How to Verify a Live Account‑Value Figure for Trustworthy Trading
Learn the essential steps and controls needed to confirm that a live account‑value number is accurate, reliable, and auditable for any trading operation.
Produced with automation, then checked by deterministic quality rules and an independent source-grounded review before publication.
- 01The source, timestamp, and freshness of market data must be recorded for every valuation.
- 02Owner‑signed limits and agent‑key scopes provide a clear audit trail for who can affect account balances.
- 03Durable mutation identifiers allow reconciliation of updates even when timeouts occur.
- 04Explicit error states and reconciliation logs are required to distinguish failed orders from delayed confirmations.
- 05Missing or unverified money must never be silently treated as zero; it should trigger a warning and manual review.
A live account‑value figure is verifiable when the number can be traced to an authoritative data source, confirmed by immutable transaction records, and reconciled against the owner’s signed controls. This requires transparent market‑data provenance, robust key management, and explicit error handling that together eliminate hidden assumptions about the figure’s correctness.
What data sources make an account‑value figure trustworthy?
The valuation must be built from market data that includes a clear source identifier, a precise timestamp, and a freshness indicator. If any of these elements are missing, the figure should be flagged for review. For example, a price feed that does not disclose its timestamp cannot be trusted for real‑time valuation.
- Source label (exchange, aggregator, or broker).
- Timestamp in UTC with millisecond precision.
- Freshness metric indicating how recent the data is.
- Coverage warnings for missing symbols or halted markets.
How do key‑management controls support verification?
Owner keys hold ultimate authority over funds, while agent keys operate under narrowly defined scopes and limits. Each action that changes the account balance must be signed by the appropriate key, and the signature must be recorded alongside the transaction.
- Owner‑signed limits for order size, daily notional, and loss thresholds.
- Agent‑key scopes that restrict which assets or strategies can be used.
- Separate withdrawal authority that requires an explicit owner‑signed intent.
Why are durable mutation identifiers important?
When an order is submitted, the system creates a mutation identifier that persists across retries and timeouts. This identifier enables the reconciliation process to match a later confirmation with the original request, preventing false assumptions that a timeout equals a failure.
- Identifier remains unchanged across network retries.
- Allows auditors to trace the full lifecycle of a trade.
- Supports automated reconciliation without manual guesswork.
What role do explicit error states play in verification?
Every operation should return a distinct error code that describes the failure reason. Generic timeouts are insufficient because they do not indicate whether the order was rejected, partially filled, or still pending.
- Reject codes for insufficient margin, limit breaches, or invalid signatures.
- Pending states that require follow‑up status checks.
- Clear success codes that confirm settlement and balance update.
How can traders ensure that missing money is not silently ignored?
If market data or transaction logs contain gaps, the valuation engine must raise a warning rather than substitute zero. This forces a manual review and prevents the illusion of a balanced account when funds are actually unaccounted for.
- Automated alerts for data gaps or mismatched totals.
- Mandatory manual reconciliation steps for any warning.
- Documentation of the review outcome in an audit log.
What practical steps can be taken today?
- 01Integrate market‑data feeds that provide source, timestamp, and freshness fields.
- 02Configure owner‑signed limits and agent‑key scopes according to your risk policy.
- 03Implement durable mutation identifiers for every order submission.
- 04Log explicit error codes and set up monitoring for timeout vs. failure distinctions.
- 05Create an audit workflow that flags any missing or unverified money for manual review.
Verification is not a single check; it is a layered process that combines data provenance, key control, immutable identifiers, and clear error handling.
Frequently asked questions
No. A single feed may lack redundancy and can hide outages. Using multiple sources and comparing them improves confidence, as discussed in [When Should a Trading Agent Compare Multiple Price Sources?](/blog/when-compare-multiple-price-sources).
A timeout does not prove failure. You must query the order status using its mutation identifier to determine whether it succeeded, is pending, or was rejected.
Limits should be reviewed whenever your risk appetite changes, after major market events, or when you add new strategies that could affect exposure.
The official [Felix documentation](/docs) provides detailed guidance on key management, limits, and error handling. Additional reading on verification processes can be found in [How an AI Trading Order Preflight Must Verify](/blog/ai-trading-order-preflight-verification).
Sources and verification
Product claims in this article were checked against these first-party references. Runtime status remains authoritative for current availability.
- Felix documentationfirst party
- Felix machine referencefirst party
- Felix runtime statuslive status
Build with Felix now.
Felix infrastructure is live through MCP and the API. The full trading app launches September 17.
A trading agent that relies on a single quote can be misled by outages, latency, or erroneous data. This article explains when and why agents should compare multiple price sources, what controls to apply, and how to handle conflicting information.
A clear freshness signal helps users trust market data and avoid costly mistakes. This article explains what information to show, how to format it, and why transparency matters for risk‑aware trading.