Alpha, Degen & learning

Alpha Scoring turns current market and on-chain inputs into a screen-specific research rank. Holder Analytics supplies part of that evidence by measuring ownership concentration and established-wallet participation. Live Learning Data then compares recorded scores and components with later 24-hour outcomes.

Use the three together: open a score to see why a token ranked, inspect holder evidence to judge who controls supply, then check Learning to see how much historical evidence exists behind each factor. None of the three is a standalone buy signal.

Top New looks for recent launches gaining real trading traction. Overlooked looks for older tokens showing renewed interest after falling from their peak. Potential CTO looks for quiet tokens with characteristics that could interest a community takeover. These screens ask different questions, so a score is not a universal forecast.

Use the available age window and expand the reasons or component breakdown to understand a result. The detailed Alpha methodology documents the exact gates, weights, and treatment of missing inputs.

Degen Mode focuses on launches less than one hour old. The launch-age display shows how much time has elapsed, including minutes and seconds. These tokens often lack the trading history used by Ranked mode.

Degen checks a limited set of early on-chain signals and known-whale participation. Missing data is not verification, and passing the checks does not make a token safe. Supply can be spread across wallets, and activity can be manipulated.

Scores combine disclosed inputs into a research ranking. A score of 80 is not an 80% chance of profit, a price target, or a risk guarantee. Read the underlying components and their availability labels.

Some missing inputs receive a neutral value; others are omitted or affect eligibility. The exact rule depends on the screen. A neutral component is not evidence that an unknown value was checked and found healthy.

Track record starts from a token’s recorded first Alpha sighting and compares it with later observed market caps. It does not retroactively record every token that existed before tracking began.

A peak multiplier is not an achievable or realized return. It does not establish that you could buy at the sighting price and sell at the peak, and does not include your fees, tax, slippage, or position size. Missing or pending observations should not be read as a zero return.

Learning compares recorded component values with later outcomes after enough time and samples exist. It adjusts influence within bounded limits rather than rewriting the system from one successful coin.

The Learning page shows sample sizes, correlations, and adjustments. Correlation is historical evidence, not proof of causation or a promise that a signal will keep working. Small samples and unresolved sightings limit what can be concluded.