High Intelligence utilizes this deterministic voting algorithm to aggregate explicit stances from multiple AI agents into a unified verdict. By implementing the domain-weighted 2/3 majority tie-breaking mechanism in the 0xNyk/council-of-high-intelligence repository, the system balances expert influence with collective agreement, preventing arbitrary decisions when the panel faces evenly divided opinions.

Domain-Weighted Seat Designation (STEP 0)

At the initiation of every session, the coordinator executes STEP 0 to establish weighted voting before any analysis occurs.

Pre-Session Expert Selection

According to SKILL.md (line 75), the coordinator designates a single "on-domain" member whose expertise most closely matches the problem domain. This designation happens before any substantive analysis, ensuring the coordinator cannot bias outcomes by selecting the expert after seeing preliminary votes.

The 1.5× Weight Multiplier

The designated domain expert receives a base weight of 1.5, while all other members contribute a standard weight of 1.0. This creates an asymmetry where the expert's stance carries 50% more influence than regular members, but cannot single-handedly override a united opposition.

Weighted Tally Calculation

The aggregation process converts individual stances into a mathematical tally based on assigned weights.

Base Weights and Abstentions

As implemented in SKILL.md (STEP 6, lines 56-66), every active member contributes their full weight to the total pool, even if they abstain. Abstentions count toward W_total but do not add weight to any specific option, effectively diluting the consensus threshold.

Total Weight Computation

For a typical three-member triad with one domain-weighted seat:

• Domain expert: 1.5
• Regular members: 1.0 + 1.0
• W_total = 3.5

This total becomes the denominator for the consensus calculation.

The Two-Thirds Consensus Threshold

The mechanism applies a strict mathematical threshold to prevent minority-rule scenarios.

Mathematical Formula

The consensus condition requires:

W_option ≥ (2/3) × W_total

Where W_option represents the accumulated weight voting for a specific choice. As documented in SKILL.md (lines 60-66), this formula ensures that an option must command substantial supermajority support among the weighted votes.

Validation Logic

The system evaluates each option's accumulated weight against this threshold. Only options meeting or exceeding the 2/3 bar qualify for victory. This requirement prevents scenarios where a simple majority (50%+1) could force a decision despite significant opposition.

Handling Indecision and Ties

When no option clears the rigorous threshold, the mechanism preserves uncertainty rather than forcing arbitrary resolutions.

Genuine Splits

If no option reaches the 2/3 threshold, the council reports a genuine split to the user rather than manufacturing consensus. As noted in SKILL.md (lines 65-66), this transparency prevents the system from presenting uncertain decisions as definitive answers.

Sub-Threshold Ties

When two options accumulate identical weights below the threshold, both are presented as an unresolved split (lines 66-67). This distinguishes between "no clear winner" and "genuine disagreement," allowing users to see when the panel is truly divided.

Python Implementation Example

The following implementation demonstrates the weight calculation and threshold validation:

# Example: 3-member triad configuration

# A = domain-weight seat (weight 1.5)

# B, C = regular seats (weight 1.0 each)

weights = {"A": 1.5, "B": 1.0, "C": 1.0}
stances = {
    "A": "option_x",
    "B": "option_x",
    "C": "option_y"
}

# Calculate total available weight

W_total = sum(weights.values())   # 3.5

# Aggregate weighted votes per option

tally = {}
for member, option in stances.items():
    tally.setdefault(option, 0)
    tally[option] += weights[member]

# Establish 2/3 majority threshold

threshold = (2/3) * W_total      # ≈ 2.33

print(f"Tally: {tally}")
print(f"Threshold (2/3 of total): {threshold}")

# Determine consensus

consensus = [opt for opt, w in tally.items() if w >= threshold]
if consensus:
    print(f"Consensus reached on: {consensus[0]}")
else:
    print("No consensus – split presented to user")

Output:

Tally: {'option_x': 2.5, 'option_y': 1.0}
Threshold (2/3 of total): 2.3333333333333335
Consensus reached on: option_x

In this example, the domain-weighted seat (A) aligns with member B on option_x, accumulating 2.5 weighted votes against the 2.33 threshold. If A had instead supported option_y, that option would achieve 2.5 votes and win, demonstrating how the expert can tip balanced panels without overriding strong majorities.

Summary

• Pre-designated expertise: The domain-weighted seat is selected at STEP 0 before analysis begins, preventing coordinator bias (SKILL.md, line 75).
• Asymmetric weighting: Domain experts wield 1.5× votes versus 1.0× for regular members, creating meaningful but not absolute influence.
• Supermajority requirement: Options must secure at least 2/3 of total weighted votes (W_option ≥ 0.666... × W_total) to achieve consensus.
• Transparent uncertainty: When no option meets the threshold or ties occur below it, the system reports genuine splits rather than forced decisions (SKILL.md, lines 65-67).
• Deterministic outcomes: The mechanism provides consistent, reproducible results based on explicit stance lines and mathematical thresholds.

Frequently Asked Questions

How is the domain-weighted seat selected?

The coordinator designates the domain-weighted seat during STEP 0 based on which member's expertise most closely matches the problem domain. This selection occurs before any analysis or voting, ensuring the choice is based on relevance rather than anticipated outcomes, as specified in SKILL.md (line 75).

What happens if the domain expert abstains?

Abstentions count toward the total weight (W_total) but do not contribute to any option's tally. If the domain expert abstains, their 1.5 weight still increases the threshold required for consensus (2/3 of a larger total), but they do not influence which option wins, effectively diluting their own impact while maintaining the mathematical integrity of the system.

Can a single member force a decision?

No. Even with the 1.5× weight multiplier, a single domain expert cannot reach the 2/3 threshold alone. In a three-member panel, the expert controls only 42.8% of total weight (1.5/3.5), requiring at least one additional member's support to approach the 66.6% threshold. This prevents unilateral control while preserving the expert's ability to break ties between evenly divided regular members.

Where is this logic implemented in the codebase?

The tie-breaking algorithm is implemented in SKILL.md (STEP 6, lines 56-66), with the domain-weight designation defined at line 75. High-level documentation appears in README.md (line 257), and the structured stance voting system was introduced in CHANGELOG.md (line 10) according to the repository history.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}