Skip to main content
Strategy Lab / Day-Trade / BULLRANDOM

BULLRANDOM BullRandom v1.0 - Random-Selection Control

Required comparison baseline: randomly selecting among stocks that merely pass basic tradability (has a price, a sane spread) - the floor every real engine must clear to be worth running at all. Version v1 - status SHADOW.

Equity
$9992.66
Closed Trades
1
Win Rate
0.0%
Evidence Status
Insufficient Evidence
N=1 closed trades

NOT a strategy - a comparison baseline. No scoring, no ranking, no pattern logic: a seeded random draw (stable per session, not re-randomized on every 60-second poll) among symbols passing the same basic tradability gate every engine shares.

Same entry window, risk model, sizing, and force-flat deadline as BullPulse - any performance gap vs. the real engines is attributable to their actual entry logic, not a different execution model.

Deliberately excluded from the adaptive-learning layer (learning.py) - same reasoning as BullRaw.

Known limitations:

  • Synthetic bid/ask spread and 5-minute-bar execution resolution, same as every other engine here.

This is a research hypothesis, not a proven profitable methodology. See docs/strategy_lab_day_trade_plan.md for the full repo-impact plan.

Evidence (Insufficient Evidence):

Needs at least 20 closed trades before any read is meaningful (1 so far). No significance testing is applied even above that floor - see module docstring for what's deferred (PBO, deflated Sharpe, bootstrap CIs).