Arithmetic fills the last part of the page. The number that decides whether any rule can make money at all is expectancy, the average result per trade after costs. For most fast-trading approaches on a CFD, costs are big enough to settle that question before anyone even tests the rule.
Key points
- Every strategy is a bet on a type of market. Trend following needs moves that last. Mean reversion and breakout each need a range, but breakout needs one that’s about to widen. Carry needs calm, and hedging needs an exposure you already have.
- Each strategy’s failure is another strategy’s market. Trend following loses in ranges and mean reversion loses in trends. A breakout fails when the price slips back into its old range. Carry fails in weeks when investors run from risk.
- Expectancy = (win rate × average win) − (loss rate × average loss) − costs per trade. Even if a strategy wins most of the time, it can still have a negative expectancy.
- Costs per trade come from the spread, the gap between the buy and sell price, and any commission. The less time you hold a trade, the more of its expected move those costs eat up.
- A published win rate tells you nothing about expectancy unless you also have the average win, the average loss and the costs.
The five strategy families
The five strategy families that retail CFD traders use differ in the market condition each one depends on and in how each one fails when that condition ends.
Trend following
Trend following opens a trade in the direction of a recent move, expecting it to continue, and closes it when the move reverses by a set amount. It needs trends that last. It fails in sideways markets, entering late, exiting late and paying the spread on every false start. Its usual record is a low win rate with the odd large win. That’s hard to stick with, and impossible to judge from a few trades (technical analysis).
Mean reversion
Mean reversion opens a trade against a recent move and closes it when the price returns towards an average or a range. The strategy expects that return to happen. It needs a market moving sideways in a range and fails in trends, where the price keeps moving further against each new entry. Its usual record is a high win rate with the odd large loss. That’s the opposite profile, and the more dangerous one on a leveraged account, where you control a trade much bigger than your deposit. The large loss arrives after a run of wins has built up your confidence and your trade size.
Breakout
Breakout opens a trade when the price pushes past a level that has held before, expecting the move to keep going. It needs a market that’s about to widen its range. False breakouts, where the price crosses the level and comes back, make it fail. On a CFD that failure is expensive, because you get in at the busiest moment, when the spread is widest and slippage is largest (how a trade works). Slippage is the difference between the price you asked for and the price you got.
Carry
With carry, you hold a currency trade that earns the difference between two interest rates, so every night you collect the swap, an overnight interest payment. Carry needs a calm market, one where the exchange rate moves by less than the interest difference earns (forex trading). It fails in risk-off periods. Investors dump riskier assets then, and high-interest currencies fall together, fast. A few days of that can wipe out what the interest earned over months. Even before you count the risk, the mark-up your provider adds to the swap also cuts the income.
Hedging
Hedging is when you open a CFD trade opposite to something you already hold. You do it so a loss on one is offset by a gain on the other for a set period. One example is a short index CFD held against a portfolio of shares. A short index CFD gains if the index falls. Hedging needs an existing exposure and a reason to keep it. The strategy fails when the hedge is imperfect, because an index doesn’t move exactly like your portfolio. It also fails when you leave the hedge on after the reason has passed. Then it becomes a bet of its own that pays financing every night (what are CFDs). Holding a buy and a sell on the same currency pair is a narrower meaning of the word hedging (hedging in forex).
Conditions needed and failure modes
Conditions and failure modes line up across the five families. Set side by side, they show a pattern where each strategy’s failure mode is another strategy’s working condition.
| Strategy | Market it needs | How it fails | Usual shape of results |
|---|---|---|---|
| Trend following | Moves that keep going one way | Ranges: repeated false starts, each paying the spread | Low win rate, a few large wins |
| Mean reversion | Prices moving back and forth in a range | Trends: each entry against the move adds to the loss | High win rate, rare large losses |
| Breakout | A quiet, narrow range that’s about to widen | False breakouts at the widest spread of the day | Middling win rate, high costs |
| Carry | Calm markets, steady interest-rate differences | Risk-off weeks: high-interest currencies fall together | Steady small income, sudden large loss |
| Hedging | An existing exposure to protect | An imperfect hedge, or a hedge left on too long | Smaller swings in value, plus a financing cost |
Expectancy arithmetic
Worked arithmetic: the numbers follow from the formula given in this section.
Expectancy tells you what a strategy earns or loses on average per trade. You get it from four numbers: expectancy = (win rate × average win) − (loss rate × average loss) − average cost per trade. The two examples below use made-up numbers.
Take a strategy that wins 60 % of the time, with an average win of 100 and an average loss of 150. Before costs, expectancy is 0.6 × 100 − 0.4 × 150 = 60 − 60 = 0. Add a cost of 15 per trade in spread and commission, and the expectancy becomes −15. The strategy loses on average, whatever its win rate.
Now take one that wins only 35 % of the time, with an average win of 300 and an average loss of 100. Before costs the sum is 0.35 × 300 − 0.65 × 100 = 105 − 65 = 40. After the same cost, that leaves 25.
A win rate means nothing on its own, and the arithmetic shows why. A mean-reversion profile, a high win rate paired with rare large losses, can hide a negative expectancy for months. The number of trades matters too, because you can only estimate expectancy from a large sample. Fifty trades give an estimate too noisy to tell a small positive expectancy from a small negative one (risk management).
Why costs decide whether frequent trading can work
Worked arithmetic: the numbers follow from the formula given in this section.
Costs decide whether a frequent-trading approach can work. Each trade costs roughly the same, the spread plus any commission, however long you hold, but the move you can expect shrinks as the holding time gets shorter.
Say a major currency pair has a spread of one pip, the standard unit of price movement in forex, and a typical hourly range of ten pips. A day’s range might then be roughly 50 to 100 pips. The price may swing across those ten pips within an hour, yet a trade held for that hour usually ends only a few pips from where it started. Set that small move against a cost of one pip just to get in. The strategy needs a big edge just to pay for itself. Held for a week, the same pair usually moves a hundred pips or more against the same one-pip spread, and the cost barely matters.
Suppose a rule trades twenty times a day using one standard lot, the usual trade size in forex, with each pip worth 10. The cost comes to twenty spreads a day: 20 × 1 pip × 10 = 200. Before any trade has gone wrong, that’s about 5 % of the 3,663 of margin the same trade needs in the forex trading example, paid every day (costs and fees). Margin is the deposit your provider holds while a trade is open.
The arithmetic doesn’t support holding trades longer, since longer holds bring financing and gap risk, the chance the price jumps past your stop while the market is shut. The arithmetic means you have to pick a strategy’s timeframe knowing the cost per trade. The strategies advertised with the most trades are the ones with the least room to work (trading styles).
What a strategy claim usually leaves out
A strategy claim in an advert or on a forum usually leaves out the five numbers you need to judge it.
- The average loss. Without it, the win rate means nothing.
- The costs per trade, including the spread at the time of day the strategy trades and the financing on trades held overnight.
- The sample size, because a record of thirty trades fits almost any expectancy.
- The market conditions of the test period. If a trend-following record comes from a year of strong trends, it says nothing about a sideways year.
- The worst losing stretch. Whether a leveraged account survives long enough to collect the expectancy comes down to the maximum drawdown, the biggest fall from a peak in the account (risk management).
Ask the same five numbers of a paid forex signal and of a trading robot sold as an expert advisor.
Questions about trading strategies
Questions about trading strategies mostly ask what a strategy is, which one earns the most money and whether software can run it.
What is a trading strategy?
A trading strategy is a set of rules that says when to open a trade, which way to open it, how big to make it and when to close it. Those rules are fixed before the trade, so you can test them and count the results. A strategy without written rules can’t be tested at all.
Which trading strategy is the most profitable?
No trading strategy has been shown to be the most profitable. Each family performs well in one type of market but fails in another, and nobody knows in advance which type will come next. The outcome depends on expectancy after costs, and you can only measure that after many trades.
What is a trend following strategy?
A trend following strategy opens trades in the direction of a recent move and closes them when the move reverses by a set amount. It needs trends that last and loses in sideways markets. Its usual record is many small losses and a few large wins (trend following).
What is a mean reversion strategy?
A mean reversion strategy opens trades against a recent move, expecting the price to return towards its average. A range trading strategy is one example of this. It often succeeds in sideways markets but suffers large losses when a trend begins (mean reversion).
Is scalping a trading strategy?
Scalping is a trading style because it defines how long you hold a trade, not when to open one. A scalper still needs a strategy to pick entries, and that could be a breakout or a range rule. The trading styles page covers scalping, day trading and swing trading.
Can a trading strategy be automated?
A trading strategy with exact rules can be automated, and MetaTrader calls the program that automates it an expert advisor. The automation follows those rules without judgement, whether during news events or across price gaps. It doesn’t alter the strategy’s expectancy (expert advisors).
Risk warning
No strategy changes what leverage does to a losing streak
No strategy changes what leverage does to a losing streak. Runs of losses are built into the results of every family on this page. Without leverage such a run would only be a setback, but on a leveraged account it ends in a close-out. In a close-out, your provider closes your trades because your account has run too low. In 2018 ESMA reported national regulators’ findings: between 74 % and 89 % of retail accounts lost money, across clients using every strategy described here (press release, 27 March 2018).
Sources
The original documents this page cites, in the order they first appear.