Most don't. That is the point.
Idea generation
Macro themes, market structure, flows and recurring behaviours we observe on the desk.
Hypothesis & data
Every idea is written as a testable rule — entry, exit and invalidation — before any data is touched.
Backtesting
Realistic costs, slippage, out-of-sample and walk-forward testing. Robustness, not the best-looking curve.
Incubation
Paper trading, then small live size, to confirm the edge survives real execution.
Capital allocation
Strategies earn size gradually against predefined drawdown and volatility limits.
Monitoring & retirement
Live behaviour is compared with expectations. Strategies that decay are cut or reworked.
Three ways an idea dies.
Overfitting
A rule tuned to past data explains history, not the future.
Cost sensitivity
Many edges disappear once spreads, commissions and slippage are counted.
Regime dependence
Some ideas only work in one environment. We need to know when they stop.
Three approaches. One standard.
Macro / discretionary
Positions built around macro themes and policy, with rules for sizing and invalidation.
Systematic intraday
Rule-based strategies executed within the trading day, with strict cost control.
Event-driven
Structured approaches to scheduled data releases and market-moving events.