Day-trading styles

Definition

Day trading opens and closes positions within one session, holding no overnight risk. Styles differ by holding period and signal: scalping (seconds-minutes, spread/microstructure-scale), momentum/breakout intraday (minutes-hours, catalyst- and volume-driven), VWAP-anchored mean reversion, and opening-range strategies. This entry defines the styles and their mechanics — and carries the KB’s bluntest evidence section.

How it works / structure

• Common structure: no overnight positions (gap risk exchanged for intraday noise), high trade counts, friction as the first-order cost (ms-slippage-friction compounds per round trip), session-structure awareness mandatory (ms-sessions-auctions — open/close auctions and lunch liquidity are different markets).
• Style parameters (engine-executable): session window, setup definition (e.g. opening-range break: range minutes, confirmation rel_volume; VWAP fade: vwap_dist_pct bands), per-trade risk (risk-fixed-fractional, stop distance in intraday ATR), daily loss limit (the session-scale form of risk-max-drawdown-budget — the style’s load-bearing control), maximum trades/day, flat-by time.
• Regulatory frame: US margin accounts making 4+ day trades in 5 business days are pattern day traders — $25,000 minimum equity (FINRA; acct-pdt-rule).

When it applies

Liquid, volatile instruments during high-participation windows; catalyst days (event-earnings aftermath, macro releases) where intraday ranges expand; account structures meeting PDT requirements. Simulation replay of intraday styles needs bar-level or finer data — a data-fidelity constraint the engine enforces on backtest claims.

Risk profile & failure modes

• The documented base rate is failure: Barber-Lee-Liu-Odean (2014), using complete Taiwan trading records, found the large majority of day traders lose after costs, and predictably skilled traders were a small single-digit percentage — the strongest negative evidence attached to any strategy family in this KB.
• Friction is the house edge: per-trade costs recur dozens of times daily; a style must clear them on every trade the way position strategies clear them once.
• Overtrading feedback: losses provoke revenge trades within the same session (bias-overconfidence); the daily loss limit exists because in-session judgment degrades.
• Leverage + speed: intraday margin invites size that one fast market converts into outsized loss.

Evidence & limits

The Taiwan studies (Barber et al 2014 and predecessors) are the canonical evidence: comprehensive data, consistent finding — aggregate day-trading losses, a thin skilled tail, and persistence of both. US PDT mechanics are FINRA-documented. Marketed day-trading win rates and income claims are folklore. The platform documents these styles for completeness and replay fidelity, not endorsement; the evidence section IS the counterweight.

Falsifiable-thesis examples

Illustrations only, not signals:

• “An opening-range-break rule (30-minute range, 2x relative volume gate) on universe U will be net profitable after modeled friction this quarter in bar-level replay” — falsified by the replay P&L.
• “X, gapping over 5% at the open on earnings, will close above its first-30-minute VWAP” — falsified by the session’s close.

Cross-references

• Tools: indicator-vwap, indicator-volume-profile, indicator-atr
• Microstructure: ms-liquidity, ms-sessions-auctions, ms-bid-ask-spread
• Controls: risk-max-drawdown-budget (daily form), mgmt-stop-loss, bias-overconfidence
• Regulatory: acct-pdt-rule, acct-margin-rules