Does the RSI(2) strategy work on the S&P 500?

Key results

Total return

+18.1%

Buy & hold: +255.4%

Annual return (CAGR)

+1.7%

Buy & hold: +13.5%

Max drawdown

-21.5%

Buy & hold: -35%

Win rate

71.23%

Profit factor

1.43

Trades

73

Longest losing streak

3 trades

Period

Oct 3, 2016 to Oct 1, 2026

10 years

Market

SPX500/USD

Daily

Equity curve vs buy & hold

Returns year by year

Calendar years, compounded from monthly results. The first and last years can be partial.

RSI(2) is the best-known short-term setup from Larry Connors: a 2-period RSI that drops below 5 flags an oversold pullback, the 200-day moving average decides whether buying it is allowed, and a close above the 5-day average ends the trade. StockCharts' ChartSchool describes it in detail, and that is the version we built. The S&P 500 is one of the markets this setup is most often shown on, so it is the fairest place to check whether the published logic survives ten recent years with costs included.

The rules we tested

• Market and period: S&P 500 (SPX500/USD), daily candles, October 2016 to October 2026.
• Long entry: RSI(2) below 5 and the close above the 200-day simple moving average (SMA).
• Long exit: the close is back above the 5-day SMA, or below the 200-day SMA.
• Short entry: RSI(2) above 95 and the close below the 200-day SMA.
• Short exit: the close is below the 5-day SMA, or back above the 200-day SMA.
• Stop and target: a wide protective stop-loss, hit once in 73 trades, and no take-profit.
• Sizing and costs: 100% of equity per trade, 0.02% commission per side, $10,000 starting capital.

What the backtest shows

The edge is there, but it is thin. The average win was 1.22% and the average loss 2.10%, so the strategy needs about 63% winners to break even. It got 71.2%, a margin of roughly eight points, which is what a profit factor of 1.43 looks like.

The total, +18.1% in ten years, is 1.7% a year against 13.5% a year for the index. The drawdown tells the same story from the other side: the strategy's worst fall was 21.5%, larger than everything it earned over the decade, while buy and hold fell 35.0% at worst on the way to +255.4%. Being out of the market most of the time did not make the account safer in proportion to what it gave up.

The years are uneven. 2018 was the worst (-9.8%, 3 winners out of 7 trades), a year in which pullbacks above the 200-day average turned into deeper sell-offs. The best were 2021 (+8.63%, all 6 trades winners) and 2024 (+8.58%, 9 winners out of 10), followed by +6.4% in 2025 and +6.5% in the first nine months of 2026. With 73 trades in ten years, commissions are not what held it back.

For comparison, the same rules made +26.9% on the Nasdaq 100 and +14.4% on the DAX over the same period: positive on all three indices, behind buy and hold on all three.

Why it works, and why it is not enough

The mechanism is simple: inside a long uptrend, a two-day dump tends to be followed by a bounce, and RSI(2) below 5 is a strict way of catching only the sharpest dumps. The S&P 500 is less volatile than the Nasdaq, so its pullbacks are shallower and the bounce to the 5-day average is smaller. That shows up directly in the average win of 1.22%. Meanwhile, a losing trade waits for either the 5-day average or the 200-day average, and in a sell-off like 2018 that wait is expensive.

What you could test next

• Use RSI(2) below 10 to trade more often and see whether the margin over break-even holds.
• Exit on a close above the previous day's high instead of the 5-day SMA.
• Run it long only, since the S&P spent most of the decade above its 200-day average.
• Try the same logic on 4-hour candles to get a larger sample of trades.

Methodology and assumptions

Frequently asked questions

Reproduce it, then change it

Every number on this page comes from the Backtrex engine. Rebuild the strategy in the app, then change the market, the timeframe or a parameter and see whether the result holds.

Other backtests