[Session 2] Shoukinryu: Design philosophy of M5 scalping achieving an 84% win rate
When talking about the Ascending Gold Dragon, the first question often asked is “why M5 (5-minute chart)?”
M1 (1-minute) offers more trading opportunities. H1 (1-hour) has less price movement noise. M5 might appear like a mid-range, somewhat ambiguous choice. In fact, at the early development stage I confronted this question and backtested across multiple timeframes from M1 to H4.
As a result, I concluded that M5 is the most suitable for gold scalping. In this piece, I will explain the reasons from both data and design philosophy. You can understand how the figure of 84.28% win rate is built on the underlying logic.
First, let’s organize the relationship between time frame and noise. In FX trading, “noise” refers to short-term price movements that are unrelated to the actual market direction.
M1 confirms candles every minute, so there are more trading opportunities. However, most of the price movement is noise, and entry accuracy drops significantly. While many scalping methods use M1, for gold the spread has a big impact, making it very difficult to profit while wading through M1 noise.
On the other hand, above H1 the signal quality improves, but the number of entries decreases, erasing the advantage of scalping. A strategy that quickly enters when RSI indicates oversold and stacks small rebounds requires a certain trading frequency. M5 offers the best balance of noise and signal for gold scalping. That was my conclusion.
The 84.28% win rate of Ascending Gold Dragon is built from a combination of three logics. Remove any one of them and this win rate cannot be achieved.
Ascending Gold Dragon enters a long position only when short-term RSI (period 2) falls below a configured value. This is contrarian logic expecting a price rebound from oversold conditions.
RSI(2) is highly sensitive, calculated over just two candles. Using M5 allows real-time detection of short-term oversold conditions within 5–10 minutes. On M1, noise is too high and RSI(2) reacts too often, lowering entry accuracy. M5 time frame makes RSI(2) filter most effective.
Even after entering via RSI, the market may fall further. In that case, we add positions (averaging down) but Ascending Gold Dragon does not fix the averaging interval. It dynamically calculates the interval using ATR (Average True Range).
When gold is choppy, price movement is large, so the averaging interval widens automatically. In calm ranges, the interval tightens, enabling quicker average entry price reduction. Fixed-width averaging ignores market volatility. Using ATR enables adapting averaging to market conditions.
The core of the 84% win rate lies in this settlement logic. Instead of individually taking profits or stopping out each position, all open positions are settled the moment their combined profit becomes positive.
For example, if you hold three positions with -10k, -20k, and +50k, the total is +20k, so all positions are closed at that moment. Even positions with unrealized losses are included in the profit-taking. This mechanism ensures that even when averaging down the drawdown, a slight market recovery can push the entire portfolio into positive territory.
The combination of “use RSI on M5 and average down with ATR” may look simple, but these three elements are deliberately integrated.
RSI(2) on M5 responds → considers oversold and enters → if it falls further, adds positions using the ATR-calculated averaging width → once the market recovers and the total P/L becomes positive, settle all at once.
This is precisely why RSI(2) is effective on M5, gold’s volatility enables ATR to function, and total P/L settlement makes averaging down less intimidating. These three components are inseparable in this design.
It may look complex, but the parameters are kept to a minimum: RSI settings, ATR multiplier, maximum number of positions, and profit target line. With just these four numbers, the logic has navigated 11 years of market conditions. Simplicity yields robustness—that is the design philosophy.
Here are backtest results for the same logic across multiple timeframes. All figures are based on backtesting from 2015 to 2026.
| Timeframe | Net Profit | Win Rate | PF | RF | Max DD |
|---|---|---|---|---|---|
| M1 (1-minute) | +9.0 million JPY | 71.3% | 1.42 | 3.1 | 28% |
| M5 (5-minute) ★Recommended | +25.0 million JPY | 84.28% | 1.93 | 9.95 | 12% |
| M15 (15-minute) | +16.3 million JPY | 78.6% | 1.71 | 6.2 | 17% |
| H1 (1-hour) | +9.7 million JPY | 68.9% | 1.58 | 4.4 | 14% |
M5 significantly outperforms other timeframes. Net profit is about 1.5 times that of the second-best M15, and the Recovery Factor is an impressive 9.95.
In particular, compare with M1. Although M1 has higher trade frequency, the selected entries on M5 outperform in net profit, win rate, and RF. “More trades mean more profit” does not hold. Entry quality determines long-term results. This data proves it.
Whether the logic described so far actually works can be verified in YouTube live streams.
Asc Gold Dragon reveals a 24/7 real-time operation screen. Entries occur the moment RSI reaches the configured value, averaging-down entries occur with ATR-calculated widths, and total P/L turns positive so all positions are settled. You can watch all of it live.
“Saying the win rate is 84% in backtests is easy. But I want you to see how that number behaves in real market conditions. I stream moments when drawdowns occur and ATR averaging down happens, as well as when the total P/L turns positive and settlements occur. Rather than explaining the design philosophy in words, watching the live screen conveys it more accurately.”
- M5 is the time frame with the best balance of noise and signal for gold scalping
- The 84.28% win rate arises from the combination of RSI filter, ATR dynamic averaging, and total P/L settlement
- M5 outperforms M1 in net profit, win rate, and recovery factor despite fewer trades
- A simple design (RSI + ATR) yields adaptability to 11 years of market conditions
- The actual movement of the logic can be confirmed in the YouTube 24-hour live stream
※This article is provided for information purposes only and is not an investment solicitation. The performance results shown are past results and do not guarantee future profits. FX/CFD trading involves risk. Please make investment decisions at your own risk.
