Parts 1 to 6 covered reading and execution: what liquidity is, where it sits, how it gets taken, and how to frame a trade around it. This part is different in kind. Almost none of it is about the chart. It is arithmetic — the arithmetic that decides whether a method survives its own losing runs, and the small number of rules that keep a decision made calmly on Sunday from being unmade at 09:04 on Tuesday.
Everything is measured in R
R is the distance between your entry and your stop, expressed as the money you lose if the stop fills at the stop price. Everything else is quoted as a multiple of it. A target three times the stop distance is a 3R target. A day that ended down two stop distances is a −2R day.
The point of the convention is that it makes trades comparable. A twelve-point stop on an index and a forty-pip stop on a currency pair are the same size of decision if both were sized to the same risk. Without R you end up comparing pips across instruments where a pip means different amounts of money, which tells you nothing.
The sizing arithmetic that produces a constant R is the one from part 6:
Fix the risk amount first as a percentage of the account, then let the volume fall out of the stop distance. Sizing in the other order — picking a familiar lot size and accepting whatever stop distance the chart offers — means your R changes on every trade, and the arithmetic in the rest of this article stops applying to you.
One consequence is worth stating plainly, because it is where the convention breaks: a stop is an instruction, not a guarantee. It becomes a market order when touched and fills at the next available price. At session opens, around scheduled releases and in thin hours, that price can be materially worse than the stop level, and the realised loss can exceed the 1R you measured. Everything below assumes a loss of 1R; treat that as the optimistic case rather than the expected one.
Expectancy is the equation that decides the outcome
A method's expectancy is the average result per trade, in R:
where W is the win rate as a decimal and R is the average reward-to-risk ratio actually realised — not the ratio you aimed at.
Setting E to zero and solving gives the breakeven win rate for any ratio:
That single expression is the reason the reward-to-risk ratio matters at all, and it is worth reading as a table rather than a formula.
| Average realised ratio | Win rate needed to break even | Expectancy at a 40% win rate |
|---|---|---|
| 1R | 50.0% | −0.20R |
| 1.5R | 40.0% | 0.00R |
| 2R | 33.3% | +0.20R |
| 3R | 25.0% | +0.60R |
| 5R | 16.7% | +1.40R |
The third column is arithmetic under one stated assumption — a 40% win rate, chosen because it is the figure at which a 1.5R method exactly breaks even. It is an input to the calculation, not a claim about anyone's results.
Two things follow. The first is that a win rate on its own is meaningless: 40% is a losing method at 1R and a strongly positive one at 3R. The second matters more for the setups in this guide. A liquidity-based approach that waits for a defined invalidation and targets the opposite pocket tends to produce a ratio well above 1, which is what makes a win rate below half survivable — but only if the ratio you realise matches the one you planned. Closing winners early while letting losers run the full stop distance collapses the realised ratio toward 1 and quietly moves you into the first row of that table, whatever your entries looked like.
Losing runs are a property of the arithmetic
If a method wins 40% of the time, the chance of any given trade losing is 0.6. Assuming each trade's outcome is independent of the last, the chance of k losses in a row is 0.6k:
| Consecutive losses | Probability of that exact run | Roughly once every |
|---|---|---|
| 3 | 21.6% | 5 sequences |
| 5 | 7.8% | 13 sequences |
| 8 | 1.7% | 60 sequences |
| 10 | 0.6% | 165 sequences |
Read that table carefully, because it is easy to misread. It assumes independence and a fixed win rate. Real trading gives you neither: outcomes cluster because market conditions cluster, and your own execution degrades during a run of losses, which is the opposite of independence. The table is therefore a floor on how common long losing runs are, not an estimate of them. A run of eight is not evidence that a method has stopped working, and it is not evidence that it still works either. It is simply not informative on its own.
The practical consequence is the reason this section exists. If you size such that eight consecutive losses would breach a limit — your own or a funding programme's — then the arithmetic says you will eventually breach it, without anything having gone wrong with your reading of the chart.
Drawdown is not symmetric
A loss and the gain needed to undo it are not the same number, because the gain is calculated on a smaller balance. The relationship is exact:
| Drawdown | Gain needed to return to breakeven |
|---|---|
| 5% | 5.3% |
| 10% | 11.1% |
| 20% | 25.0% |
| 33% | 49.3% |
| 50% | 100.0% |
The curve is gentle and then it is not. Below roughly 10% the asymmetry is a rounding error. Past 30% it starts to dominate every other decision you make, because the size of the required recovery pushes people toward exactly the behaviour that caused the drawdown: bigger positions, looser criteria, more trades.
This is the argument for a risk-per-trade figure small enough that a long losing run stays inside the gentle part of the curve. At 1% per trade, the ten-loss run in the table above costs about 9.6% and needs about 10.6% to recover. At 3% per trade, the same run costs about 26% and needs about 36%. The chart reading was identical in both cases; only the sizing differed.
Correlation: three trades, one bet
Risk limits are usually written per trade, and that is where the arithmetic quietly fails. If you are long EUR/USD, long GBP/USD and short USD/CHF at 1% each, you do not have three positions at 1%. You have one position on a weaker dollar, at roughly 3%, split across three tickets.
Currency pairs sharing a leg tend to move together, and the setups in this guide make that worse rather than better: a dollar-driven liquidity take prints a similar-looking sweep on every dollar pair at the same moment, so the method itself will hand you three chart-perfect entries on what is economically a single trade.
- Cap total open risk, not just risk per trade. A limit on simultaneous exposure to any one currency, or any one driver, is the version of the rule that binds.
- Count a shared leg as the same position. If two setups would both lose on the same dollar move, they are one bet and should share one allocation of risk.
- Gold and indices are not exempt. They frequently share a risk driver with each other and with the dollar; they are simply less obvious about it.
What funded-account rules do to the arithmetic
Traders working under a funding programme have two limits imposed on top of their own, and the two behave very differently.
- The daily loss limit — a cap on how much the account may lose within one trading day, usually measured from the day's starting balance or equity, and usually reset at the provider's own daily rollover, which is often not midnight in your timezone.
- The maximum drawdown — a floor on the account, either static (fixed at a percentage of the initial balance) or trailing (following the account's high-water mark upward, so profits raise the floor and can never be given back below it).
The daily limit is the one that binds in practice, and the arithmetic is worth doing before rather than after. If the daily limit is 5% and you risk 1% per trade, three losses put you at 3% and a fourth at 4% — but a fourth loss that slips beyond its stop can breach 5% on a position you believed was sized to stay inside it. Sizing to the limit rather than to the limit minus a slippage allowance is what turns an ordinary losing morning into a failed account.
A trailing drawdown deserves its own note because it inverts a familiar habit. Under a trailing floor, an unrealised gain that you give back has moved the floor up and then eaten the distance to it. Rules written for a static floor — hold for the full target, add on strength — can be actively wrong under a trailing one. Read which variant applies before sizing anything.
Turning discipline into rules
Most of what gets filed under trading psychology is better handled as a rule written in advance, because a rule can be checked and a state of mind cannot. Each of the following is a common failure paired with the mechanical constraint that removes the decision from the moment.
- Trading to recover a loss. The constraint is a hard daily stop, in R, decided before the session and enforced by closing the platform. Note what the arithmetic above says about this: after a loss your account is smaller, so a fixed-percentage risk automatically reduces position size. Increasing it instead is the exact inversion of what the drawdown table requires.
- Entering because the move is leaving without you. The constraint is the confirmation set from part 6. If the confirmations are not present, the setup does not exist, however convincing the candle looks in real time. A missed move costs nothing; it is the one outcome with a guaranteed result of zero.
- Moving the stop away from price. The constraint is that the stop distance is decided before entry and is an input to the position size. Widening it after entry does not reduce the loss — it increases the risk on a position that was sized for the original distance, which means the trade is now larger than any rule you set.
- Trading outside the plan's conditions. The constraint is a written session window and a maximum number of trades per day. Both are checkable against a clock and a count; neither requires you to assess your own state.
None of this makes the discomfort go away. It makes the discomfort irrelevant to what happens next, which is the only part that can be engineered.
The journal, and what makes it measurable
A journal that records how a trade felt produces material you cannot act on. Record instead the fields that let you check a rule was followed, and the ones that let you compute the two numbers this article depends on — your realised win rate and your realised ratio.
- Date, instrument, direction, session — and whether the entry fell inside your written window.
- Which confirmations were present, listed individually, so a rule breach is visible rather than remembered.
- Planned R and realised R, separately. The gap between them is where most methods lose their edge, and it is invisible if you record only the outcome.
- Slippage on entry and exit, in points. This is what tells you whether your stop-distance assumption survives contact with your broker at the times you actually trade.
- Whether the trade broke a rule — a single yes/no field. Sorted by this column, a journal answers the only question worth asking after a losing month: was the method wrong, or was it not followed?
Thirty entries is not a sample from which to conclude anything about a method. It is, however, enough to see whether your realised ratio resembles your planned one, and that comparison is the one that feeds directly back into the expectancy equation at the top of this article.
What this part does not do
The arithmetic here is exact, and its inputs are not. Expectancy computed from your last thirty trades is an estimate with wide error bars, and it will move as conditions change. None of these equations can tell you whether a method has stopped working, because a losing run consistent with a working method and a losing run produced by a broken one look identical for far longer than most people's patience.
What the arithmetic can do is bound the damage while you find out: keep the risk per trade small enough that the drawdown table stays gentle, keep total exposure honest about correlation, and keep the daily stop mechanical. Those three constraints do not improve a method. They keep a workable one from being ended by an ordinary losing run — which, as the streak table shows, is not a rare event but a scheduled one.