Performance Methodology
This page describes how TRADEOS.tech measures, records, and will publish its live performance track record — including the standards we hold ourselves to before making any performance claims public.
TRADEOS.tech does not publish live performance numbers until an independent audit of the on-chain attestation record is complete. This page describes the methodology that record is built on and the verification process any auditor or institutional evaluator can run.
The problem with trading performance claims
Performance claims in algorithmic trading are easy to fabricate and hard to verify. Common misleading practices include:
- Backtest overfitting — parameters tuned to historical data produce impressive backtests that fail immediately in live trading
- Selective disclosure — showing only the winning periods, hiding the drawdown periods
- Paper trading presented as live trading — simulated results with no real capital at risk
- No slippage accounting — backtests that assume perfect fills at signal prices
- Survivorship bias — the strategy you're shown is the one that survived; failed strategies are not disclosed
TRADEOS.tech addresses each of these directly — not through claims, but through architecture.
How performance is recorded
Every trade is immutable from inception
When a trade intent is created, it is written to the audit ledger with a unique ID, timestamp, signal attribution, regime classification, and HMAC hash. This creates a tamper-evident chain — any modification to a historical record breaks the hash chain and is immediately detectable.
The ledger records:
- Signal that triggered the trade (with its score, confluence factors, and regime)
- Intended entry price and size
- Actual fill price, size, and exchange-reported fees
- Exit trigger (signal, stop, take-profit, manual, risk event)
- Actual exit price and fill
- Net P&L including fees and funding costs
This is actual execution data, not simulated fills.
Daily on-chain attestation
Every day, a cryptographic Merkle root of all signal outcomes is committed to Base L2 — a public, immutable blockchain. The on-chain root cannot be altered retroactively. Any auditor can:
- Read the daily root from the
SignalAttestationcontract - Request the Merkle proof for any specific signal from the API
- Recompute the leaf hash from the signal's raw fields
- Verify the proof independently — no TRADEOS.tech infrastructure access required
See On-Chain Attestation for the full technical specification.
This means: the signal activity record for every day TRADEOS.tech has been live is permanently anchored to a public blockchain. TRADEOS.tech cannot selectively disclose only good days. Every day is recorded. Every record is verifiable.
Slippage and market impact are recorded, not assumed
TRADEOS.tech records both the signal-at-time-of-generation price and the actual exchange fill price. The difference — execution slippage — is tracked by the TCA (Transaction Cost Analysis) monitor service. Performance numbers reflect actual fills, not theoretical prices.
What "audited performance" means here
Before TRADEOS.tech publishes a live track record as a marketing claim, it will be independently audited through the following process:
Step 1 — On-chain record verification An independent auditor reads the full sequence of daily Merkle roots from the Base L2 contract and verifies:
- A root exists for every day in the claimed period (no gaps)
- The roots are consistent with the database export provided
Step 2 — Database export reconciliation
The auditor receives a full export of the signals, orders, fills, and audit_ledger tables for the audit period. They verify:
- HMAC hash chain integrity (no tampered records)
- Every on-chain root matches the Merkle tree computed from that day's signal export
- Fill prices match exchange-reported trade history (cross-referenced via exchange APIs)
Step 3 — P&L calculation P&L is computed by the auditor from raw fill data — not from TRADEOS.tech-reported numbers. This ensures the performance figure is not a function of any internal calculation TRADEOS.tech controls.
Step 4 — Publication The audited period, auditor identity, methodology, and final performance figures are published. The on-chain roots for that period remain publicly verifiable indefinitely.
Standards we apply to performance claims
We will not publish a performance number unless:
- It covers a minimum of 90 consecutive live trading days with real capital
- It is accompanied by the maximum intraday drawdown in that period, not just the end result
- It specifies the strategy profile (conservative / balanced / aggressive) used
- It is accompanied by the Sharpe ratio and Sortino ratio calculated on actual daily returns
- The underlying data has been independently audited as described above
- The on-chain attestation record for the full period is publicly accessible for independent verification
We consider any performance claim that does not meet all six standards to be misleading — regardless of how impressive the number is.
Current status
TRADEOS.tech is actively accumulating a live track record. The on-chain attestation system is operational and committed daily. The internal audit ledger is active and hash-verified.
We are not yet at the 90-day minimum for a publishable audited track record. When we are, the audit process described above will be initiated, and results will be published here with the full methodology and verification instructions.
How to verify independently today
Any evaluator can already verify that the system is live and operating:
- Request the contract address for the
SignalAttestationcontract on Base L2 mainnet from the team - Call
roots(dateKey)for any recent date (format:YYYYMMDDasuint32) - A non-zero
bytes32value confirms an attestation was submitted for that date - Cross-reference the signal count via
GET /attestations?date=YYYY-MM-DDfrom the metrics API
This is not a performance claim. It is proof that the system ran on that date and generated signals — verifiable by anyone, without trust.