Crypto Scalping Strategy Backtested: 6 Months of 15-Minute Data on BTC

A crypto scalping strategy sounds compelling. Fast trades. Frequent wins. Small profits that compound. The theory is clean. The reality, as the data shows, is more complicated.

We ran a full backtest of a 15-minute BTC scalping strategy on Kaiko institutional data sourced from Binance. Six months of data. 92 trades. The results challenge one of the most common misconceptions in short-term crypto trading: that a high win rate means a profitable strategy.

The Strategy: RSI Oversold Bounce on 15-Minute BTC

This backtest uses a single-indicator scalping approach built entirely on RSI(14). The logic targets oversold conditions on Bitcoin and enters when the market shows signs of recovery.

How signals are generated

A buy signal fires in two steps:

1. RSI(14) crosses below 30, confirming Bitcoin has entered oversold territory on the 15-minute chart

2. RSI(14) then crosses back above 30, confirming the bounce has started

This two-step sequence filters out premature entries. Rather than buying at the moment RSI drops below 30, the strategy waits for confirmation that selling pressure is fading and momentum is turning.

The exit condition is equally straightforward: close the position when RSI(14) crosses above 60, capturing the move from oversold recovery to neutral-to-bullish territory.

Why this qualifies as a scalping strategy

With 15-minute candles and RSI(14) governing both entry and exit, positions typically last between 1 and 6 hours. The strategy takes short, targeted positions around local price reversals rather than holding through extended trends. This is the core of scalping logic: precise entries on short-term dislocations with defined, rule-based exits.

Backtest Configuration

Backtest Results: 6 Months of 15-Minute BTC Data

Key Metrics: Total Return: -16.88% | Win Rate: 66.30% | Total Trades: 92 | Max Drawdown: 30.60%

The numbers create an immediate puzzle. A 66.3% win rate means the strategy was right on direction more than two out of every three trades. That is a better accuracy rate than most discretionary traders achieve. Yet the strategy lost 16.88% of the starting capital over the test period.

This is the win rate paradox in crypto scalping, and it is worth understanding before committing to any short-term strategy.

Why High Win Rate Does Not Equal Profitability

The issue is not accuracy. It is the distribution of wins versus losses.

When a scalping strategy exits on RSI(14) crossing above 60, the average winning trade captures a relatively modest price move. RSI moves from 30 back to 60 on BTC's 15-minute chart often correspond to a percentage gain of 0.5% to 2%. When the exit fires quickly, the gain is small.

When a losing trade occurs, the position holds through extended weakness, waiting for RSI to either recover and trigger the exit condition or continue lower without a clean bounce signal. In a trending bear move, that can mean sitting through a significant drawdown before the position eventually closes.

The result: many small wins, fewer but proportionally larger losses. The 30.60% maximum drawdown confirms that the worst losing sequences were severe relative to overall capital, even though most individual trades were profitable.

This pattern is the core challenge of mean-reversion scalping in trending markets. During the December 2025 to June 2026 backtest window, BTC experienced extended directional moves that punished oversold-bounce strategies. RSI touched below 30 frequently during these periods, generating entry signals that led into further declines rather than clean recoveries.

What the Data Shows About Crypto Scalping

Win rate alone is not a valid performance metric. 92 trades with 66.3% accuracy and a net loss of 16.88% demonstrates that accuracy and profitability are separate measurements. Profitable strategies need both reasonable win rates and favorable risk-reward ratios.

Oversold conditions are not automatic reversal signals on 15-minute data. In trending markets, RSI can stay oversold for extended periods. An asset can close below RSI(30) on 15-minute bars repeatedly without triggering a sustained bounce. Mean-reversion logic works best in range-bound, sideways conditions.

Scalping frequency affects risk exposure differently than it appears. With 92 trades across 6 months, this strategy averages roughly 15 to 16 trades per month. Each trade exposes the full capital position (100% sizing) to a potential losing sequence, which explains how the 30.60% drawdown developed despite a high win rate.

Market regime matters more at short timeframes. At the 15-minute level, BTC's market structure during the test period, periods of sharp selling, ETF flow sensitivity, and macro-driven volatility, created conditions where oversold bounces frequently failed to sustain.

Frequently Asked Questions

Q: How did the crypto scalping strategy perform on BTC/USDT 15-minute data?

Over the six-month backtest from December 9, 2025 to June 9, 2026, the RSI oversold bounce scalping strategy generated a total return of -16.88% on a $10,000 starting capital. The win rate was 66.30% across 92 trades. The maximum drawdown reached 30.60%.

Q: Why does the strategy have a high win rate but negative returns?

The average winning trade captures a relatively small price move (RSI recovering from oversold to 60), while losing trades can hold through extended declines before closing. This creates an asymmetric outcome where many small wins are offset by fewer, larger losses.

Q: Is RSI scalping a reliable crypto trading strategy?

RSI-based scalping can be effective in range-bound market conditions where price regularly oscillates and oversold conditions lead to reliable bounces. It tends to underperform in trending markets where price continues lower after entering oversold territory. Backtesting across multiple market conditions is essential before applying any scalping strategy to live trading.

Q: How can I test this strategy on CoinQuant?

Type the following prompt into CoinQuant's strategy builder:

BTCUSDT 15m. Enter long when RSI(14) crosses below 30, then crosses back above 30. Exit when RSI(14) crosses above 60. Backtest the last 6 months.

CoinQuant will build the strategy logic, connect to Kaiko institutional data from Binance, and run the full backtest automatically. No coding required. No Python. No Pine Script.

Q: What improvements can make this scalping strategy more robust?

Adding a trend filter, such as requiring price to be above the 50-period SMA before taking RSI oversold entries, can reduce losing trades in downtrending markets. Adjusting position sizing to a fixed percentage risk per trade (rather than 100% of capital) also limits the impact of losing sequences on overall capital. CoinQuant lets you modify any of these parameters and rerun the backtest instantly.

Try This Exact Strategy on CoinQuant

To run this exact backtest and see the results yourself, type the following into the CoinQuant strategy builder:

BTCUSDT 15m. Enter long when RSI(14) crosses below 30, then crosses back above 30 (oversold bounce confirmed). Exit when RSI(14) crosses above 60. Backtest the last 6 months.

No coding required. No Pine Script. No Python. CoinQuant builds the strategy from your natural language description and runs the backtest against Kaiko institutional data from Binance automatically.

Backtest scalping strategies on CoinQuant