True Accrumance data analysis dashboard displayed on a workstation
AI-managed crypto portfolio analysis

AI portfolio analysis, checked against a public log

True Accrumance runs predictive models against live market data and records every recommendation. The outcome of each one — win or loss — is written to a log that registered users can review in full.

Representative model curve, not a forecast. Actual entry-by-entry results are published in the Performance Logs section below.
What the platform is

A tool for analysis, not a promise of return

True Accrumance ingests price, volume and volatility data across major crypto markets and runs it through a set of predictive models. The output is a risk-adjusted recommendation, not an instruction.

You keep control of execution. The system is built to reduce the volume of low-quality decisions an investor has to filter through manually, using data rather than intuition as the starting point.

Every recommendation the model produces is timestamped and logged before its outcome is known, which is the basis of the verification process described below.

True Accrumance analyst reviewing model output on screen
Methodology

How results get verified

A performance claim is only useful if it can be checked. This is the process we use to make that possible.

01 / Capture

Data capture

Every recommendation and its outcome is timestamped and written to an append-only log the moment it occurs. Entries cannot be edited after the fact, only appended to.

02 / Snapshot

Independent snapshot

A daily snapshot of the log is stored outside True Accrumance's own infrastructure. This separates the record of what happened from the system that generated it.

03 / Audit

Public audit trail

Snapshots are made available to registered users for download. Any historical entry can be checked against the published record at the time it was logged.

Log integrity checked daily

A daily reconciliation compares the live log against the previous day's snapshot. Discrepancies are flagged before the next trading session and disclosed, not silently corrected.

Losses are published too

The log includes losing positions alongside winning ones. A performance log that only shows favourable outcomes is not a verification log — it is marketing.

Mechanics

What the system does

No single component of this platform makes a decision on its own. Each stage below checks the output of the one before it.

Predictive modelling

The model reviews historical price movement, volume and volatility across several time frames. It outputs a probability-weighted range for short-term movement, not a single fixed prediction. Where the data does not support a confident range, the model returns no recommendation.

Market data
Model
Weighted range

Risk mitigation engine

Before any recommendation is surfaced, a separate process checks position size against account-level volatility limits. Recommendations that breach a preset risk threshold are suppressed at source. They are not shown as an optional, riskier alternative.

Recommendation
Risk check
Pass / suppress

Real-time data handling

Market data feeds refresh continuously. Recommendations carry a fixed validity window and expire automatically rather than remaining on screen once the underlying conditions have changed.

Live feed
Refresh cycle
Expiry check
Community-verified results

Live performance logs

This table mirrors the structure of the log made available to registered users. It does not display live figures — access to the current log is granted on request.

Accuracy figures are calculated weekly against the closed-position log, including losing positions, and published in full alongside the raw entries.
Date Asset pair Model signal Logged outcome Verification status
Available on request
Available on request
Available on request
Before you commit capital

Questions investors ask

Direct answers to the questions we are asked most often, without qualifying language.

Is my capital at risk?

Yes. Crypto markets are volatile and the model's recommendations are statistically probable outcomes, not guaranteed ones. The system is designed to reduce exposure to low-quality decisions, not to remove market risk.

What does True Accrumance charge?

Fee structure is disclosed in full before you commit any capital, and it is available on request through the contact page. We do not vary disclosed fees based on account size after the fact.

Who checks the algorithm?

Model output is reviewed against the independent log snapshot on a fixed schedule. Any deviation between what the model recommended and what was logged is investigated and disclosed, not adjusted quietly.

Can the model be wrong?

Yes, regularly. The model returns a probability-weighted range, not a certainty. Losing positions are recorded in the same log as winning ones, and both count toward the published accuracy figures.

Does the platform execute trades automatically?

The model produces a recommendation and a risk assessment. Execution requires a separate, deliberate action from the account holder — the system does not place trades without that step.

Next step

Review the data before you decide

Request access to the full performance log, or read the methodology paper first. There is no obligation attached to either.