Order-Flow Imbalance (research)
researchRESEARCH MODE — does not trade. Reconstructs the L2 order book and records (OFI z-score → forward return) observations for offline analysis. Graduates to a trading candidate only if forward return rises monotonically across OFI buckets net of modelled Alpaca cost (core assessOfiMonotonicity). Requires the orderbooks channel (CRYPTO_ORDERBOOK_ENABLED).
Pros
- + Order-flow imbalance is one of the best-documented short-horizon predictors of price on venues with real depth.
- + Signal-only means it can be validated safely with zero execution risk before a single dollar is committed.
- + Produces a falsifiable, bucketed forward-return curve — a clear go/no-go gate rather than a vibe.
Cons
- − Alpaca gives no queue position and a coarse book, so the signal is a weaker proxy than the playbook assumes.
- − Taker-only crypto fees will likely swamp the raw edge even if the signal is real — hence research-only here.
- − Book snapshots can lag or cross; observations are only recorded on a valid, uncrossed book.
Rejection gate
Reject if the recorded (OFI z → forward return) buckets are not monotonically increasing, OR if the monotone edge does not exceed the modelled round-trip taker cost — i.e. even a perfect signal would not pay for its own execution on Alpaca.
Research verdict
Entry gate funnel (last 24h)
No gate telemetry yet — the engine flushes funnel windows every few minutes while running.
Adjustable parameters
Only |OFI z| above this is logged as a signal observation; lower captures more of the distribution. · candidate range 0.5–2.0
How far ahead the forward return is measured for the monotonicity test. · candidate range 1–10
Skip observations when the book is abnormally wide or stale. · candidate range symbol-relative
Throttle observation logging on very active pairs. · candidate range 0–60s
Recent markouts (0)
No markouts recorded yet.
Recent orders (0)
No orders yet.