Finding a Profitable Edge

In plain English

An 'edge' just means a reason you'll make money over time, not by luck. The key number is expectancy — your average profit per trade after wins, losses, and costs. A high win rate can still lose money if the rare losses are huge. This lesson shows how to tell a real edge from a backtest that just got lucky.

A profitable strategy is not a clever setup — it is a positive expectancy that survives costs and is repeatable. Most retail "edges" are noise dressed up as signal, or real patterns that vanish once STT and slippage are counted. This lesson is the framework for telling a genuine edge from a flattering backtest.

Expectancy is the only thing that matters

Expectancy is the average profit you expect per trade: (win rate × average win) − (loss rate × average loss). A strategy can win 90% of the time and still lose money if the 10% of losses are large enough — the classic option-seller profile. Another can win just 35% of the time and be highly profitable if winners dwarf losers. Stop asking "what is the win rate" and start asking "what is the expectancy, net of costs". Win rate alone tells you nothing.

Expectancy = (Win% × Avg Win) − (Loss% × Avg Loss). Positive and large enough to clear costs = an edge. Everything else is gambling with extra steps.

Win-rate vs payoff: the trade-off

There are two honest ways to make money: win often for small amounts (high win rate, small payoff — premium selling) or win rarely for large amounts (low win rate, large payoff — trend buying). Both work. What does not work is wanting both at once: a high win rate AND a high payoff with no tail risk does not exist in liquid markets, because someone on the other side would take the free money. If a strategy seems to offer both, you have not yet found where the risk is hiding.

Costs decide who survives

A thin per-trade edge gets eaten alive by frequency. Every NIFTY/BANKNIFTY trade pays STT, exchange charges, GST, stamp duty and brokerage, plus the invisible cost of slippage and the bid-ask spread. A setup that looks profitable on raw backtested prices can flip net-negative across hundreds of trades once costs are subtracted. Always backtest net of realistic costs — the Cost Calculator exists to quantify exactly this bite — and prefer fewer, higher-conviction trades over churning a tiny edge.

Run any candidate strategy through the Cost & Tax Calculator at your real trade frequency. If the edge does not survive costs, it is not an edge.

Overfitting — the backtest that lies

Test enough rules on enough history and some will look brilliant by pure chance. That is overfitting: a strategy tuned to the noise of the past, not the structure of the market. Defences: keep rules few and simple (every extra parameter is a chance to fit noise), test on data you did not optimise on (out-of-sample), demand an economic reason the edge should exist (why does someone pay you?), and be suspicious of any equity curve that is too smooth or any result that needed many tweaks to look good.

If a strategy only works with very specific parameters and breaks when you nudge them, it is fitted to history, not to the market. A real edge is robust to small changes.

The validation checklist

Before trusting a strategy with real capital, demand: (1) a clear economic reason the edge exists — who is paying you and why; (2) positive expectancy net of realistic costs; (3) a max drawdown inside your tolerance; (4) results that hold out-of-sample, not just on the data you tuned; (5) enough trades to be statistically meaningful, not three lucky wins. Anything that fails these is a hypothesis, not an edge — paper-trade it before it touches your account.

Key takeaway: an edge is positive expectancy that survives costs, has an economic reason to exist, holds out-of-sample, and keeps drawdown inside your budget. Validate ruthlessly before risking capital.

Validate ideas on the backtester