Setups › RSI(2) dip

RSI(2) dip

From Larry Connors. In one line: short, sharp dip (2-day RSI under 10) inside an up-trend.

About the names. The famous names are credit for publicly described ideas. Every test here uses our own simplified, fully coded version of the idea, so the numbers show how our code behaved on past prices, not how the named person trades or traded. No affiliation with, or endorsement by, any named person.

11,644

trades tested

38%

win rate

-0.14R

average result

-0.87%

vs Nifty per trade

What it looks like

Larry Connors' RSI(2) strategy uses a 2-day RSI, a very short-term measure of how oversold a stock is. When a stock in a long-term up-trend has two or three sharp down days, the 2-day RSI drops under 10. Connors buys that dip and sells on the first bounce.

Why traders use it

Over a few days, stocks often swing back after a sharp move (short-term mean reversion), while over months they tend to keep their trend. RSI(2) tries to use both: only in up-trends, and only for a few days.

The exact rule we tested

Close above the 200-day average.
2-day RSI below 10.

Connors' own exit is to sell on the first close above the 5-day average, not at a fixed target. We tested that exit too: in the November 2026 backtest it won more often (47% vs 39%) and lost less on average across all dip setups.

Mistakes to avoid

Using a wide target. This is a 1-2 day trade; waiting for a big target turns a bounce trade into a hope trade.
Buying dips in down-trends. Without the 200-day filter, "oversold" can stay oversold for a long time.

How we tested it

Every day from 05 Nov 2024 to 05 Nov 2026, every main-board NSE stock (price ≥ ₹50, turnover ≥ ₹5 crore a day) that showed this setup was bought the next day above its last high, with a stop 1.5 ATR below and a target 2.5 ATR above, and sold after 10 trading days if neither was hit; 0.1% costs per trade. Traders usually hold this setup 1-2 days.

The maths behind the result

Average winner +1.10R, average loser −0.88R, win rate 38%. Expectancy = 38% × 1.10 − 62% × 0.88 = -0.14R per trade. To break even with this payoff you need to win at least 44% of trades. How expectancy works.

A historical simulation, not live trading and not a recommendation.