Connors RSI for Indian Markets: Formula, Worked Example and Levels

What Connors RSI Actually Measures

Connors RSI was designed by Larry Connors and his team to answer one narrow question well: how stretched is price right now, relative to its own recent behaviour? A normal Relative Strength Index only looks at the size of up moves versus down moves. Connors RSI adds two more lenses. It looks at how long the current run of up or down days has lasted, and it ranks today's single-day return against a long history of single-day returns. The three numbers are then averaged into one reading from 0 to 100.

This matters for Indian traders because our most liquid names, like Reliance, HDFC Bank, TCS and Infosys, and the indices Nifty and Bank Nifty, tend to mean-revert sharply after short bursts of selling or buying. Connors RSI is built precisely for that rhythm. It is a short-term indicator. It is not a trend filter, and using it as one is the fastest way to lose money with it.

The key mental model: Connors RSI is a stretched-rubber-band meter. A reading near 0 says the band is stretched hard to the downside and a snap back is statistically likely. A reading near 100 says the opposite. It never tells you the band will snap, only that it is unusually stretched.

The Three Components, Stated Correctly

Most articles, including older versions of this page, describe the third component as a 100-day percentage change. That is wrong. Here is each component stated the way the formula actually works, so your calculation matches what TradingView, Amibroker or a Python script will produce.

• Component 1, RSI of price, default period 3. This is an ordinary Wilder RSI applied to closing prices, but over a very short 3-day window so it reacts almost instantly.
• Component 2, RSI of the streak, default period 2. First you build a streak series: +1, +2, +3 for consecutive up-close days, and -1, -2, -3 for consecutive down-close days, resetting to 0 on an unchanged close. Then you run a 2-period RSI on that streak series. This measures how persistent the current run is.
• Component 3, PercentRank of the 1-day return, default lookback 100. Take today's 1-day percentage return. Count how many of the prior 100 daily returns were smaller than it. Express that count as a percentage. A value of 5 means today's move was at the 5th percentile, that is, among the most negative one-day moves in the last 100 sessions.

The final value is a simple average: Connors RSI = ( RSI(close,3) + RSI(streak,2) + PercentRank(return,100) ) / 3. All three pieces already live on a 0 to 100 scale, so averaging them is valid. Getting component 3 right is what separates a correct Connors RSI from a broken one.

A Fully Worked Connors RSI on Reliance

Let us compute Connors RSI by hand on an illustrative Reliance Industries daily series after three straight down sessions. All prices below are illustrative round numbers chosen to make the arithmetic clear, not live quotes. The method is exactly what you would apply to real NSE end-of-day data.

Assume these recent Reliance closes, oldest to newest: 2986, 2972, 2960, 2940, 2920, 2905, 2868. The last three daily changes are therefore -20, -15 and -37 rupees, three consecutive down days into today's close of 2868.

Now average the three: Connors RSI = (0.00 + 9.09 + 10.0) / 3 = 6.36. A reading of 6.36 is deeply oversold on the Connors scale. Note why RSI(close,3) pinned at exactly 0: when every one of the last three closes is lower than the one before, the average gain is zero, so the formula returns 0 by construction. That is correct behaviour, not a bug, and it is a feature of using such a short 3-day window.

In the real world your PercentRank uses the full 100-day lookback rather than the short window used here for teaching, so the exact number will differ. The shape of the result will not: a sharp three-day drop in a liquid large cap typically drives Connors RSI into single digits. That is the signal the indicator exists to flag.

How to Read the Levels (And Why 30 and 70 Are Wrong Here)

Standard RSI uses 30 and 70. Connors RSI is far more reactive because two of its three parts use 2 and 3 period windows, so it spends much more time near the extremes. Connors himself tested entries around readings below 10 to 15 for longs and above 85 to 90 for shorts, not 30 and 70. Applying ordinary RSI thresholds to Connors RSI will flood you with signals that mean very little.

• Below 10: deeply oversold. Connors' research on US equities favoured buying weakness here within an established uptrend, then exiting on a bounce.
• 10 to 30: oversold, weaker edge. Worth watching but not a standalone trigger.
• 30 to 70: the middle zone. Connors RSI gives little usable edge here. Stand aside.
• 70 to 90: overbought, weaker edge on the short side.
• Above 90: deeply overbought. A candidate to fade or to book profits on existing longs.

For Indian names this still holds, but our market opens with gaps driven by overnight US and Asian moves, so a Connors RSI extreme at the previous close can be erased by the opening tick. Always re-read the indicator after the first 15 to 30 minutes rather than acting blindly on the prior day's print.

Connors RSI Versus Standard RSI

A quick comparison clarifies when to reach for each tool. Standard RSI is steadier and works as a trend and momentum gauge across many timeframes. Connors RSI is jumpier and is built for one job: timing short-term mean-reversion entries in liquid instruments.

A practical combination many Indian swing traders use: take direction from a longer tool such as the 50-day moving average or standard RSI, then use Connors RSI only for the entry trigger in that direction. Buy a sub-10 Connors RSI print only when the stock is above its 50-day average, and the win rate improves markedly versus buying every oversold reading.

Settings for Indian Stocks and Indices

The default CRSI(3, 2, 100) is a sensible starting point and you should resist the urge to over-tune it. That said, instrument liquidity matters more than the exact numbers. Connors RSI is only reliable on names that trade enough to make short-term price changes meaningful rather than noise.

• Nifty and Bank Nifty: defaults work well. These are deep, liquid, and mean-revert cleanly intraday and over a few days.
• Large-cap single stocks (Reliance, HDFC Bank, TCS, Infosys, ICICI Bank): defaults work. Re-confirm signals with volume.
• Mid and small caps: be cautious. Thin volume creates erratic streaks and unreliable PercentRanks. Connors RSI is weakest exactly where spreads are widest.
• Intraday on 5 or 15 minute charts: some traders shorten the lookback (for example CRSI(3, 2, 50)) so the PercentRank reflects the current session's behaviour rather than weeks of history.

Whatever you choose, backtest it on NSE end-of-day data across at least one bull phase, one bear phase and one sideways phase before risking capital. Connors RSI strategies that look brilliant only in trending data usually fall apart in the choppy ranges that dominate Indian markets between events.

Trading the Signal: A Sized Example With Costs

Suppose the Reliance Connors RSI print of 6.36 above triggers a mean-reversion long, and you express it with one lot of Reliance futures. Take an illustrative lot size of 500 shares (always confirm the live NSE lot size, as the exchange revises these). You buy at 2868 and the band snaps back to your target of 2920, then you exit. This is a teaching example, not a prediction, and there is no guaranteed outcome.

• Entry: buy 1 lot Reliance futures at 2868. Notional = 2868 x 500 = Rs 14,34,000.
• Exit: sell at 2920. Gross gain = (2920 - 2868) x 500 = 52 x 500 = Rs 26,000.
• STT on futures sell side at 0.02 percent of sell value: 0.0002 x (2920 x 500) = 0.0002 x 14,60,000 = Rs 292.
• Brokerage, exchange charges, GST and stamp duty together are illustratively around Rs 200 to 300 for a discount broker.
• Net gain after costs is roughly Rs 26,000 - 292 - 250 = about Rs 25,458, before income tax.

Because this is an F&O trade, the profit is business income and is taxed at your income-tax slab rate, not at the equity STCG or LTCG rates. If instead you had taken delivery of Reliance shares and sold within a year, the gain would be short-term capital gain taxed at 20 percent; held beyond a year it would be long-term capital gain taxed at 12.5 percent on the amount above the Rs 1.25 lakh annual exemption. Logging the trade, the Connors RSI reading at entry, and the actual costs in your trading journal is what turns a single signal into a measurable edge over time.

Limitations and Where It Breaks

Connors RSI is a mean-reversion tool, so its single biggest weakness is a strong trend. In a powerful downtrend the indicator can sit below 10 for days while price keeps falling, and every oversold buy gets run over. This is why pairing it with a trend filter is not optional for most traders, it is the difference between an edge and a slow bleed.

• Strong trends: repeated extreme readings that never revert. Use a trend filter to avoid fading a freight train.
• Illiquid stocks: noisy streaks and meaningless PercentRanks. Stick to liquid names.
• Event risk: results, RBI policy, Budget and global shocks can gap price past your stop overnight. F&O and overnight delivery positions both carry this risk.
• Over-optimisation: tuning the (3, 2, 100) parameters until a backtest looks perfect almost always fails forward. Keep it simple.

None of this makes Connors RSI useless. It makes it a specialist. Used for its intended job, timing short-term snapbacks in liquid instruments with a trend filter and hard risk limits, it is a sharp tool. Used as a general-purpose buy or sell signal, it will disappoint.

Sources and Further Reading

For the original methodology see the work of Larry Connors and Cesar Alvarez. For Indian rules, contract specifications and current charges, always confirm on official sources: Zerodha Varsity, Investopedia and NSE India. Lot sizes, STT rates and tax rules change, so verify the live numbers before you trade.

Sources and Further Reading

For authoritative data and further reading on this topic, refer to Zerodha Varsity, Investopedia and NSE India. Always confirm current rules, rates and contract specifications on the official source before you trade.