Scoring Methodology
Transparency is core to DRepScore. This page explains exactly how we calculate scores so you can trust and verify our methodology.
We believe the most important thing a DRep can do is explain their reasoning. Participation matters, but a DRep who shows up and votes without explanation is less accountable than one who clearly articulates why they voted the way they did.
Our scoring model reflects this belief: Rationale is the highest-weighted pillar at 35%. This rewards DReps who invest in transparency and gives delegators the information they need to evaluate whether a DRep truly represents their values.
Formula
DRep Score = (Effective Participation × 0.30) + (Adjusted Rationale × 0.35) + (Reliability × 0.20) + (Profile Completeness × 0.15)Effective Participation (30% weight)
Raw participation rate adjusted by a Deliberation Modifier that penalizes uniform voting patterns.
Effective Participation = Participation Rate × Deliberation ModifierDeliberation Modifier:
- >95% same direction: 0.70 (30% discount)
- >90% same direction: 0.85 (15% discount)
- >85% same direction: 0.95 (5% discount)
- ≤85% same direction: 1.00 (no discount)
Why? A DRep who votes Yes on 98% of proposals isn't demonstrating thoughtful deliberation — they're rubber-stamping. We reward DReps who engage with proposals individually.
Rationale Rate (35% weight)
Measures how often a DRep provides meaningful on-chain rationale for binding governance decisions, weighted by proposal importance and adjusted with a forgiving curve.
Proposal Importance Weights:
- Critical (3×): Hard forks, no confidence motions, constitutional committee changes, constitution updates
- Important (2×): Significant/major treasury withdrawals, parameter changes
- Standard (1×): Routine treasury withdrawals
- Excluded: InfoActions (non-binding sentiment polls)
Forgiving Curve:
- 0-20% raw → 0-30 adjusted (rewards initial effort)
- 20-60% raw → 30-70 adjusted (linear middle)
- 60-100% raw → 70-100 adjusted (diminishing returns)
The displayed rate is curve-adjusted to match what the algorithm uses.
Quality Threshold:
Rationale must be at least 50 characters to count. Votes with externally-hosted rationale (IPFS/HTTP) that hasn't been fetched yet are given benefit of the doubt.
Why 35%? Rationale is the highest-weighted pillar because it's the single best signal of accountability. A DRep who explains their reasoning enables delegators to verify alignment — without rationale, votes are opaque. We want to heavily incentivize the practice of public reasoning.
Reliability (20% weight)
Measures whether a DRep can be counted on to keep showing up — distinct from participation (which measures how many proposals they voted on). Reliability tracks the pattern of engagement over time.
Four Components:
- Active Streak (35%): Consecutive recent epochs where the DRep voted. Epochs with no active proposals are skipped. Scores 10 points per consecutive epoch, capped at 100.
- Recency (30%): Exponential decay based on how many epochs since the last vote. Voting this epoch = 100; each epoch of inactivity decays the score (half-life ~5 epochs).
- Gap Penalty (20%): Penalizes the longest continuous stretch of inactivity during proposal epochs. Each epoch of gap costs 12 points from a starting 100.
- Tenure (15%): Time since first vote, with diminishing returns (logarithmic curve). Rewards sustained presence in governance without over-weighting longevity alone.
Reliability = (Streak × 0.35) + (Recency × 0.30) + (GapPenalty × 0.20) + (Tenure × 0.15)Why? A DRep who voted on every proposal (100% participation) but only during a single month before disappearing would score high on participation but low on reliability. Delegators need to know their representative will be there for the next vote, not just past ones. Reliability measures that forward-looking trustworthiness.
Profile Completeness (15% weight)
Measures how thoroughly a DRep has filled out their CIP-119 governance metadata profile.
Point Allocation (100 total):
- Name (givenName): 15 points
- Governance objectives: 20 points
- Motivations: 15 points
- Qualifications: 10 points
- Bio: 10 points
- Verified social links: 25 points (1 link) or 30 points (2+ links)
Broken Link Detection:
Social links are validated periodically. Links that return errors or point to non-existent accounts (including soft-404s on platforms like X/Twitter) are flagged as broken and reduce the profile score.
Why? DReps who invest in their public profile provide delegators with context for informed delegation decisions. Listing working communication channels signals ongoing engagement and gives delegators ways to follow a DRep's reasoning beyond on-chain activity.
How It Works
- You select value preferences that matter to you (e.g., Treasury Conservative, Decentralization First)
- We classify each governance proposal using its CIP-1694 type and on-chain data
- For each preference you selected, we calculate how aligned a DRep's voting record is (0-100%)
- Your Match Score is the average of only the categories you selected
Preference Categories
Treasury Conservative
Favors DReps who vote No on large (>20M ADA) treasury withdrawals and show restraint with ecosystem funds. Routine withdrawals (<1M ADA) have minimal impact.
Treasury Growth
Favors DReps who vote Yes on treasury funding proposals, supporting ecosystem development and innovation investment.
Decentralization First
Favors DReps with moderate voting power (not whales) who don't concentrate too much influence in one entity. Based on the DRep's size tier.
Security Focus
Favors DReps who take conservative positions on protocol parameter changes and hard forks, prioritizing network stability.
Innovation Forward
Favors DReps who vote Yes on hard forks and protocol upgrades, embracing technological progress.
Transparency Advocate
Favors DReps with high rationale rates who consistently explain their voting decisions.
Treasury Amount Tiering
We classify treasury withdrawals by amount to provide nuanced scoring:
- Routine: <1M ADA — minimal impact on alignment
- Significant: 1M-20M ADA — moderate impact
- Major: >20M ADA — high impact on treasury preferences
When you connect your wallet, we monitor your delegated and watchlisted DReps to surface the information that matters most. All alerts appear in the bell icon in the header.
Alignment Shift Alerts
You'll be notified when a DRep's alignment score drops by more than 8 percentage points from their previous score. This filters out noise from individual votes while catching meaningful pattern changes.
Inactivity Warnings
If your delegated DRep hasn't voted in over 30 days, you'll see a warning so you can review their activity and consider whether they're still representing your interests.
New Proposals
See how many new governance proposals have been submitted since your last visit.
Vote Activity Summary
Recent votes by your DRep are evaluated against your preferences so you can see whether their voting aligns with your values.
A persistent insight banner below the header summarizes your current DRep's compatibility and highlights how many alternatives score higher for your preferences.
- DRep Data:Koios API (mainnet) — refreshed periodically
- Vote History:Koios /drep_votes endpoint
- Proposals:Koios /proposal_list endpoint with CIP-1694 type classification
- Metadata:IPFS/HTTP URLs from on-chain metadata references
- Catalyst voting is on a separate system and not yet integrated into our scoring.
- Rationale quality uses a minimum length threshold (50 characters) but does not assess depth or thoughtfulness beyond that.
- New DReps may have low reliability scores due to limited voting history rather than genuine disengagement. The tenure component partially mitigates this, but new DReps should focus on building their streak.
- Rationale text is fetched from IPFS and cached on demand; the first time you view a DRep, some rationales may show as "hosted externally" until they are cached.
- Proposal classification relies on CIP-1694 types and treasury amounts; some edge cases may be imperfectly categorized.
- Social link validation uses platform-specific detection for soft-404s (e.g., X/Twitter accounts that return HTTP 200 for non-existent profiles). Detection may have false positives if platforms change their response format.
Questions about our methodology? Suggestions for improvement?
Reach out via $drepscore on Cardano or open an issue on our GitHub.