Buy Gold When RSI Hits 70 — A 3-Year Backtest of the Contrarian Momentum Strategy

RSI hits 70 on gold and most traders scream "overbought." Time to sell. But what if you did the exact opposite?

I ran a dead-simple experiment from September 2022 to June 2025 using a 7-period RSI on the daily chart. Buy when RSI crosses above 70. Sell when it drops below 30. No stop-loss, no take-profit. Just hold until the signal flips. The results were, honestly, surprising.

The Core Idea — Chasing Momentum Instead of Fading It

Key finding: A momentum strategy that buys RSI overbought and sells RSI oversold on gold produced a 61% win rate and a profit factor of 2.0 over nearly three years of daily chart data.

Textbooks teach RSI as a mean-reversion tool. Above 70 means overbought, so sell. Below 30 means oversold, so buy. Logically sound.

The problem is that gold doesn't care about logic.

Gold is notorious for starting a move in one direction and just... not stopping. Traders who short because "it's already gone too far" get burned repeatedly. That's the entire premise behind this test: what if instead of fighting gold's momentum, you simply followed it?

The Setup — Brutally Simple

Here's the full configuration:

• Indicator: RSI (7-period, simple)
• Timeframe: Daily chart
• Buy signal: RSI ≥ 70 (on close)
• Sell signal: RSI ≤ 30 (on close)
• Position management: Close existing trade and reverse on opposite signal
• Stop-loss / Take-profit: None

That's it. No moving averages. No news filters. No additional confirmation. Pure, raw momentum — if gold is moving up, buy. If it's moving down, sell.

Backtest Results — $10,000 to $17,000

The test ran from September 1, 2022 through June 2025, covering roughly three years of gold price action.

Metric	Value
Starting capital	$10,000
Peak equity	$17,000
Net profit	$7,100
Win rate	61%
Profit factor	2.0
Average trade	$397
Max drawdown	33% ($3,350)
Median hold time	~30 days

The numbers look solid. But behind them lies a process that was anything but comfortable.

Why This Works on Gold

Gold has one characteristic that separates it from many other assets: when it trends, it trends hard.

There was a position entered in early 2024 when RSI crossed above 70. The conventional read would be "overbought, due for a pullback." Instead, that position stayed open for months as the RSI never dipped back below 30. Gold moved from $2,390 per ounce to over $2,635 while this simple strategy just... sat there.

The final profit on that single trade: $6,144.

These "monster winners" are the engine of this approach. You endure frequent small losses in exchange for capturing the full run when gold goes on a tear.

The Weakness — Choppy Markets Eat You Alive

Honesty matters here: the early stretch of this test was painful.

Right after starting, the strategy ate consecutive losses. RSI hit 70, triggered a buy, then gold immediately reversed. RSI hit 30, triggered a sell, gold bounced right back. In ranging markets, this whipsaw pattern repeats relentlessly.

This is the structural weakness of every trend-following strategy. No trend, no profit. But gold tends to produce powerful trends more frequently than many other assets, which over time offsets the chop losses.

Practical Takeaways

This isn't a recommendation to trade this strategy as-is. The backtest revealed clear areas for refinement.

First, position sizing needs reduction. A 33% drawdown is psychologically brutal for most traders. Capping risk at 2% of capital per trade is the realistic approach.

Second, adding stop-losses is worth exploring. Fixed stops on every trade would significantly reduce drawdown depth. The backtest optimizer suggested that adding both stop-loss and take-profit levels would actually improve overall returns.

Third, time-based exits could help. Putting a maximum holding period on positions reduces opportunity cost during sideways markets.

The central takeaway: the instinct to sell gold when it looks "overbought" has historically been the losing move. Respecting the trend — at least on gold — has produced statistically better outcomes.