Frozen study
Is signed order flow a usable signal?
No. 51.7% hit rate at the same five-minute, no-delay cell (Wilson lower bound 51.3%), and the gross edge is negative before any cost.
Over 93,868 firings this signal was right 51.7 percent of the time, barely above the coin flip and far below the 55.5 percent the resting book managed. Its gross edge is -0.002 cents, already negative before a single cost is subtracted, which means there is nothing here for costs to eat.
Data · 93,868 observations · 370,423 book snapshots · 4,489 tokens · 11 days · 2026-07-18 to 2026-07-28
| Hit rate | 51.7 % |
|---|---|
| Wilson lower bound (95%) | 51.3 % |
| Gross edgealready negative before any cost is subtracted | -0.0016 cents per firing |
| Firings | 93,868 observations |
Payload snapshot 2026-08-17. These figures are frozen; the interactive page reads the same file.
Full report (docs/research/orderflow_rest-2026-07.md) · Source module (src/orderflow_study.py)
Research tool only: no investment advice, no order placement, no venue affiliation. Public Polymarket and Kalshi data, provided as-is.