Why Unverified Money Must Stay Unknown Rather Than Become Zero
Explains why missing or unverified funds should be treated as unknown data, not silently set to zero, to preserve integrity and risk management.
Produced with automation, then checked by deterministic quality rules and an independent source-grounded review before publication.
- 01Treating missing funds as zero masks errors and can lead to incorrect position sizing.
- 02Explicitly marking money as unknown preserves auditability and supports reconciliation.
- 03Unknown balances force operators to investigate data gaps before committing capital.
- 04Systems must surface source, timestamp, and freshness of market data to avoid silent zeroing.
- 05Robust controls and clear error states reduce the chance of unintended loss.
Unverified money must remain unknown because silently treating it as zero hides potential data problems and can lead to incorrect trading decisions. When a balance cannot be confirmed, the system should flag it as unknown and require investigation. This approach preserves auditability and protects against hidden losses. For related context, see What an AI Trading Order Preflight Must Verify.
What does “unverified money” mean in a trading system?
Unverified money refers to any amount that the platform cannot confirm through its authoritative sources, such as missing timestamps, absent venue confirmations, or incomplete account state. The uncertainty may arise from network delays, API failures, or mismatched reconciliation data. Because the platform’s runtime status and account state are the ultimate source of truth, any deviation must be highlighted rather than ignored.
Why can’t we simply set unknown balances to zero?
Setting an unknown balance to zero creates a false sense of certainty. It may cause the system to believe there is no capital available, leading to under‑allocation of risk or, conversely, to assume a position is fully funded when it is not. This silent conversion also removes a critical signal that an investigation is needed, increasing the likelihood of downstream errors.
- Zeroing masks the root cause of the data gap.
- Risk limits may be incorrectly applied.
- Reconciliation reports will show mismatches without clear origin.
- Operators lose a clear audit trail for compliance and debugging.
How should a platform handle unknown money?
The platform should retain the unknown status and surface it in the user interface and API responses. It must include the data source, timestamp, and any freshness warnings. This aligns with best practices for market data provenance and supports the principle that missing information is a risk, not a neutral value.
- 01Mark the balance as “unknown” with a clear flag.
- 02Display the last known source and timestamp.
- 03Provide a warning about data freshness or verification failure.
- 04Require operator acknowledgment before using the balance for new orders.
What controls can mitigate the impact of unknown balances?
Owner‑signed limits and agent scopes can be configured to prevent orders that exceed verified capital. By tying order size, daily notional, and loss limits to only confirmed funds, the system reduces exposure when some balances are unknown. However, these controls are not a guarantee; they depend on accurate verification and timely updates.
- Owner‑signed daily notional caps apply only to verified capital.
- Agent keys can be limited to a fraction of confirmed balances.
- Emergency stop can revoke the calling key, halting further activity until verification is restored.
How does this practice relate to market data integrity?
Market data must always expose its source, timestamp, coverage, and any warnings. Treating missing money as zero violates the same principle applied to price feeds: both require explicit acknowledgment of uncertainty. For further reading on data provenance, see Why Financial Market Data Must Show Its Source.
The same discipline is described in the component guide for trading APIs. See Key Components an AI Trading API Must Provide for details on required data fields and error handling.
“Missing or unverified money must never be silently treated as zero.” - platform documentation
Frequently asked questions
Zero‑valuing hides data gaps, can cause mis‑allocation of risk, and removes a clear audit trail, increasing the chance of hidden losses.
The system should flag balances as unknown, include source and timestamp information, and present a warning that verification failed.
Limits apply only to verified capital; they reduce exposure but do not eliminate risk if verification is incomplete.
Investigate the data source, reconcile with venue reports, and obtain operator acknowledgment before using the balance for new trades.
See the article Why Financial Market Data Must Show Its Source for related principles.
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
Build with Felix now.
Felix infrastructure is live through MCP and the API. The full trading app launches September 17.
A running trading agent depends on clear health signals to stay reliable. This article explains the most important signals, why they matter, and how to act on them.
Model reasoning and deterministic policies are two distinct approaches to automated trading. This article explains their mechanics, risk considerations, and when each is appropriate for a trading operation.