Active public registry — Protocol 04.2
A financial analysis that documents each of its calculations.
Mideravel analyzes large volumes of financial data, quantifies risk before each recommendation and records each output in a verifiable public log. You manage the capital at your own pace, including when your time zones change.
Exposure reduction — filtered volatility
Risk score 0.62 · Configured threshold 0.60 · Proposed action
Delayed rebalancing — summary 07:00 CET
Risk score 0.34 · Within tolerance · Asynchronous digest
Macro hedging — currency correlation
Risk score 0.71 · Exceedance · Validation required
01 — Operational context
Investing remotely requires continuity, not constant monitoring.
When activity moves between time zones, manual market monitoring becomes fragmented. Information arrives outside of working hours, arbitrations are postponed and risk thresholds are applied irregularly.
Mideravel treats this constraint as a design parameter. The platform continuously observes the data, applies your risk limits systematically and sends you a readable summary at the time you choose.
Risk containment
No recommendation is issued without a risk score, associated drawdown limit and registration in the register. If a threshold is exceeded, the system proposes an exposure reduction before any other action.
Dispersed watch
Multiple sources consulted separately, interpretation dependent on availability, history difficult to reconstruct.
Unified ingestion and filtering
Aggregation of flows, normalization of series and filtering of short-term volatility before any modeling.
Thresholds applied irregularly
Loss limits are adjusted after the fact, often under time or connectivity constraints.
Settings applied continuously
Your drawdown, exposure and concentration limits are monitored with each calculation, with full traceability.
The comparison concerns the organization of the analytical work. It does not prejudge any financial results.
02 — Public performance register
Verifiable results, not claims.
Each model output — signal, risk score, recommendation — is entered into a public registry with timestamp, model version and verification fingerprint. Independent community nodes replay calculations from the published data schema.
| Timestamp (UTC) | Model version | Signal type | Risk score | Verification Hash |
|---|---|---|---|---|
| 2026-11-07 14:30:00 | v2.4.1 | Exposure reduction | 0.62 — threshold 0.60 | 9f2a 77c1 ··· 41cd |
| 2026-11-07 07:00:00 | v2.4.1 | Asynchronous digest | 0.34 — content | b81e 02aa ··· 9f04 |
| 2026-11-06 21:15:00 | v2.4.0 | Macro coverage | 0.71 — validation required | 44d0 c9e2 ··· 17b8 |
| 2026-11-06 09:45:00 | v2.4.0 | Maintenance of allowance | 0.28 — content | 7a3c 5d11 ··· e2f6 |
| 2026-11-05 16:05:00 | v2.3.9 | Concentration alert | 0.58 — under supervision | c90b 41f7 ··· 03ad |
Reading the table. The lines above illustrate the registry structure, not past performance. In production, each hash allows you to find the inputs, the version of the model and the risk rule applied.
Community audit. Verification nodes compare the published fingerprint with an independent recalculation. Any discrepancy is reported publicly and blocks the certification of the line concerned.
03 — Methodology
A simple pipeline, from data to traceable decision.
The system operates without continuous presence on your part. You set the limits. Mideravel applies the method and documents each step.
Multi-source ingestion
Aggregation of financial flows, standardization of formats and quality control before any calculation.
- Market data and macro indicators
- Deduplication and time alignment
Modeling and filtering
Predictive models estimate trajectories and isolate short-term noise from usable information.
- Risk score 0 to 1 per scenario
- Volatility and concentration filtering
Execution or digest
Depending on your authorizations, transmission of a recommendation or supervised execution via your intermediaries.
- Asynchronous digest per chosen zone
- Validation required beyond thresholds
Registration in the register
Each decision is timestamped, versioned and sealed with a hash that can be viewed by the community.
- Immutable public audit log
- Computation-independent replay
04 — Operational scenarios
Three common situations for remote control.
Explicit rules, adapted to capital management distributed between several jurisdictions and work rhythms.
Capital preservation and drawdown limits
You set a maximum acceptable loss per portfolio. If the risk score approaches the limit, the system first suggests a calibrated exposure reduction and then logs the decision.
volatilité 21j · corrélation devises · exposition nettesi score > seuil_utilisateur → proposer réduction proportionnellerecommandation + score 0,62 + limite drawdown + hashdigest 07:00 CET · validation requise si > 0,70Illustrative block. It describes the applied logic, without predicting performance.
Asynchronous rebalancing by time zone
You receive a summary at a fixed time, in your reference zone. Minor adjustments are grouped together. Significant deviations trigger a separate alert with context and history.
Coverage during macro events
In times of uncertainty, the engine assesses currency and rate exposure, then documents hedging options compatible with your liquidity and jurisdiction constraints.
05 — Due diligence
Practical questions before setting up tracking.
Direct answers on capital control, security and time required.
Do you maintain control of your capital?
Yes. Mideravel is non-custodial. Your funds remain with your banks and brokers. The platform analyzes, recommends and, only if you authorize it, transmits instructions within your limits. You can pause any automation at any time.
How secure are my account logins?
Connections use official API interfaces with default read-only permissions and scope-restricted tokens. Write access requires explicit validation and remains capped by your exposure thresholds.
What happens if thresholds are exceeded?
The system interrupts the current sequence, offers exposure reduction and requires validation for high scores. Each event is entered in the register with the score, the rule triggered and the timestamp.
How much weekly monitoring time should you expect?
The initial configuration requires a precise definition of your limits and your synthesis zone. Then, reading the periodic digest and occasionally validating alerts is sufficient in most organizations. No continuous screen monitoring is required.
How do you check a prediction after the fact?
Each line in the registry contains the model version, referenced inputs, and hash. You can compare the recommendation, score and documented outcome, or replay the calculation from the published schema.
06 — Measured access
Start with the registry. Then configure at your own pace.
Review public logs, verify data format, then test a sandbox model with your own risk thresholds. No commitment and no transfer of funds.
Operational system · UTC reference data · Documentation available in French.
Configure your risk settings
Describe your monitoring organization. We prepare a sandbox space aligned with your time zone and your limits.