How to Use the RSI Indicator in Indian Markets

What RSI Actually Measures

The Relative Strength Index is a momentum oscillator built by J. Welles Wilder in 1978. It does not measure the price itself. It measures the speed and ratio of upward closes to downward closes over a chosen lookback, almost always 14 periods. The output is squeezed into a 0 to 100 band, which makes it easy to compare a Rs 1,600 HDFC Bank share with a Rs 3,800 Reliance share or with the Nifty 50 index on the same scale.

The single most common error, including in older versions of this very guide, is to feed RSI a tidy price series that rises by the same amount every day, such as 200, 205, 210, 215, and then claim the RSI is around 71. That is mathematically impossible. If every single day closes higher than the day before, there are zero down days, the average loss is 0, the relative strength term blows up, and the RSI is exactly 100. Real markets never move in a clean straight line, so a useful example must use a realistic mix of up days and down days. That is what we do below.

RSI is calculated on closing prices by default. On NSE, that means the closing price stamped at the 3:30 PM close for cash stocks, or the relevant settlement close for index and F&O charts. You can run RSI on any timeframe, from a 5 minute candle for intraday Bank Nifty to a weekly candle for long term position trades. The maths is identical. Only the lookback period changes meaning.

The Formula, Stated Correctly

There are two stages, and skipping the second stage is where most online examples go wrong. The base formula is RSI = 100 minus (100 divided by (1 plus RS)), where RS is the average gain divided by the average loss over the period.

Stage one is the first RSI value. You take the first 14 periods, add up all the gains on up days and divide by 14 to get the first average gain. You add up all the losses on down days, treat them as positive numbers, and divide by 14 to get the first average loss. RS is the ratio of those two.

Stage two is Wilder's smoothing, which every charting platform like TradingView, Zerodha Kite and Upstox uses for every candle after the first. For each new period: new average gain equals ((previous average gain times 13) plus current gain) divided by 14. The same formula applies to losses. This smoothing is why RSI on your screen does not jump wildly, and it is why a plain simple average of the last 14 days will not match what Kite shows you.

A Real Worked Example: HDFC Bank 14-Day RSI

Let us use a realistic, mixed series for HDFC Bank on NSE. These daily closes are illustrative levels in the plausible Rs 1,580 to Rs 1,660 zone the stock has traded, chosen to show a normal mix of up and down days rather than a fake straight line. We need 15 closes to produce 14 daily changes.

Now sum the columns over the 14 changes. Total gains equal 12 plus 13 plus 6 plus 12 plus 7 plus 12 plus 12 plus 9, which is 83.0. Total losses equal 7 plus 8 plus 5 plus 4 plus 4 plus 6, which is 34.0. The first average gain is 83.0 divided by 14, which is about 5.929. The first average loss is 34.0 divided by 14, which is about 2.429.

RS is 5.929 divided by 2.429, which is about 2.441. Plug into the formula: RSI equals 100 minus (100 divided by (1 plus 2.441)), which is 100 minus (100 divided by 3.441), which is 100 minus 29.06, giving a first RSI of about 60.9. That is a genuinely healthy, mildly bullish reading. It is not overbought, which makes sense because the stock climbed from 1,600 to 1,649 with several pullbacks along the way rather than rocketing in a straight line.

Updating RSI for the Next Day With Wilder Smoothing

Suppose on Day 15 HDFC Bank closes at Rs 1,661, a gain of 12.0 over the Day 14 close of 1,649. We do not recompute from scratch. We apply Wilder smoothing using the previous averages. New average gain equals ((5.929 times 13) plus 12.0) divided by 14, which is (77.077 plus 12.0) divided by 14, which is 89.077 divided by 14, giving about 6.363.

There was no loss on Day 15, so the current loss is 0. New average loss equals ((2.429 times 13) plus 0) divided by 14, which is 31.577 divided by 14, giving about 2.256. New RS is 6.363 divided by 2.256, which is about 2.821. New RSI equals 100 minus (100 divided by 3.821), which is 100 minus 26.17, giving about 73.8. The single strong up day pushed RSI from 60.9 into overbought territory, which is exactly the kind of momentum shift the indicator is designed to flag.

• Compute the first average gain and loss as plain averages of the first 14 changes.
• For every later candle, carry 13 parts of the prior average and add 1 part of the new change.
• Recompute RS and RSI each candle. Never re-average the whole window from scratch, or you will not match your broker chart.
• On a day with no gain, the current gain is 0 in the smoothing formula, and vice versa for losses.

Reading RSI Levels in the Indian Market Context

The textbook thresholds are 70 for overbought and 30 for oversold. In practice, large cap Indian names that trend strongly, such as a momentum leader in a bull phase, can sit above 70 for weeks without correcting. Selling Reliance or an index simply because RSI crossed 70 in a strong uptrend is a classic way to exit a winner too early. The level is a warning, not a sell button.

A more durable use is the 50 line. In an uptrend, RSI tends to bounce off the 45 to 50 zone and rarely breaks far below it. In a downtrend, RSI struggles to push above 55 to 60. Watching whether momentum holds above or below 50 often tells you more about trend health than the 70 and 30 extremes. For range-bound midcaps, the 30 and 70 bounces work better because price has no strong trend to override them.

RSI Divergence: The Highest-Value Signal

Divergence is where RSI earns its keep. Bullish divergence happens when price prints a lower low but RSI prints a higher low. The new price low was made on weaker momentum, which often precedes a bounce. Bearish divergence is the mirror: price makes a higher high while RSI makes a lower high, hinting the rally is running out of fuel. On NSE this shows up often near event days such as RBI policy, Budget, or quarterly results, when price spikes but momentum quietly fades.

Divergence is a context signal, not a trigger by itself. A bearish divergence on the Nifty daily chart can persist for several sessions before price actually rolls over. The disciplined approach is to wait for confirmation, such as a break of a short term swing low or a close back below a moving average, before acting. Used alone, divergence will get you short into a market that keeps grinding higher.

• Bullish divergence: price lower low, RSI higher low. Look for a long once price confirms a turn.
• Bearish divergence: price higher high, RSI lower high. Look to trim or hedge, not to short blindly.
• Hidden bullish divergence: price higher low, RSI lower low, often a trend-continuation buy in an uptrend.
• Always pair divergence with a price trigger such as a swing break or moving average reclaim.

Setting Up RSI on Zerodha, Upstox and TradingView

On most Indian platforms the steps are nearly identical. Open the chart, find the technical indicators or studies menu, search RSI, and add it. The default period of 14 is correct for most uses and matches what other traders see, which matters because RSI is partly a self-fulfilling, crowd-watched level. Leave the smoothing method as Wilder's, which is the platform default.

• Open the stock or index chart on Kite, Upstox Pro, or TradingView.
• Open Indicators or Studies and search for RSI or Relative Strength Index.
• Keep the period at 14 unless you have a tested reason to change it.
• Optionally add horizontal lines at 30, 50 and 70 for quick visual reference.
• For intraday Bank Nifty, apply RSI on the 5 or 15 minute candle. For positional trades, use the daily or weekly candle.

Avoid over-tuning. Shortening RSI to a 7 period makes it twitchy and throws far more false signals, which suits fast scalpers but punishes swing traders. Lengthening it to 21 smooths the line and lags more, which suits position trading. Pick the period that matches your holding time and then stop fiddling, because constantly changing settings is just curve fitting to the recent past.

Worked F&O Trade: Bank Nifty RSI Bounce With Real Costs

RSI signals are most often traded through options or futures in India, so the rupee result depends heavily on costs and tax. Here is an illustrative, not guaranteed, example. Suppose Bank Nifty falls into a support zone and the hourly RSI shows bullish divergence near 28. You decide to buy one lot of a weekly at-the-money call. Bank Nifty lot size is 30. Say the index is around 48,000 and you buy the 48,000 CE at a premium of Rs 320.

Your entry cost is 320 times 15, which is Rs 4,800 of premium. Suppose the RSI bounce plays out and the option rises to Rs 470, so you sell at 470 times 15, which is Rs 7,050. The gross gain is Rs 2,250. Now apply the real costs. STT on options is charged on the sell side. For options it is 0.1 percent of the premium value on sell, so 0.1 percent of 7,050 is about Rs 7. Discount broker flat fees are typically about Rs 20 per executed order, so roughly Rs 40 for the buy and sell legs combined.

Exchange transaction charges, SEBI charges, GST on brokerage and stamp duty add a small amount more, realistically another Rs 25 to Rs 35 for a single lot round trip. Rounding the total frictional cost to roughly Rs 85, your net profit before income tax is about Rs 2,250 minus Rs 85, which is around Rs 2,165 on one lot. The point is not the exact paisa. It is that costs are real and a thin RSI scalp can have its edge eaten by fees if you over-trade.

RSI on Delivery Equity and the Tax That Applies

If you instead use RSI to time a delivery purchase of an NSE stock, the tax treatment is different. Say RSI on HDFC Bank dips below 30, you buy 100 shares at Rs 1,590 for an outlay of Rs 1,59,000, and you sell at Rs 1,720 within a few months for Rs 1,72,000. Your gross gain is Rs 13,000. Delivery STT is 0.1 percent on both buy and sell, so roughly Rs 159 plus Rs 172, about Rs 331 in STT, plus small brokerage and statutory charges.

Because you held under 12 months, this is a short term capital gain taxed at 20 percent under the current rules. Twenty percent of roughly Rs 12,600 net gain is about Rs 2,520 in tax. Had you held the shares more than 12 months, it would be a long term capital gain, where gains are tax free up to Rs 1.25 lakh in a financial year and taxed at 12.5 percent above that. These numbers are illustrative and depend on your full year totals, so confirm with a tax professional.

RSI vs MACD vs Stochastic

RSI is one of several momentum tools, and each answers a slightly different question. RSI tells you whether momentum is stretched. MACD tells you about the relationship and crossovers between two moving averages, which is better for confirming trend direction. The Stochastic oscillator compares the close to its recent high-low range and is more sensitive, which makes it popular for fast intraday signals but noisier.

The practical takeaway is to use RSI for the momentum read and let a second tool confirm. A common, robust pairing on NSE is RSI with a moving average. Take RSI signals only in the direction of the longer term average. If price is above its 50 day average, favour RSI buy signals and ignore most RSI sell signals, and the reverse in a downtrend. This single filter removes a large share of losing counter-trend trades.

Common Mistakes That Cost Indian Traders Money

The biggest mistake is treating 70 and 30 as automatic sell and buy triggers. In a strong trend, an index or a leading large cap can hold above 70 for a long stretch, and shorting it just because RSI is high is a fast way to bleed. The second mistake is using a simple average instead of Wilder smoothing, which makes your own RSI disagree with your broker's chart and corrupts any backtest.

• Blindly shorting at RSI 70 or buying at RSI 30 without trend context or a price trigger.
• Computing RSI with a plain simple average rather than Wilder smoothing, so your numbers do not match Kite or TradingView.
• Ignoring divergence, which is usually the most reliable RSI signal.
• Over-trading thin RSI scalps in options, where STT and brokerage quietly erase the edge.
• Forgetting that F&O profit is taxed as business income at slab rate, not at capital gains rates.

Finally, do not over-optimise the period. Traders who keep switching between 7, 9, 14 and 21 looking for the setting that would have caught the last move are curve fitting. Pick a period that matches your holding time, confirm signals with one other tool, respect your stop loss, and journal every trade so you can see whether RSI is actually adding edge to your specific strategy.

Sources and Further Reading

For authoritative material on RSI mechanics, Wilder smoothing and Indian charges, see Zerodha Varsity, Investopedia, and NSE India. Always confirm current STT rates, lot sizes, brokerage and tax rules on the official source before you trade, since contract specifications and charges change.

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.