What is a trading strategy?

A clear definition

A trading strategy is a complete, repeatable plan for participating in a market: it specifies what you trade, when you enter, when you exit, and how much you put at risk, all justified by a reason you expect it to be profitable. The word "complete" matters. "Buy when RSI is oversold" is an entry idea, not a strategy, because it never says when to sell, how much to buy, on which market, or why the idea should pay off. A strategy turns a loose intuition into a specification precise enough that you could hand it to someone else — or a computer — and have them follow it identically.

The justification — the hypothesis about why the strategy should work — is the part beginners skip and professionals dwell on. Without a plausible reason for an edge, even a profitable-looking backtest is probably luck or curve-fitting in disguise.

Strategy vs algorithm vs bot

These three words describe three different layers of the same thing, and confusing them causes a lot of muddled thinking. Pin them down:

• Strategy — the idea and the plan. "Ride strong trends and cut losers quickly."
• Algorithm — that plan written as exact, testable rules a machine can follow. See what is a trading algorithm.
• Bot — the running program that executes the algorithm live against an exchange API. See what is an AI trading bot and what is a trading bot.

You design a strategy, encode it as an algorithm, and a bot runs it. The same underlying strategy can be expressed by many algorithms and run by many bots — but a bad strategy ruins all of them, no matter how clever the code. For the broader picture, see what is algorithmic trading.

The building blocks of a strategy

Strip any complete strategy down and you find the same components. Get all of them on paper before you write a line of code:

1. Edge hypothesis — the reason this should make money. "Markets that have trended tend to keep trending for a while." Without this, you are gambling.
2. Market and timeframe — what you trade (a stock, a crypto pair, an index) and on what bar size (1-minute, hourly, daily). The same idea behaves completely differently on different timeframes.
3. Entry rule — the precise condition that opens a position.
4. Exit rule — when to close, usually a stop-loss to cap the loss and a target or trailing rule to take the gain.
5. Position sizing — how much to risk per trade, so a single loss cannot sink you. Use a tool like the position sizing calculator and read risk management and position sizing.

Miss any one and you have an idea, not a runnable strategy. The exit and sizing rules are where most beginners are weakest — they obsess over entries and ignore the parts that actually control risk.

The main types of strategy

Most strategies fall into a handful of families, each betting on a different market behaviour:

• Trend-following — bets that a move in progress continues. A moving-average crossover or a trend-following bot are classic examples. Wins big in sustained trends, bleeds in choppy markets.
• Mean-reversion — bets that price stretched far from an average snaps back. See mean-reversion strategy and the RSI indicator. Wins in range-bound markets, gets steamrolled by strong trends.
• Breakout — bets that price escaping a defined range keeps going in the breakout direction. Profits from new moves, suffers from false breakouts.
• Arbitrage — exploits price differences between related instruments or venues. Edges are small, fleeting, and usually demand speed and infrastructure retail lacks.

None of these is universally best. Each is a bet on a market condition, so the right strategy depends on the market, the timeframe, and where you genuinely believe an inefficiency exists. Browse more in trading bot strategies and automated trading strategies.

Why a strategy needs a real edge — and most don't

Here is the truth that separates serious thinking from wishful thinking: a strategy only makes money over time if it has a genuine, repeatable edge — a reason its expected value is positive after every cost. The market is a fierce competition, and after fees and slippage it is roughly zero-sum, so for you to win consistently, someone else has to consistently be on the other side of a measurable inefficiency.

Most strategies do not have this. Many that look profitable in a backtest are simply overfitted to past noise — they describe history, not the future. A strategy without a defensible edge hypothesis is a coin flip dressed up in indicators. The discipline is to be sceptical of your own ideas and demand evidence before believing them.

How to evaluate a strategy honestly

Evaluation is where good intentions meet hard data. The honest process:

1. Backtest with realistic costs on the backtester — always include fees, spread and slippage, or the result is fiction.
2. Reserve out-of-sample data the strategy never saw during design, and check it still holds up there.
3. Forward test and paper trade on fresh, live data before risking money — see backtesting vs forward testing and paper trading explained.
4. Read the right metrics together — return alongside drawdown and Sharpe, never in isolation. See drawdown and Sharpe explained and the full checklist in how to evaluate a trading strategy.

If a strategy survives all of that, you have something worth a small, careful live allocation. If it does not — and most do not — you have saved yourself real money by killing it on paper. That is a win, not a failure. For where strategies most often go wrong, read why most trading bots fail.

Frequently asked questions