Understand how M2M works — from the signals we detect to the methodology behind our backtests.
M2M evaluates and trades both directions — bullish (calls) and bearish (puts) under the same gates, equally strict. As of 2026-05-15 the bear path is in shadow mode: bear-direction setups are tracked and paper-traded, but held off live execution until our re-backtest of the 2024-05 → 2026-05 hold-out confirms PF ≥ 1.3. In sharp drawdowns the regime gates suppress bullish setups independently of direction — you may see fewer or no signals when the regime is unsuitable, by design.
Confidence A (POP), B (edge), C (EV) — why three numbers, not one
Field-by-field walkthrough of every element on a setup card
Three-gate sequence: backtest, shadow mode, live release
Market, short-interest, short-volume, and index data — what each source covers and what it does not
Explicit limitations, what M2M sees vs. what it doesn't, inference vs. measurement
5 curated scenarios — would you Take, Skip, or Watch? See M2M's call and the actual outcome.
21 terms explained in plain English
Walk-forward validation, execution assumptions, regime gating
6-factor quality engine and squeeze scoring explained
Parameters are optimized only on in-sample data (Jan 2020 – Dec 2022). All published results come from the out-of-sample period (Jan 2023 – present) — data the optimizer never saw. Strategies that degrade significantly between periods are shelved, not published.
Every simulated trade includes 0.05% slippage and round-trip transaction costs. Entries execute at next-day open (no same-bar entry). Gap-through stops are modeled — if the market opens past your stop, you get filled at the open price.
Our 5-state trend regime engine classifies the market using SPY, SMA(50/200), and ADX(14). Signals only publish during strong trends (ADX > 25).
Stop losses scale to each stock's volatility using ATR. Low-volatility stocks get tighter stops; volatile names get wider ones (4%–16% range). This prevents the common mistake of fixed percentage stops that destroy performance on high-beta names.
Every strategy signal is scored on 6 weighted factors, each contributing to a total score of 0-100. The primary publish gate is 65+ AND 4+ factors above their midpoint. A narrow tier-3 fallback also publishes signals at 40+ with at least 2 factors above midpoint when the detector class warrants it — those rows are markedpasses_multi_factor=false and carry the preliminary label until they accumulate enough sample size to be calibrated.
The flow scanner uses a composite squeeze score (0-100) measuring how many factors align for a potential squeeze: unusual options activity, short interest, gamma positioning, IV dynamics, and technical compression. Higher scores indicate stronger confluence.