RSI

A number can describe how strong recent momentum has been without ever telling me what will happen tomorrow.

• Explain what RSI measures — the ratio of average recent gains to average recent losses, rescaled to a 0-100 number.
• Calculate RSI by hand from a series of daily price changes: separating gains and losses, averaging each, forming RS, and applying the RSI formula.
• Interpret a high or low RSI reading as a description of recent momentum only, and identify why it is not a guaranteed buy or sell signal.

RSI matters because you'll see it flashing on almost every trading app you open — usually with a scary label like "overbought" or "oversold" sitting right next to it. If you don't know what builds that number, you either trust it blindly or fear it blindly. Neither is a good place to make decisions from.

Here's what RSI actually is: a ratio of how much a stock has gained on its up-days compared to how much it has lost on its down-days recently, squeezed into a single number between 0 and 100. That's it. No prediction, no crystal ball — just a description of recent behavior.

To build it, take a series of daily price changes and split them into two piles: gains (the up-days) and losses (the down-days). Any day that doesn't move in that direction counts as zero for that pile. Average each pile separately. Divide Average Gain by Average Loss — that ratio is called RS, or Relative Strength. Then rescale RS into the familiar 0-100 number using RSI = 100 − (100 ÷ (1 + RS)).

RSI is only reporting what already happened — so the question changes.

In our worked example, average gain was ₹1.20 and average loss was ₹0.60. RS = 1.20 ÷ 0.60 = 2.0 — gains were twice as strong as losses. Plug that into the formula: RSI = 100 − (100 ÷ 3) = 100 − 33.33 = 66.67. Notice that 66.67 didn't come from nowhere — it's just a stretched-out way of saying "gains have recently outweighed losses, roughly two-to-one."

That's the whole calculation. Nothing hidden, nothing mysterious.

Once you see that RSI is only reporting what already happened, the question changes. A high RSI doesn't say "sell now" and a low RSI doesn't say "buy now" — those labels ("overbought," "oversold") describe zones traders have flagged as worth a closer look, not verdicts. RSI is a clue about recent momentum, and a clue is something you investigate, not something you obey.

Say you're watching a stock and its RSI is showing 66.67 — the exact number from our calculation above, built from five days where the price moved +₹2, +₹1, −₹1, +₹3, −₹2. It's not near the commonly-flagged 70 line, but it's close, and it's climbing.

Here's the moment of choice: what do you actually do with that 66.67?

The tempting move is to treat it like a countdown — "it's near 70, so it's about to be overbought, I should sell before it drops." But you now know exactly what built that 66.67: an average gain of ₹1.20 that was twice the average loss of ₹0.60. That's it. It's a description of a week where up-moves won more than down-moves. It says nothing about tomorrow.

So instead of reacting to the number, you pause and ask: why has this stock been gaining more than it's been losing lately? Maybe there was good news — a strong quarterly result, a new order win. Maybe it's just riding a broader market rally with nothing stock-specific behind it. Maybe volume has been unusually low, meaning the moves aren't backed by much conviction. You don't know yet — and that's exactly the point. The RSI number handed you a clue worth investigating, not a signal to act on. Only after you've looked into the "why" do you decide what, if anything, to do next.

Point: RSI is built from transparent, hand-computable steps — once a learner sees average gain (₹1.20) is double average loss (₹0.60), the resulting RSI of 66.67 stops being mysterious and simply confirms that up-moves have recently outweighed down-moves, describing the past rather than predicting the future.

• Believing an RSI above 70 guarantees a fall, or below 30 guarantees a rise. — Trading apps label these zones 'overbought' and 'oversold', which sounds like a forecast rather than a description of the past. Fix: Remind yourself RSI only summarizes what recent gains and losses have already done — treat a 70+ or 30- reading as a prompt to investigate why, not a command to trade.
• Treating RSI as a mysterious black-box number only experts can understand. — Charts show RSI as a ready-made line with no visible working, so it feels hidden. Fix: Remember it's just Average Gain ÷ Average Loss, rescaled with a simple formula — you can compute it by hand from five numbers, as shown above.
• Assuming a bigger average loss than average gain pushes RSI higher. — Learners assume 'more loss' means a bigger score, without tracing through the actual ratio. Fix: Trace the direction of RS: RSI rises only when Average Gain is larger relative to Average Loss, and falls when losses dominate — the ratio's direction matters, not the raw size of either number.

• RSI is just Average Gain ÷ Average Loss, rescaled onto a 0-100 scale — nothing more.
• A high RSI means recent gains have outweighed recent losses; a low RSI means the reverse — both describe the past.
• No RSI reading, however extreme, guarantees what happens next — it's a clue, not a verdict.
• When you see RSI near 70 or 30, pause and ask why momentum has been that strong or weak before deciding anything.
• Once you can calculate RSI by hand, it stops being a mysterious signal and becomes a tool you can question.

Q1. What two ingredients is RSI actually built from?

• Average Gain and Average Loss over a recent period
• Total trading volume and stock price
• The stock's opening price and closing price
• The number of buyers and sellers that day Answer: Average Gain and Average Loss over a recent period — RSI is a rescaled ratio of Average Gain to Average Loss — nothing about volume or buyer/seller counts goes into the formula.

Q2. In our worked example, a stock's daily changes were +₹2, +₹1, −₹1, +₹3, −₹2. What was the Average Loss over these 5 days?

• ₹0.60
• ₹1.20
• ₹3.00
• ₹1.00 Answer: ₹0.60 — The losing days were ₹0, ₹0, ₹1, ₹0, ₹2. Adding them gives ₹3, and dividing by 5 days gives an Average Loss of ₹0.60.

Q3. If a stock's RSI reading is 75 today, this means the stock is guaranteed to fall in price soon. Answer: False — RSI only describes how strong recent gains have been compared to recent losses. It never guarantees what will happen next — a high reading is a prompt to investigate, not a forecast.

Q4. A trader says: 'This week's average loss was bigger than the average gain, so the RSI must be a very high number.' What is wrong with this thinking?

• RSI rises when Average Gain is larger relative to Average Loss, not when losses are bigger — a bigger loss actually pulls RSI down
• RSI only depends on the total number of trading days, not on gains or losses
• Nothing is wrong — bigger losses always push RSI higher
• RSI can only be calculated over a 14-day period, so a 5-day comparison is meaningless Answer: RSI rises when Average Gain is larger relative to Average Loss, not when losses are bigger — a bigger loss actually pulls RSI down — RSI moves based on the ratio Average Gain ÷ Average Loss. When losses dominate, that ratio (RS) shrinks, which pulls RSI down, not up.

Q5. Over a different 5-day stretch, a stock had an Average Gain of ₹0.60 and an Average Loss of ₹1.20 (the reverse of our worked example). Roughly what would you expect the RSI to be — closer to a high number like 67, or a low number like 33? Answer: Closer to 33 (a low RSI) — Here RS = 0.60 ÷ 1.20 = 0.5, which is the mirror image of our worked example's RS = 2.0. Since losses now outweigh gains, RSI comes out low instead of high — around 33, showing recent momentum has been weak or negative.

Q6. A stock's RSI crosses 30, and a trader buys immediately, thinking: "Oversold means it has to bounce back now." Does an RSI of 30 mean the stock has to bounce? Reveal: Weak: yes, oversold means a bounce is due. Strong: an RSI near 30 only shows that recent losses have outweighed gains — it's a clue worth investigating, not an automatic buy signal; a genuinely troubled company can stay "oversold" a long time.