Build a repeatable process for observation, scenarios, execution and review. A candle or pattern does not guarantee the future. Start with historical charts and a practice environment; all numbers and diagrams in this course are synthetic and educational.
What price action is—and is not
Price action reads price movement in relation to time, important zones and market structure. Open, high, low and close are its raw material. The aim is not certainty about the next candle, but a scenario whose entry, invalidation and risk are defined before the trade.
The same pattern does not mean the same thing in a trend, range or news release. Volume can supplement analysis, but OTC forex volume or one exchange’s volume is not necessarily total market volume. Do not attribute every wick to the definite intention of large participants; a chart alone reveals neither identity nor intent.
Read candles, not just their colour
The body is the distance between open and close; wicks show the interval’s high and low. A large body can describe a substantial move during that interval, but does not by itself guarantee continuation. A close near the high differs from a close near the middle.
OHLC does not specify the order of high and low. A single candle cannot tell you whether stop or target came first. If both lie within one test candle, use finer data or a predefined conservative assumption. Selecting the better outcome after viewing the chart corrupts the test.
Measure the candle: body, wick and closing location
Replace “strong candle” with three measurements. Range = high minus low; body = absolute close minus open; closing location = (close minus low) divided by range. With open 100, high 110, low 96 and close 107, range is 14, body is 7 and body/range is 50%. The close is about 78.6% up the range; upper wick is 3 and lower wick is 4. Record the ratio as undefined for a zero range.
Compare it with preceding candles. A range of 14 after several 3-unit candles is expansion; the same 14 after 25-unit ranges is smaller. Same-direction bodies with little overlap describe directional movement; heavy overlap and alternating closes describe rotation. Record these observations separately from entry decisions so that observation does not become a prediction.
Exercise: choose two candles with equal bodies of 4, but ranges of 5 and 12. Their body ratios are 80% and 33.3%. Before revealing the next candle, note where each closed relative to the previous zone. The output is a repeatable description, not an automatic purchase of the higher-ratio candle.
Market structure: trend, range and transition
A simple uptrend has rising significant highs and lows; a downtrend has falling ones. Swing importance depends on timeframe and selection rules. Specify in advance how many candles or what distance defines an important low, and do not change the rule during testing.
In a range, price repeatedly moves between upper and lower zones. Buying the middle can leave poor distances to resistance and stop. Transitions are often ambiguous; a minor break alone does not confirm a complete reversal. If classification is unclear, no trade is a recordable decision.
Confirmed swings and confirmation delay
For a practice definition, a swing high exceeds the highs of the two candles before and the two after it; reverse the rule for a swing low. In highs 98, 100, 105, 102 and 101, the third candle is a candidate high, confirmed only after the fifth candle closes. Specify equal-high treatment too: under this strict definition, equality rejects the candidate.
HH means higher high, HL higher low, LH lower high and LL lower low relative to the previous swing of the same type. “Low 90, high 100, low 95, high 106” is a simple rising structure. Trading below 95 challenges that sequence; calling a downtrend still needs an additional rule, such as a lower high and lower low. Small internal structure is not the same as higher-timeframe major structure.
Replay exercise: record candidate time and confirmation time in separate columns. An entry that depends on that swing cannot precede confirmation. Other definitions, such as minimum swing distance, are possible; results are comparable only if the swing definition stays fixed throughout the sample.
Treat support and resistance as zones
Mark areas of prior meaningful reactions: major highs and lows, range edges and role reversals are starting points. A line fitted perfectly to a wick after the outcome may be hindsight selection. Keep zone width consistent with volatility and timeframe.
Ask more than whether price touched the zone. Consider approach speed, candle size, news timing, closing location and distance to the next barrier. More touches do not necessarily strengthen a level. Zones can break, and resistance becoming support remains a testable scenario.
Volatility and ATR: put stop distance in context
ATR does not predict direction; it averages true range. A candle’s true range is the greatest of high minus low, absolute high minus previous close, and absolute low minus previous close. If the previous close is 100 and current high and low are 107 and 103, ordinary range is 4 but true range is 7: the gap is included.
With Wilder smoothing, new ATR = (previous ATR × (n−1) + new true range) ÷ n. For period 14, previous ATR 3 and true range 7, the update is about 3.29. Initialisation usually starts from an average of true ranges; check the platform’s implementation and data. 14 is a common convention, not a period guaranteed to be optimal everywhere.
Use it for comparison: a 0.5-unit stop in a market with ATR near 3 may lie within ordinary fluctuation. This does not establish an automatic ATR-multiple stop. Locate structural invalidation first, then consider volatility, costs and executable size. If necessary distance at minimum size exceeds the budget, reject the scenario rather than squeezing the stop to fit.
Assign clear roles to timeframes
Choose one timeframe for context, another for the setup and optionally another for execution. A teaching combination might use daily direction and zones, four-hour entry structure and hourly execution; it is not a prescription for every person and market.
Moving to a smaller timeframe does not erase larger context. A small bullish candle beneath major higher-timeframe resistance does not remove that barrier. If rules rely on closing prices, wait for the candle to close; an unfinished candle can change completely.
Multi-timeframe workshop: when to reject a small signal
Suppose daily highs and lows are rising, but a premarked resistance is at 110. A four-hour pullback reaches 100–102. An hourly trigger suggests entry 103 and structural stop 99. Distance to resistance is 7 and risk distance is 4, giving gross room of 1.75 times risk. If the example’s previously tested rule requires at least 2, reject the trade despite an attractive entry candle.
Save three simultaneous screenshots and label each timeframe’s role: context, zone and trigger. Switching timeframes after entry to justify holding a loss changes the process. Broker daily closes and chart time zones can differ; use consistent time definitions in replay and practice execution.
No-trade exercise: find a bullish small timeframe inside the middle of a higher-timeframe range. Before continuing the chart, specify the missing condition: reaching an edge, forming clear structure or gaining room. If it never occurs, keep the non-entry as a complete journal decision; a later price move is no reason to delete it.
Breakouts, pullbacks and failed breaks
Before trading, define a break by a price crossing, a close outside a zone or another condition. Immediate entry and waiting for a pullback differ: the first may fill worse and the second may never fill. A valid break need not return to your entry level.
On a retest, examine price behaviour and the relevant latest low or high. Renewed acceptance inside the range may invalidate the breakout. A wick beyond a level followed by a rapid return may describe a failed break, but cannot establish manipulation or the next direction with certainty.
Setup one: trend continuation after a pullback
This is a testable practice setup, not a proven profitable method. Context: confirmed rising swings without nearby resistance. Location: a return to the earlier breakout zone at 100–101. Sample trigger: after reaching the zone, a completed candle closes above the previous candle’s high. Hypothetical entry is 102, stop 99 and next barrier 108; risk distance is 3 and target distance 6, before costs.
Specify cancellation equally precisely: a close below 99 before entry, expiry after four execution-timeframe candles, or reaching the target before the order fills. Four candles is only a sample parameter to test. If large bearish candles pass through the zone, do not rename the event “a cheaper price” and turn it into a different trade.
Anticipate two outcomes. In the successful path, target 108 is reached for gross gain 6 per unit. In the unsuccessful path, an actual exit at 98.5 loses 3.5 per unit. Both belong to the same scenario. Assess process validity by whether conditions held before entry and exits followed the plan, then examine the performance of the full sample.
Setup two: trade the range edge, not its middle
Premark a hypothetical range from 90 to 100. Price approaches the floor and completes the defined reversal condition. Entry 92, stop 89 and target before the ceiling at 98 give risk 3 and target room 6. If the same trigger appears only at 96, the old stop risks 7 for target room of just 2. Identical candle shapes need not have identical location quality.
A range does not remain a range forever. A compressed approach to its edge, changed volatility or news may create a new problem. In this sample plan, a close below the predefined boundary invalidates the idea; “it must return to the middle” is not an exit rule. Buying every touch and buying after confirmation are separate methods whose records should not be mixed.
Setup three: failed breakout and hindsight traps
Suppose resistance is 100; price reaches 102 but closes at 99. This shows only a return below the level. A practice short scenario may add a retest of 100 from below and a completed bearish trigger. Hypothetical entry 99, stop 103 and target 91 give risk 4 and reward distance 8. Reverse the structure for a long after a failed downside break; actual execution depends on shorting availability and instrument terms.
If price instead drops straight to 94, chasing with stop 103 risks 9 for target room of only 3. A gain that occurred without your entry is not an account loss. If price returns above 103, the short scenario failed; raising the stop and calling it a “second trap” rewrites the definition after the outcome.
Treat “liquidity sweep” as an interpretation. Stops may cluster around highs and lows, but OHLC cannot prove their size or owners. For testing, record observable events: distance beyond the level, closing location, bars until retest and invalidation level. A compelling story does not replace a measurable variable.
Put candle patterns in context
A pin bar usually has a prominent wick and smaller body; engulfing refers to one body covering the previous body; an inside bar lies within the preceding candle’s range. Define exact ratios and coverage before testing, because the names have varying definitions.
Evaluate the pattern alongside zone, trend and available room. A pin bar in congestion is not the same problem as one at a premarked level. Require three separate conditions: explainable context, a defined trigger and sufficient target room relative to risk. Be ready not to trade if any is missing.
Define BOS, CHOCH, order blocks and FVG precisely
Some approaches use BOS for a structural break with the trend and CHOCH for the first break against it. Definitions vary between educators. Specify which swing breaks, whether a wick or close qualifies, and any confirmation delay. Breaking a small internal low is not necessarily breaking the major daily low; terminology should not conceal the difference.
One three-candle bullish FVG definition is that the third candle’s low exceeds the first candle’s high. A first high of 100 and third low of 102 mark the geometric interval 100–102. This does not mean no trades occurred there: the middle candle may have traded through it. A return and complete fill are not guaranteed.
An order block is sometimes defined as the last opposite candle before a move and break. Specify body versus full-range boundaries, validity after a revisit and deletion rules. A candle cannot establish that institutional orders remain there. Exercise: apply the same definition to failed examples. If you can select a zone only after seeing a large move, the rule is not yet operational.
From idea to order: a complete scenario
In the diagram, price breaks hypothetical resistance 100 and retests it. After the predefined confirmation, entry is 101, invalidation 98 and initial target 107. Risk per unit is 3 and reward distance 6, giving gross reward/risk of 2:1. These numbers teach calculation and are not a live market signal.
If the next barrier is 103, do not choose 107 merely to improve the paper ratio. If price reaches 104 before entry, the old stop and target no longer give the original ratio. Write order type, cancellation, expiry and partial-fill handling before submitting.
Order type, slippage and partial fills: chart versus execution
Test three entries separately: a market order after confirmation, a limit on a return, or a stop after crossing the trigger. Market orders may execute sooner but at a worse price; limits constrain price but may not fill; stops usually become market orders once triggered. “Price touched the level, so my entire order filled” is not a reliable assumption, especially with coarse data.
For planned entry 101 and stop 98, adverse entry slippage of 0.4 and exit slippage of 0.3 raise price risk from 3 to 3.7 per unit, before commission. A budget of 60 then permits at most 16 whole units before commission, not 20. Do not double-count spread: model actual bid/ask fills or add an equivalent cost to reference prices, using one consistent convention.
If only part fills, size protective and target orders for the actual open quantity and cancel the remainder under the expiry rule. Check this in practice: linked or OCO orders can have specific behaviour. Disconnection, rejected orders and market closure also need written responses; a line on a chart does not prove server acceptance.
Position size, costs and management
With hypothetical loss budget 60 and stop distance 3, linear size before costs is 20 units. At round-trip cost 0.20 per unit, risk per unit is 3.20 and theoretical size 18.75; if only whole units are allowed, round down to 18. Planned risk is 57.60 and net target profit in the example is 104.40.
This excludes gaps and additional slippage. A normal stop does not guarantee exit price. Breakeven moves, partial exits and trailing stops need predefined tested rules; moving a stop to entry may not even cover costs. Do not change invalidation after entry merely to avoid accepting a loss.
Trade management: calculate partial exits and breakeven
Keep initial R fixed after entry so statistics remain comparable. Exiting half at +1R and half at +3R gives +2R overall before costs, not +4R. Half at +1R and half at −1R gives zero gross. Partial exits change the distribution of outcomes; they are not always better or worse and should be compared with full exits on the same sample.
Price breakeven differs from monetary breakeven. Entry and exit at 100 with round-trip cost 0.2 per unit still lose 0.2. A trailing stop under new lows must respect their confirmation rule; future lows cannot be used early. A time stop, such as exit after a specified number of bars without progress, is also a strategy parameter that must be included in testing from the start.
Do not treat concurrent position budgets as independent. Two correlated trades with planned risk of 50 dollars each carry 100 dollars of combined planned risk in a shared scenario, with gaps potentially exceeding it. Three symbols are not necessarily three independent risks. Record total open risk, currency direction and shared industry alongside individual trade risk.
Backtest without seeing the future
Make entry, exit, timing, news and size rules measurable. Replay candles using only information available then. A swing identified with right-side candles is confirmed with delay; an unconfirmed historical signal cannot be traded earlier.
Reserve one period for design and a separate one for out-of-sample evaluation. Cover trends, ranges and high volatility. Model spreads, commissions, slippage, unfilled limits and events. Repeated parameter tuning until results look desirable risks overfitting. No fixed magic sample size proves profitability.
Backtest protocol: sampling through out-of-sample evaluation
Before starting, version the rules: symbol and data source, time zone, timeframe, swing definition, zone, trigger, entry, stop, target, costs, expiry and news filter. Separate development and evaluation by date. Once evaluation results influence rule changes, that period is no longer untouched; the next evaluation needs fresh data.
Record every qualifying opportunity, including unfilled orders and non-entries. If candle data touch both stop and target with unknown order, use finer data or a fixed conservative rule. For shares, excluding failed companies creates survivorship bias; split and dividend adjustments must be consistent. For futures, distinguish continuous charts from actual contract prices.
Run three sensitivity checks: higher costs, a plausible execution delay and parameters near the original choice. If results are positive only in one narrow setting, investigate why; do not simply select the best parameter after seeing all outcomes. Practice samples of 30 or 100 trades may teach recordkeeping, but neither proves a lasting edge. Market conditions and trade dependence affect estimate reliability.
Expectancy, drawdown and journaling
Define R as the initial planned loss per trade. If sample win rate is 40%, average win 2R and average loss 1R, gross expectancy is 0.2R. Average costs of 0.1R reduce net expectancy to 0.1R. Rates and averages estimate the past and can change in the future.
Record drawdown, losing streak length, rule adherence and cost sensitivity alongside net results. Journal date, symbol, timeframe, context, before/after images, entry, stop, target, size, costs and exit reason. Keep rejected and failed trades too; selecting only beautiful examples distorts the method.
Numerical report: profit factor, drawdown and recovery
A hypothetical 20-trade sample has 8 +2R wins and 12 −1R losses. Gross wins total 16R, absolute losses 12R, and the result is +4R, averaging +0.2R. At cost 0.1R per trade, total cost is 2R and net result +2R. Gross profit factor is 16/12, about 1.33. With those same costs, net wins total 15.2R and net losses 13.2R, giving net profit factor about 1.15.
Measure drawdown from the preceding equity peak, not just starting balance. The path 100, 105, 102, 96 and 108 has maximum decline of 9 from peak 105, about 8.57%. Recovering a 20% loss needs a 25% gain because 80 must return to 100. Recovery = drawdown/(1−drawdown), using decimals. Win/loss ordering changes the path even when the final total is identical.
Breakeven win rate also depends on costs. With gross win 2R, loss 1R and fixed cost 0.1R per trade, 2p − (1−p) − 0.1 = 0 gives about 36.7%, versus 33.3% without costs. This model applies only to those assumptions. Show cost, sample count and period beside performance; a number without that context cannot be assessed.
Practice plan and final checklist
First annotate historical candles and swings with a fixed rule. Second define and record one setup, such as breakout and retest. Third test unchanged rules on separate data. Fourth compare practice execution with realistic costs and constraints. Each stage aims at sound records and execution, not a promised profit number.
Before an order, answer: what regime is this? Was the zone premarked? Is the trigger complete? Where is the next barrier? What invalidates the idea? Does size account for costs and correlation? Have I considered news and overnight risk? If an answer is unclear, complete the scenario or do not trade.
Final exercise
Swing recognition: five highs are 98, 100, 105, 102 and 101. With two candles required on each side, when is high 105 usable?
Show answer
Only after candle five closes; at candle three the two right-hand candles did not exist. Trading with future confirmation introduces look-ahead bias.
True range: previous close 100, current high 107 and low 103. Why is the answer not 4?
Show answer
The maximum of 4, 7 and 3 is 7. High-to-low range does not capture the gap from the previous close. ATR averages ranges and does not specify the next direction.
Scenario rejection: entry 103, stop 99, next resistance 110. The practice rule requires minimum gross reward/risk of 2:1. What follows?
Show answer
Target room is 7 and risk 4, a ratio of 1.75, so the rule fails. Do not raise the target without justification or squeeze the stop merely to improve the ratio. Reject this version of the scenario.
Partial exit: half closes at +1R and half at −1R. With total costs of 0.1R, what is the net result?
Show answer
Half of +1 plus half of −1 is zero gross; after costs, −0.1R. A positive first exit does not make the whole trade profitable. R remains based on the full position’s initial risk.
Drawdown: equity drops from peak 105 to 96. What percentage is lost, and why not divide by 100?
Show answer
Use the preceding peak: (105−96)/105, about 8.57%. Starting balance is not the denominator for this drawdown. State whether the curve includes open profit/loss or only closed trades, and keep that convention consistent.
What happens if the planned 101 entry fills at 104 while the stop stays 98 and the target stays 107?
Show answer
Risk becomes 6 and potential reward 3 per unit. The gross ratio drops from 2:1 to 0.5:1; recalculate the scenario and size.