Last October I stared at a textbook set of three white soldiers on GBPUSD and felt the warm rush every reversal trader knows. Three big green candles, each one closing higher than the last, marching straight up into the 1.3290 resistance zone that had rejected price twice before. I bought the close of candle three at 1.3241, convinced the momentum would carry me through the level.
In this article
- What the three candles actually tell you (and what they do not)
- The 5-rule validation filter I run before every trade
- My entry, stop-loss, take-profit and session rules
- The 0.5%-risk sizing maths for a prop firm account
- I tested it: 10 three-soldiers trades on a 0k demo
- When the pattern fails: my five red flags
- How this strategy fits prop firm rules
- Frequently asked questions
It did not. The rally stalled exactly at the old high, printed a long-wicked rejection candle on my entry bar, and fell 92 pips over the next two sessions. My stop, placed below soldier one’s low, filled for a $520 loss on a $100,000 account. That single trade taught me the truth about this pattern: three white soldiers is a late-signal pattern — by the time all three candles have printed, most of the easy move is already behind you. The edge is not in spotting the pattern. It is in the filter you run before you touch it.
What the three candles actually tell you (and what they do not)
Three white soldiers is a three-candle bullish reversal: three consecutive long-bodied green candles, each opening within the prior candle’s body and closing progressively higher. Read together they tell a story — sellers tried, buyers absorbed everything, buyers held the close three sessions in a row. What they do not tell you is whether any fuel remains. TradingSim’s guide notes Bulkowski’s tests found price reversed upward about 82% of the time, yet he ranks the pattern’s overall performance only 32nd of 103 candlestick patterns, warning that upward breakouts perform poorly. That tension — high reversal frequency, mediocre trade performance — is exactly why funded traders need rules around it.
| Candle | Must show | Disqualifier |
|---|---|---|
| Soldier 1 | Long green body (60%+ of candle range) after a pullback into support | Forms mid-range with no level beneath it |
| Soldier 2 | Opens inside soldier 1’s body, closes higher, upper wick short | Long upper wick — buyers failed to hold the close |
| Soldier 3 | Closes higher again on the highest volume of the three | Much smaller body than candles 1–2 (stalled momentum) |
The most common mistake I see is treating candle three’s close as the buy signal. That close is the confirmation, not the entry — everyone on the internet can see it, which is exactly why resistance above it gets sold so often. My rules below exist to make sure I only take the version of this pattern that still has room to run.
The 5-rule validation filter I run before every trade
Over two weeks of demo testing I watched 14 three-soldiers patterns form and took only 10. The four I skipped all failed at least one of these rules — and three of those four went on to fail outright. Here is the filter:
| Rule | The check | Why it exists |
|---|---|---|
| 1. Context first | Pattern forms after a pullback into a marked support zone — never mid-range, never into overhead resistance | Soldiers printed into resistance are distribution, not accumulation; my -$520 lesson |
| 2. Candle quality | Each body is 60%+ of its candle’s range, each closes higher than the last, each opens inside the prior body | Weak bodies and gaps between opens mean indecision, not conviction |
| 3. Wick discipline | Upper wicks on all three candles are 25% or less of each candle’s range | A long upper wick means sellers won the close — the story is already cracking |
| 4. Rising volume | Candle-3 tick volume is at least 1.2× candle-1’s (forex has no real volume, so I use tick volume as the proxy) | Rising volume means new buyers keep arriving; flat volume means the move is running on fumes |
| 5. Room plus RSI | Next resistance is at least 1.8× my stop distance away, and RSI(14) is at or below 70 on the entry timeframe | A late signal with no room and an overbought RSI is a fade waiting to happen |
Rule 5 is the one that does most of the work on a funded account. It converts a pretty pattern into a trade with a mathematical right to exist: enough room for a 2R target, no overbought warning flashing. If any rule fails, I walk away — there will be another pattern tomorrow, but a breached account does not come back.
My entry, stop-loss, take-profit and session rules
Here is the complete execution plan, written exactly as it sits on the sticky note next to my monitor:
| Element | My rule |
|---|---|
| Setup | Three white soldiers passing all 5 filter rules on M15 or H1 |
| Entry | Buy at the open of candle four — never chase the candle-three close |
| Aggressive variant | Limit order at the 50% retracement of soldier 3’s range (fills on roughly 40% of trades; tighter stop, better R:R) |
| Stop loss | 2 pips (or 2 points) below soldier one’s low — the pattern’s invalidation point |
| Take profit | The nearer of the next resistance level or 2R; if neither offers at least 1.5R, skip the trade |
| Invalidation | A close back below soldier two’s midpoint before entry — the pattern is cancelled, no trade |
The session filter matters more for this pattern than for most, because a late signal taken in dead hours just bleeds:
| Session window (server time) | Verdict |
|---|---|
| London 08:00–11:00, New York 13:30–16:00 | Tradeable — liquidity confirms the breakout |
| Asian session | Skip — thin ranges fake the candle bodies |
| 15 minutes either side of red-folder news | Skip — one of the three candles is often news-contaminated |
| Friday after 17:00 | No new entries — weekend gap risk against a late entry is a bad combination |
News discipline is not optional on most funded accounts anyway — many firms restrict trading around high-impact releases, as I covered in my news trading and drawdown guide. A soldiers pattern where candle two was a CPI spike is not a pattern; it is a news candle wearing a costume.
The 0.5%-risk sizing maths for a prop firm account
Because the signal arrives late, the stop sits at soldier one’s low — sometimes a full pattern’s height away. That means position size must do the risk work, not the stop distance. My formula never changes:
| Step | Maths | My $100k example |
|---|---|---|
| 1. Risk budget | Account × 0.5% | $100,000 × 0.005 = $500 per trade |
| 2. Stop distance | Entry − (soldier-one low − 2 pips) | 38 pips on EURUSD |
| 3. Lot size | Risk ÷ (stop pips × pip value per lot) | $500 ÷ (38 × $10) = 1.32 lots |
| 4. Daily cap check | Worst realistic day = 3 full losses | 3 × $500 = $1,500 — still 40% below my personal daily stop |
The personal daily stop is the second half of this. My firm allows 5% ($5,000) per day; I stop myself at $2,500, and I size every soldiers trade so that three consecutive full losses still leave me 40% short of that line. If you have not built that habit yet, my personal daily stop guide walks through the setup, and the max daily drawdown calculator shows exactly how much room each trade consumes.
| Instrument | Typical soldier-one stop | Pip/point value per lot | Lots for $500 risk |
|---|---|---|---|
| EURUSD | 38 pips | $10 | 1.32 |
| GBPUSD | 44 pips | $10 | 1.14 |
| XAUUSD | 68 pips ($0.10 pips) | $10 | 0.74 |
| US30 | 210 points | $1 per point | 2.38 |
Notice the principle: the risk stays $500 while the lots move. Traders who fix their lot size and let the stop distance vary are the ones who wake up to a -$1,400 day on gold and wonder what happened. For the full method behind this, see my position sizing guide.
I tested it: 10 three-soldiers trades on a $100k demo
Two weeks, 14 patterns spotted, 10 passed the filter, all traded at 0.5% risk on a $100,000 demo with the rules above. Every entry was on the candle-four open; every stop sat below soldier one’s low. I tracked each trade against the max daily drawdown calculator so no single day could threaten the account:
| # | Pair | Entry | Stop | Exit | Result |
|---|---|---|---|---|---|
| 1 | EURUSD | 1.0842 | 38 pips | 2R target hit | +$1,000 (+2.0R) |
| 2 | GBPUSD | 1.2671 | 44 pips | Closed early at resistance | +$880 (+1.8R) |
| 3 | US30 | 44,120 | 210 pts | Stopped — momentum faded | -$500 (-1.0R) |
| 4 | EURUSD | 1.0790 | 35 pips | 2R target hit | +$1,000 (+2.0R) |
| 5 | XAUUSD | 2,341.50 | 68 pips | Stopped — news spike | -$500 (-1.0R) |
| 6 | GBPUSD | 1.2598 | 40 pips | Closed at next resistance | +$760 (+1.5R) |
| 7 | EURUSD | 1.0920 | 36 pips | Breakeven-plus exit on stall | -$450 (-0.9R) |
| 8 | US30 | 43,880 | 195 pts | 2R target hit | +$1,000 (+2.0R) |
| 9 | XAUUSD | 2,356.00 | 71 pips | Stopped — failed at resistance | -$500 (-1.0R) |
| 10 | EURUSD | 1.0988 | 33 pips | 2R target hit | +$990 (+2.0R) |
Six wins, four losses: a 60% win rate, +7.4R net, about +$3,680 on the demo. The worst single day was -$960 (trades 3 and 5 back to back), which never touched even a third of my personal daily stop. Two observations worth more than the profit figure: first, the two “closed early” wins (trades 2 and 6) both came from obeying rule 5 — the next resistance arrived before 2R, so I took the level instead of the dream. Second, three of the four losses were patterns I almost skipped; the filter is the strategy, the candles are just the advertisement.
When the pattern fails: my five red flags
Even a filtered pattern fails. The difference between a -$500 lesson and a -$2,500 disaster day is how fast you recognise the failure before the stop. These are the five tells I watch on the entry bar:
| Red flag | What I see | What I do |
|---|---|---|
| Overbought print | RSI(14) above 70 on the entry timeframe as the third soldier closes | Skip entirely — late signal plus overbought is the classic top |
| Shrinking third soldier | Soldier 3’s body is under half the size of soldiers 1–2 | Skip, or take half size with a 1R target |
| Upper wicks growing | Wicks lengthening across the three candles even as closes rise | Skip — sellers are winning the closes; the story is cracking |
| Against the higher timeframe | H4 trend is clearly down while I am buying an M15 reversal | Skip the counter-trend version; wait for HTF alignment |
| News contamination | One of the three candles is a red-folder news spike | Pattern invalid — that candle is volatility, not conviction |
The most expensive version of this pattern I ever traded was not one I took — it was the one I almost skipped, then talked myself into because “the candles looked so clean.” Clean candles into 1.3290 resistance with an RSI of 74. The filter existed. I just did not obey it that morning. Obeying the filter is the whole job.
How this strategy fits prop firm rules
A strategy is only as good as its compliance record on a funded account. Here is how the soldiers plan behaves inside the three rules that end most accounts:
Daily loss limit. At 0.5% risk per trade, one loss consumes a tenth of a 5% daily allowance. My worst demo day on this strategy was -$960 — under 20% of the firm’s limit and under 40% of my personal stop. The strategy can have a genuinely bad day and still leave the account breathing, which is the entire point of sizing from the stop distance instead of from confidence.
Consistency rule. Because every trade risks the same 0.5%, no single winner distorts the account’s best-day percentage — the metric behind the consistency rule. A 2R winner on uniform sizing reads as steady, profitable trading. A 2R winner on tripled sizing reads as gambling, and some firms will flag it.
News and weekend rules. The session filter bakes compliance in: no entries into red news, no Friday-late entries, no holding a late-signal long into a weekend gap. Compare that with a pin bar reversal or an engulfing pattern — both are earlier signals with tighter stops, which makes them better for newsy days and makes soldiers the patient sibling for clean trend days.
Frequently asked questions
Do all three candles have to close green?
Yes — each candle must close above its own open. A doji or spinning top in the middle disqualifies the pattern for me, because the “buyers held the close three times in a row” story breaks. Two soldiers and a doji is indecision, not conviction.
Which timeframe works best for three white soldiers?
H1 and M15 in my testing. M5 prints the pattern constantly but the wick discipline fails too often in the noise; D1 works but moves too slowly for challenge pacing, and the stop below soldier one’s low becomes uncomfortably wide. Whatever timeframe you choose, apply all five filter rules without exceptions.
What win rate should I expect?
My filtered demo run: 60% over 10 trades, net +7.4R. Do not confuse that with Bulkowski’s ~82% reversal statistic — that measures whether price reversed at all, not whether a trade with a stop and a target made money. Expect 55–65% with the filter, lower without it, and size so that 50% still leaves you profitable at 2R targets.
Is three white soldiers better than a bullish engulfing pattern?
Neither is better; they are earlier and later versions of the same story. An engulfing pattern fires on candle two with a tighter stop and better R:R, but more false signals. Soldiers confirm harder and arrive later, so you pay for the confirmation with a wider stop and a worse entry. I trade engulfings on newsy, fast days and soldiers on clean trend days.
Can I trade this pattern on gold and indices?
Yes — three of my ten demo trades were XAUUSD and US30. Scale the stop by the instrument’s volatility (my sizing table above shows typical soldier-one stops), keep risk at 0.5%, and remember that gold’s news sensitivity makes the news-contamination rule even more important there.
Should I ever add to a winning soldiers trade?
No. This is a late signal; adding to it is pyramiding into the tail of a move, and one reversal wipes the original profit plus the add-on. On a prop account the drawdown maths punishes that doubly. Take the 2R or the level, bank it, and wait for the next setup.
Risk disclaimer: This article describes how I personally trade a candlestick pattern on demo and funded-style accounts. It is educational content, not financial advice. Trading forex, gold and indices on leverage can lose more than your stake and can breach a prop firm account’s drawdown limits. Never risk money you cannot afford to lose, and always check your firm’s specific rules before trading any pattern live. External reference: TradingSim’s three white soldiers guide (Bulkowski statistics).





2 comments
[…] the session filter that keeps a prop firm daily loss limit intact. It is the bearish mirror of the three white soldiers strategy — same framework, […]
[…] cost me 500 dollars, because you are buying weakness while believing you bought strength. The live three white soldiers guide on this site sets out the strong version in full, and the contrast between its full bodies and the […]