01 · NO PEEKING
Walk-forward
Each year's settings are chosen from the five years before it, then scored on that unseen year.
RESEARCH / ALTERNATIVE DATA
Testing trading rules,
honestly.
I collect market prices, Chicago-area traffic and local weather on a home server, then test whether simple rules beat buying and holding. Every rule is scored on years it never saw, after trading costs.
02 · REAL FRICTION
Costs included
Every entry, exit and rebalance pays 0.05% of the money moved.
03 · THE TRAP, SHOWN
Hindsight beside it
Each rule also appears tuned after the fact, to show how much a curve-fit backtest flatters it.
WALK-FORWARD · FULL PRICE HISTORY
Does any rule beat just holding?
Loading…
| Strategy | Yearly growth | Beat holding | Profitable years | Worst drop | Time invested | With hindsight |
|---|
TRY IT YOURSELF
Move the sliders.
Change the signal threshold and the amount per trade, and watch every chart and total recalculate from the same day-by-day projections.
WHAT-IF · $30 RULE · HONEST TEST IN THE STRATEGY LAB ↗
Buy $30 at a +10% projection, sell at −10%
Calculating…
Each day uses only the prices known that day: if the 60-day trend projects +10% or more over the next 21 trading days and nothing is held, buy $30 at that close; if it projects −10% or worse, sell everything. No fees or taxes. Compared with buying $30 of every symbol at the start of the window and holding. Hypothetical, not advice.
Loading market data…
Projections continue the last 60 trading days' trend for 21 trading days, with an 80% range based on recent volatility. Exchange holidays aren't skipped. A projection, not a prediction to trade on.
FORWARD-LOGGED · TRAFFIC + WEATHER
Do busy roads predict transport stocks?
| Day | Rush-hour speed | vs. usual | Signal | Weather | Basket next day | Next 5 days | Logged |
|---|
Hypothetical research on historical data. Not investment advice; nothing here places or recommends a trade.