Where do verified trade calls fit in a Python backtest?

AlgoVault is the verified trade-call source your Python backtest reads, not a backtester. It returns one composite verdict per call — direction, confidence, regime — with a 91.6% PFE win rate over 233,000 calls, each Merkle-anchored on Base L2. Your library runs the test against that series.

Is AlgoVault a backtester?

No. AlgoVault is not a backtester and never claims to be. It returns interpreted verdicts; your Python backtesting library runs the simulation. The clean split is the point: you stop maintaining a factor engine and a track-record ledger, and you test against a series anyone can verify on-chain.

How do you get a verified call series?

Call get_trade_call for live verdicts, or read the published track record for the historical series. Map each verdict to an entry or exit in your loop:

# AlgoVault returns the verdict; your library runs the backtest.
verdict = call_tool("get_trade_call", {"coin": "BTC", "timeframe": "1h"})
# -> {"call": "BUY", "confidence": 72, "regime": "TRENDING_UP", ...}

if verdict["call"] == "BUY" and verdict["confidence"] >= 60:
    strategy.buy()        # Backtrader / vectorbt / your own loop

How are the calls verified, so the backtest is honest?

Each call is hashed and published on Base L2 before its outcome is known. That removes hindsight: you cannot tune a strategy against verdicts that were edited after the market moved, because they were fixed beforehand. Recompute any hash at /verify across 5 venues.

Which backtesting libraries pair with it?

Any Python library — Backtrader, vectorbt, or a custom loop. AlgoVault returns the verdict; the library handles fills, slippage, and metrics. See the docs for the response shape.

FAQ

Is AlgoVault a backtester?

No — it is not a backtester. It is the verified trade-call source your backtest reads; your library runs the test.

How do I get a verified call series to test against?

Use get_trade_call for live verdicts or the track record for history. Both are Merkle-anchored.

How are the calls verified, so the backtest is honest?

Each call is hashed on-chain before its outcome is known — no hindsight tuning. Check any hash at /verify.

Which backtesting libraries does it pair with?

Any Python library. Map the verdict to an entry or exit; the library handles fills and metrics.

This is call interpretation, not investment advice; agents decide execution.