Research method

Backtests are evidence, not promises

TrendRadar uses backtests to compare transparent rules over the available history. The result is a way to reject weak ideas and understand trade-offs, not a forecast of the next trade.

What the table measures

A strategy row reports the number of completed cycles, average and median return, win rate, peak gain, drawdown, and an annualised estimate where meaningful. Median return and sample size are often more informative than a high average created by a small number of exceptional winners.

How to compare strategies

  • Prefer repeated positive behaviour across a substantial sample to an impressive result based on a few names.
  • Read return together with drawdown, holding period, liquidity, and the number of simultaneously active trades.
  • Compare modest threshold changes. A result that vanishes after a small change may be over-fitted.
  • Give special attention to recent and out-of-sample periods because market structure can change.

What is not modelled completely

Historical rule tests cannot fully model bid-ask spreads, slippage, market impact, tax, trading halts, changing constituents, borrow availability, personal execution, or future company news. A strategy with attractive cycle statistics still needs portfolio sizing, diversification, and independent judgement.