Research Score explained

The Research Score is a 0 to 100 number that helps you decide which assets deserve your attention first. It ranks; it does not predict. Here is how it works, in plain English.

A score of 90 is not a 90% chance of profit.

Why a score at all

There are thousands of crypto assets and limited hours in a day. A good first filter saves time by surfacing assets whose trend, momentum, liquidity, tokenomics and risk profile stand out relative to their peers. The score is that filter. What you do with the shortlist is up to your own research.

The eight components

Percentile ranks: grading on a curve

Raw metrics live on very different scales: a trend measured in percent, volume in dollars, volatility in standard deviations. To combine them fairly, each raw metric is converted to a percentile rank between 0 and 100 across the tracked universe. A Liquidity percentile of 80 means the asset is more liquid than about 80% of the assets DUKOIN tracks.

This has an important consequence: scores are relative. In a broad bear market, the best-scoring asset may still be falling, just less than others. A high score means “stronger than peers”, not “strong in absolute terms”.

Missing data and coverage

When a component has no data, DUKOIN does not fill it with a guess. It computes the score over the available components and re-normalizes the weights so they sum to 100%. Today, Trend, Momentum, Liquidity, Tokenomics, Derivatives (for coins with a Bybit perpetual) and Risk resilience are available, which is 80% of total weight, so coverage is about 80%.

Score = Σ (weight_i × percentile_i) / Σ weight_i   (over available components only)
Coverage = Σ weight_i (available) / 100%

Each score also carries a data-quality grade from A to D, reflecting how complete and fresh the inputs are. A score with lower coverage or a weaker grade deserves more caution.

Exclusions

Some assets are left out of the ranking rather than given a low score:

• 24-hour volume below $500,000, where prices are too easy to move.
• Missing market cap.
• Stale data, so a score is never built on old prices.
• Stablecoins and wrapped assets, whose prices track something else by design.

Versioning

The current formula is score_v1. If weights or inputs change, the version number changes too, so you can tell when a score moved because the market moved and when it moved because the method changed. The exact formula is published on the Research Score methodology page.

How to use the score well

1. Start a shortlist, not a buy list. Use the score to choose what to research next.
2. Open the components. An 80 driven by momentum alone is different from an 80 balanced across trend, liquidity and risk resilience.
3. Check coverage and grade. Today’s scores include derivatives positioning (Bybit funding health and open interest relative to market cap) but not yet per-asset on-chain or catalyst data.
4. Watch changes. A score rising from 40 to 70 over a few weeks can be more informative than a static 75. Smart Alerts can notify you of research-score changes on your watchlist.
5. Pair it with risk. Look at the forecast range and your own position size before acting on anything.

A worked example

This is a hypothetical asset, used only to show the arithmetic. Suppose its percentile ranks are Trend 80, Momentum 60, Liquidity 90, Tokenomics 40 and Risk resilience 70, and the three unconnected components have no data.

Weighted sum = 18×80 + 12×60 + 15×90 + 15×40 + 10×70
             = 1440 + 720 + 1350 + 600 + 700 = 4810
Available weight = 18 + 12 + 15 + 15 + 10 = 70
Score = 4810 / 70 ≈ 68.7
Coverage = 70%

Read this as: the asset ranks in roughly the top third of the tracked universe on the evidence available, with strong liquidity and trend but weak tokenomics. That weak tokenomics percentile is worth investigating before anything else, because dilution can undo a good trend.

What the score is not

• Not a forecast. It ranks current conditions; it does not estimate future returns.
• Not a safety rating. Risk resilience is one component among eight.
• Not a judgment of the project. It does not assess technology, team quality or legal status.
• Not absolute. Because it is a percentile, a whole market can fall while scores stay the same.