Skip to content
Parth Parekh
NOTENOTE3 MIN READ
← ALL NOTES

Not every hypothesis graduates

Ideas die in backtest. That's fine. The gates are what keep bad ones away from client money.

NOT EVERY HYPOTHESIS GRADUATESPARTH PAREKH

Rough path: idea → spec → backtest → walk-forward → paper → live. Most time is spent in the backtest loop. Most ideas exit there.

Walk-forward freezes training decisions before you score out-of-sample. If it only works when you cheat that freeze, it doesn't work.

Paper isn't a separate toy program. Same routing, same timeouts, same fan-out — only the fill is simulated. If it clears paper, it's already exercised the path live capital will use.

Eight strategies cleared every gate — intraday index options and overnight stock options. The useful number is how many didn't.

MORE NOTESPARTH PAREKH