Coiling stocks, ranked — then scored on what they actually did

The question

Which S&P 500 stocks are coiling for a breakout — and do the setups actually work when every pick is scored afterward?

How it works

A daily scan looks for consolidations of 4+ sessions that have not yet fired, plus bullish pattern setups, and ranks what it finds. Setups that have already triggered are dropped, not chased. Every pick is then scored over the following 5 sessions — target hit, stop hit, or neither — and near-misses are logged and scored as a separate tier so they never inflate the main record. The scanner learns from its own record, but cautiously: scoring weights adjust only after at least 10 scored picks.

On Oct 26, 2026 the module published its own correction: targets had been set too far from the trigger. Re-anchoring targets to 1.0× ATR(14) lifted backtested target hits from 4% to 41% — and the 4% is still on the record, in the open.

Cadence

Scan runs daily on weekdays; every pick's 5-session scoring window closes automatically and feeds the learning loop.

Current status — Oct 28, 2026

The target-math fix of Oct 26, 2026 is in force, and the forward record is accumulating under it. The all-time record is stated honestly: 4% target hits before the fix; 41% backtested after re-anchoring. The founding / scorecard study — a full performance audit — is planned.

Studies

• Planned — “Do Breakout Coils Work? A full performance audit and the target-math fix” · Founding / Scorecard Study

Scorecard summary

Pre-fix 4% · post-fix backtest 41% · every live pick scored over 5 sessions. See the full scorecard →