Project Record

Diamond Hand Score

Post-IPO / ICO investor conviction score measuring token retention after distribution, including wallet balances, staking deposits, and LP exposure.

Domains

CryptoToken DistributionInvestor Analytics

Capability

Scoring & Decision Models

Methods

ScoringBehavior AnalysisOnchain AnalyticsTime Series

Executive Summary

The Diamond Hand Score is a post-IPO / ICO investor quality score. It measures what investors actually did after tokens became liquid or unlockable: did they keep exposure, stake, provide liquidity, or quickly move tokens away?

This is deliberately different from a pre-sale “value-add” score. A value-add model estimates who might help before allocation. The Diamond Hand Score measures realized conviction after distribution, using daily on-chain snapshots and project-specific protocol labels.

The score was built around token projects where retention is not just “tokens still in the wallet.” A strong holder may stake, bridge, or deposit into an LP. The pipeline therefore tracks wallet balances, staking contracts, and LP positions across each project’s configured chains before turning the history into a single score.

Problem

After a token launch, headline holder counts hide the behavior that matters. A wallet can appear inactive because it staked. Another can still hold a small balance after selling most of its allocation. A third can move tokens into an LP pair and retain economic exposure even though the wallet balance drops.

We need a scalar score that distinguishes durable post-distribution investors from early sellers while remaining auditable enough to explain why a user landed in a given tier.

Score

For investor or wallet $i$ , the score is a time-weighted average of retention:

D_i = \frac{\sum_{t \in T_i} e^{-k(t-t_{0,i})} r_{i,t}}{\sum_{t \in T_i} e^{-k(t-t_{0,i})}}

where:

$t_{0,i}$ is the first date investor or wallet $i$ received tokens.
$T_i$ is the set of snapshot dates from first receipt onward.
$k$ is the decay rate, with half-life $\ln(2) / k$ .
$r_{i,t}$ is retention at snapshot date $t$ .

Retention is defined as tokens still under economic control divided by the relevant baseline:

r_{i,t} = \frac{wallet_{i,t} + staked_{i,t} + lp_{i,t}}{baseline_i}

At user level, $baseline_i$ is the amount first received. At wallet level, where allocation and vesting data is available, the baseline can be the unlocked allocation at that date. Retention is capped before scoring so secondary purchases or airdrops do not dominate the metric.

The exponential weighting makes first-receipt behavior important. An investor who held from day zero scores better than one who sold early and bought back later.

Protocol Tracking

The numerator counts tokens still controlled by the investor:

wallet balance
staking protocol deposits
LP positions

Transfers are classified per project using contract labels. Sends to known staking contracts increase staked balance. Sends to known LP pairs increase LP balance. Generic transfers, router outflows, and unclassified destinations reduce retention unless a project-specific label marks them as control-preserving.

That project-specific classification is what makes the score usable across different token launches: each project can define its staking contracts, LP pairs, routers, and other protocol addresses without changing the scoring formula.

Implementation

The pipeline is intentionally mechanical:

Fetch ERC20 Transfer events from each configured chain.
Classify transfer intent using project-specific contract labels.
Build daily wallet and user snapshots with wallet, staked, LP, and transferred-out balances.
Merge chains when a project has multi-chain token activity.
Compute user-level and wallet-level Diamond Hand Scores from first receipt onward.
Aggregate across projects when needed, weighting each project score by accepted allocation USD.

The output is a transparent score per wallet or Legion user, plus the underlying snapshot history needed to inspect the result.

Interpretation

Score	Label	Interpretation
0.80-1.00	Diamond Hand	Retained most exposure after distribution
0.50-0.79	Holder	Partial selling or mixed holding behavior
0.20-0.49	Paper Hands	Sold the majority of received or unlocked tokens
0.00-0.19	Dumped	Sold or transferred nearly everything early

Trade-offs

The score depends on high-quality contract labeling. Unknown staking or LP addresses can make a loyal holder look like a seller, so the label set needs review per project. The decay rate is also a product choice: a short half-life makes the score react quickly to early selling, while a long half-life favors slower-moving investor behavior.

The advantage is interpretability. Every score decomposes into a time series of daily retention ratios, and every retention ratio can be traced back to wallet, staking, LP, and transfer classifications.