Predictive modelling for student investors
Eternally applies predictive modelling to publicly available market data, helping students assess risk before committing capital rather than reacting after a loss. Every output is logged and open to review.
Methodology
The process below is published in full. Each stage exists to reduce guesswork, not to produce a single confident signal.
Market data is pulled at fixed intervals from public exchanges and on-chain records. No private or paid data feeds are used, so the inputs can be independently verified.
A predictive model screens the data for volatility patterns, volume anomalies and historical precedent, producing a structured output rather than a single buy or sell call.
Each output is assigned a risk score based on volatility range and liquidity depth, giving students a reference point for position sizing rather than certainty.
The final output, risk score and timestamp are written to the public log before any market outcome is known, preventing results from being edited in hindsight.
Key capabilities
Each tool is designed around a single task. None of them claim to predict a market with certainty; they narrow the range of uninformed decisions a new investor might otherwise make.
RISK / SCORING
Translates volatility and liquidity data into a simple score, so a student can compare the relative risk of two assets before allocating funds.
ALERTS
Flags when an asset moves outside its recent trading range, giving an early signal to review a position rather than react mid-swing.
HISTORY
Places current price behaviour alongside comparable historical periods, making it easier to judge whether a move is typical or exceptional.
SIZING
Calculates a suggested allocation based on a student's stated risk tolerance and available capital, keeping individual positions proportionate.
Transparency
Every model output is written to a public log before the outcome is known. Nothing is amended after the fact, so the record reflects what the model actually said, not what would look best afterwards.
| Timestamp (UTC) | Asset pair | Model output | Risk score | Status |
|---|---|---|---|---|
| 2024-11-03 09:00 | Sample pair A | Hold — low confidence | Moderate | Logged |
| 2024-11-03 15:00 | Sample pair B | Review position | High | Logged |
| 2024-11-04 09:00 | Sample pair A | No action | Low | Logged |
| 2024-11-04 21:00 | Sample pair C | Reduce exposure | High | Logged |
Illustrative log format — figures above represent structure only and not actual market outputs.
Each entry is time-stamped and hashed at the point of publication. Students and outside reviewers can cross-reference the timestamp against public market data to confirm the output was not altered retrospectively.
The log is append-only. Past entries cannot be edited or removed, including entries that were later shown to be wrong.
Illustrative distribution of logged entries by asset category, not a performance claim.
Getting started
The onboarding sequence is deliberately slow. It is designed to delay capital commitment until a student understands the terms the model uses to describe risk.
Before accessing any output, new users read the published methodology and the current data source list, so the reasoning behind a score is never a black box.
A short glossary covers the terms used throughout the log, including drawdown, volatility and position sizing, before any real figures are shown.
Students are encouraged to follow the public log for a period before allocating any capital, treating early access as observation rather than instruction.
About the approach
Eternally was built on the view that most new investors are not short of opportunities; they are short of a consistent way to evaluate them. The platform does not promise outsized returns, because no model can guarantee one.
Instead, it standardises how risk is measured and makes that measurement visible, so a student can compare their own judgement against a documented process over time.
Next step
Access is free to request and includes the full methodology, the current public log and the risk glossary. There is no obligation to trade.
Access Data