METHODOLOGY
Opportunity Engine.
How CryptoNerd synthesises multi-signal alignment into opportunity scores — what the scores represent, what they do not, and the experimental status of this pipeline.
EXPERIMENTALLimited backtested history
The Opportunity Engine is an early-stage synthesis pipeline. Opportunity scores and conviction scores are computed from multi-signal heuristics that have limited historical validation. This surface should be treated as directional context — not as a validated predictive system with proven edge. We are actively building the track record required to change this classification.
OPPORTUNITY SCORE — COMPOSITION
Opportunity = f(Alpha Score, Narrative Momentum,
Whale Activity, Predictive Score, Conviction)
Whale Activity, Predictive Score, Conviction)
Exact weights vary by signal availability. Tokens with more aligned signal inputs receive higher scores. The engine does not output a fixed formula — it synthesises available signals using an LLM-assisted ranking step.
PROVENANCE MAP — WHAT IS VERIFIED VS MODELED
Token symbolVERIFIED DATAToken identifier, not inferred.
Alpha score inputMODELED SCOREDerived from Alpha Scanner pipeline (see Alpha methodology). Composite heuristic.
Narrative momentumMODELED SCORENarrative heat score from the NarrativeHeatmap pipeline. AI-assisted.
Whale activity scoreMODELED SCOREAggregated from WhaleSignal entity for this token in last 48h. Depends on scan recency.
Whale countVERIFIED DATACount of WhaleSignal records for this token in last 48h. Factual record count.
Smart money countVERIFIED DATACount of SmartMoneyMove records for this token in last 48h. Factual record count.
Predictive scoreEXPERIMENTALForward-looking probability estimate from the predictive model. Limited validation history.
Conviction scoreMODELED SCOREMulti-signal alignment strength. Number of independent confirming signals — not a probability.
Opportunity scoreEXPERIMENTALTop-level composite. LLM-assisted synthesis of all inputs. Experimental pipeline.
Risk level (low/mod/high)MODELED SCOREHeuristic classification based on token market cap tier, volatility proxy, and signal strength spread.
Explanation textMODELED SCORELLM-generated narrative summary. Should be read as analytical context, not factual assertion.
Key factors listMODELED SCORELLM-extracted summary of contributing factors. May omit or emphasise incorrectly.
CONVICTION SCORE — WHAT IT ACTUALLY MEANS
Conviction (0–100) measures how many independent signals point at the same token — not the probability of upside. A conviction of 80 means many signals are aligned. It does not mean there is an 80% chance of positive return.
≥75
3+ signals aligned
Alpha + Narrative + Whale all point to same token. Highest internal alignment.
50–74
2 signals aligned
Two independent signal types show activity. One may be dominant.
<50
Single or weak signal
One signal source or marginal alignment. Treat with higher skepticism.
ENTRY, EXIT, AND TIMING — IMPORTANT FRAMING
The Opportunity Engine does not generate entry prices, exit targets, stop losses, or holding duration recommendations. Where such values appear on any signal card, they are estimated reference points derived from historical similar-score signals — they are not analyst-set targets and have not been validated as trading signals.
APPROPRIATE USES
- Identifying tokens with multi-signal momentum for further research
- Prioritising which tokens to investigate in depth
- Monitoring watchlisted tokens for signal convergence
- Understanding which narratives have active capital flow
NOT APPROPRIATE FOR
- Direct buy/sell execution based on score alone
- Treating scores as price probability forecasts
- Assuming "institutional confirmation" from signal count
- Short-term trading without additional due diligence
PREDICTIVE SCORE — A SPECIFIC CAUTION
The predictive score is derived from the PredictiveScore entity — a forward-looking model that estimates upside probability based on historical alpha signal patterns and narrative momentum trajectories. This model has limited backtested history and should be treated as EXPERIMENTAL.
When you see "Predictive: 72%" on an opportunity card, this means the model estimates a 72% probability of price appreciation within the forecast window based on historical signals in similar configurations — not based on verified forward-looking data. The model has limited calibration history and confidence intervals are wide.
When you see "Predictive: 72%" on an opportunity card, this means the model estimates a 72% probability of price appreciation within the forecast window based on historical signals in similar configurations — not based on verified forward-looking data. The model has limited calibration history and confidence intervals are wide.
CLASSIFICATION & VERSION
OPPORTUNITY SCORE
EXPERIMENTAL
Limited validation history
CONVICTION SCORE
MODELED SCORE
Signal alignment count, not probability
PREDICTIVE SCORE
EXPERIMENTAL
Forward-looking, calibration in progress
EXPLANATION TEXT
AI-GENERATED
LLM synthesis — contextual only
REFRESH CADENCE
Every 6 hours
Automated pipeline run
LIMITATIONS
- This is an experimental pipeline — historical validation is limited and in progress
- Opportunity scores are heuristic composites, not calibrated probabilities
- Predictive scores have wide confidence intervals due to limited backtesting data
- LLM-generated explanations may be plausible but factually imprecise
- Signal staleness: scores reflect data at last pipeline run; market conditions change faster
- "Multi-signal alignment" does not equal institutional consensus or validated edge
- Not financial advice. All signals require independent verification before acting.