Technical Indicators

1. What you'll learn and why it matters

In Chapters 2 and 3 you learned to read price itself: candles, support and resistance, trends and patterns. An indicator is a formula applied to that same price and volume data. It turns a messy chart into a cleaner number or line that answers one specific question: Is there a trend? How strong is it? Is the move overstretched? How big is a normal day's move?

By the end of this chapter you will be able to:

• Tell leading from lagging indicators, and oscillators from trend-following tools, and know what each family is good and bad at.
• Calculate by hand (and in Python) the SMA, EMA, RSI, MACD, Bollinger Bands, ATR, ADX/DMI, Supertrend, OBV, VWAP, Stochastic, Donchian channels and relative strength.
• Read the standard signals (crossovers, overbought/oversold, divergences, squeezes, band walks, breakouts), and recognise when each one usually fails.
• Use ATR to place stops and size positions, which is the single most practical use of any indicator in this chapter.
• Combine a small set of indicators that do different jobs (trend + momentum + volatility) rather than five that say the same thing.

2. What an indicator is: leading vs lagging, oscillators vs trend-following

Varsity groups indicators into two broad camps, and the split is worth understanding before you learn any single formula.

An oscillator is an indicator that moves between fixed bounds (RSI and Stochastic run from 0 to 100) or around a centre line (MACD oscillates around zero). Because it is bounded, it can show "extremes". A trend-following indicator, such as a moving average, is plotted in price units on top of the chart and has no extremes. It simply follows price with a delay.

A third group is easy to overlook but in practice it is the most useful: volatility indicators (ATR, Bollinger bandwidth). They say nothing about direction. They tell you how large a normal move is, which is exactly what you need to place a stop that is not too tight and to size a position. A fourth group is volume-based (volume average, OBV, VWAP), which tells you whether a move has participation behind it.

3. Moving averages: SMA and EMA

3.1 The idea in plain language

Daily closes are noisy. A moving average (MA) replaces each day's close with the average of the last n closes. Every day the oldest close drops out and the newest one comes in, so the average "moves" along the chart. The result is a smooth line that shows the underlying direction. Varsity's short summary is "the trend is your friend": the MA is a way to see that trend.

The logic: averaging cancels random day-to-day noise but keeps persistent drift. If most of the last 50 closes were higher than the 50 before them, the 50-day average is rising, and that is about as simple a definition of an uptrend as you can get. The cost of smoothing is lag. The average always trails price, and the longer the window, the bigger the lag.

3.2 Simple moving average (SMA)

Every one of the n closes gets equal weight \(1/n\). One side effect is the "drop-off effect": the SMA can change direction just because a big old close falls out of the window, even if today's price did nothing unusual.

3.3 Exponential moving average (EMA)

The EMA gives more weight to recent prices. Varsity ties this to the idea that "markets discount everything", so the latest price carries the most information. It is recursive: today's EMA is yesterday's EMA nudged towards today's close by a fraction \(\alpha\), called the smoothing factor.

For a 10-day EMA, \(\alpha = 2/11 = 0.1818\): today's close gets 18.2% weight and the past gets the remaining 81.8%, which itself decays geometrically. For a 20-day EMA, \(\alpha = 2/21 = 0.0952\). For the first value, charting platforms usually start ("seed") the EMA with the SMA of the first n closes, and we do the same here.

3.4 SMA vs EMA on a real turn

3.5 Choosing the length

4. Moving-average crossovers and the golden/death cross

4.1 The crossover system

To reduce whipsaws, Varsity suggests using two averages instead of price vs one average. The shorter one is the "fast" MA and the longer one is the "slow" MA. Buy when the fast MA crosses above the slow MA. Exit when it crosses back below. Varsity lists these popular EMA pairs:

Varsity is clear about the profile of such systems: many small losses, a few large wins, and you must take every signal, because the edge comes from the rare big trend and you cannot know in advance which signal it will be.

4.2 Golden cross and death cross

The most famous pair is the 50-day and 200-day SMA. A golden cross is the 50-day crossing above the 200-day. A death cross is the 50-day crossing below it. Financial media love these events. Here is every one of them on the Nifty 50 in our data:

5. Price vs the 50/200-day average as a trend filter

Many swing traders do not trade crossovers at all. They use "close above the 200-day SMA" (often together with "50-day above 200-day") only as a filter: long setups are considered only when it is true. The idea is that buying pullbacks inside long-term uptrends puts the larger tide behind you.

6. RSI: Wilder's momentum oscillator

6.1 What it measures and why

Momentum is the speed of price change. If a stock rises 10% in 5 days it has high momentum. If it takes 5 months, momentum is low. J. Welles Wilder's Relative Strength Index (1978) compares the size of recent up-days with the size of recent down-days and squeezes the ratio into a 0–100 scale. When up-moves dominate, RSI is high. When down-moves dominate, it is low. (RSI has nothing to do with "relative strength vs the index", which is covered in Section 15. The name causes endless confusion.)

6.2 Calculation (14 periods, Wilder smoothing)

Wilder's smoothing is just an EMA with \(\alpha = 1/14\) (not \(2/15\)). Varsity's worked example uses a plain 14-day average of gains and losses (gains 29 points, losses 10 points → avg gain 2.07, avg loss 0.714, RS 2.90, RSI 74.35). That is exactly the seed step. Charting platforms then switch to Wilder's smoothing for every later day, so their RSI has a "memory" longer than 14 days.

6.3 Overbought and oversold

The classic reading is that RSI above 70 is overbought (a possible pullback) and below 30 is oversold (a possible bounce). Varsity's author prefers the stricter 80/20, which gives fewer but more extreme signals. Varsity also gives the opposite interpretation: if RSI stays overbought for a long time, that shows strong positive momentum, which is a reason to stay long, not to sell. The real data agrees with that more nuanced reading.

6.4 RSI ranges shift with the trend

A very useful practical idea, popularised by Andrew Cardwell and Constance Brown, is that in a strong uptrend RSI tends to oscillate between about 40 and 80 and rarely reaches 30. In a downtrend it oscillates between about 20 and 60 and rarely reaches 70. So in an uptrend, a dip to 40–45 is the "oversold" buy zone, and in a downtrend a rally to 55–60 is the "overbought" zone.

6.5 Divergence

A bullish divergence happens when price makes a lower low but RSI makes a higher low. The selling pushed price lower, but with less force than the previous wave. A bearish divergence is a higher high in price with a lower high in RSI. Divergence is a warning that momentum is fading. It is not an entry signal by itself, because divergences can repeat two or three times before price actually turns.

7. MACD: trend and momentum in one

7.1 Construction

Gerald Appel's Moving Average Convergence Divergence measures the gap between a fast and a slow EMA. When the 12-day EMA pulls away above the 26-day EMA, upward momentum is building (the averages diverge). When they come back together, momentum is fading (they converge).

MACD is in price units (rupees for a stock, points for Nifty), so a value of 100 on a ₹3,000 stock is not comparable with 100 on a ₹300 stock. Compare MACD only with its own history, or divide by price.

7.2 Signals

• Signal-line crossover (Varsity's main signal): MACD crosses above the signal line → bullish. Crosses below → bearish. Equivalent to the histogram changing sign.
• Zero-line (centreline) crossover: MACD crosses above 0 → the 12-EMA is now above the 26-EMA, so the medium-term trend has turned up. Slower but more reliable.
• Histogram: shrinking bars mean momentum is fading before the crossover happens. That makes the histogram the "earliest" part of MACD.
• Divergence: same idea as RSI. Price makes a lower low while the MACD line or histogram makes a higher low.

8. Bollinger Bands: volatility envelopes

8.1 Construction

John Bollinger's bands place an envelope around a 20-day SMA at ±2 standard deviations of the last 20 closes. Standard deviation (SD) measures how spread out the closes are. When the stock is calm the bands squeeze together, and when it is volatile they widen.

%B tells you where price sits inside the bands: 0 = on the lower band, 0.5 = on the middle line, 1 = on the upper band, above 1 = outside the upper band. Bandwidth measures how wide the bands are relative to price, and a multi-month low in bandwidth is a "squeeze".

The logic: if closes were normally distributed, about 95% would fall within ±2 SD. Price is not normally distributed (it trends and has fat tails), so the bands are best read as "what is unusual relative to the last month", not as a probability statement.

8.2 Varsity's way: mean reversion in ranges

Varsity teaches Bollinger Bands as a mean-reversion tool. Price at the upper band is "expensive" relative to its average, so short it with a target at the middle line. Price at the lower band is "cheap", so buy it with a target at the middle line. Varsity also warns that this fails when the bands expand during a strong trend ("envelope expansion"). Bollinger's own guidance adds two more patterns, the squeeze and the band walk.

8.3 The squeeze, and the head-fake

Volatility tends to alternate: quiet periods are followed by active ones. A squeeze (bandwidth at a multi-month low) says a big move is getting more likely. It does not say in which direction. Traders wait for a close outside a band and then follow it. Bollinger himself warned about the head-fake: a first break one way that reverses into the real move the other way.

8.4 The band walk

In a strong trend price can "walk" along the upper band, closing near or above it day after day. Shorting the upper-band touch in that situation is exactly Varsity's "envelope expansion" failure.

9. Stochastic oscillator (brief)

George Lane's Stochastic asks: where is today's close within the high–low range of the last 14 days? In an uptrend, closes tend to be near the top of the recent range, and in a downtrend near the bottom.

Above 80 is overbought and below 20 is oversold. A buy signal is %K crossing above %D while below 20, and a sell signal is %K crossing below %D while above 80.

Use it for: timing entries inside a known range or a pullback within a trend. Do not use it with RSI and %B together. Section 18 shows the three are 0.83–0.89 correlated, so they mostly repeat each other.

10. ATR: measuring volatility for stops and position size

10.1 True range

Wilder's Average True Range measures how much a stock typically moves in a day, including gaps. The plain high-minus-low range ignores overnight gaps, which is a serious omission in India, where results, global news and overnight moves in GIFT Nifty (formerly SGX Nifty) cause frequent gap opens. True range uses the largest of three distances:

10.2 Using ATR for stops

A stop placed inside a stock's normal daily noise gets hit by randomness, not by being wrong. ATR tells you how wide that noise is. Varsity's rule: place the stop at least one ATR from entry (long: entry − ATR; short: entry + ATR). Swing traders commonly use 1.5–3 × ATR, and a chandelier stop trails the stop at "highest close since entry − 3 × ATR". (Chuck LeBeau's original uses the highest high and a 22-day ATR. Using closes, as we do here, is a common variant that is slightly less likely to be stopped by one intraday spike.)

10.3 ATR as a comparison tool

Raw ATR depends on the price level, so convert it to a percentage of price to compare stocks. On 30 Sep 2026, HAL's ATR was ₹110.36 (2.36% of its ₹4,668 price) and HDFCBANK's was ₹13.96 (1.97% of ₹708.70). A fixed "5% stop" is therefore about 2.1 ATRs for HAL but 2.5 ATRs for HDFCBANK. ATR-based stops adjust automatically for each stock's personality.

11. ADX and the DMI: is there a trend at all?

11.1 The idea

Wilder's Directional Movement system has three lines. +DI measures upward pressure (how much today's high exceeds yesterday's). −DI measures downward pressure (how much today's low undercuts yesterday's). ADX is a smoothed measure of how far apart +DI and −DI are. ADX tells you the strength of a trend, not its direction. A falling market with strong selling has a high ADX too.

Varsity's trade rule: buy when ADX is above 25 and +DI is above −DI. Sell or short when ADX is above 25 and −DI is above +DI. Stop at the low (or high) of the signal candle.

12. Supertrend

Supertrend (by Olivier Seban) is an ATR-based trailing line. In an uptrend it sits below price as a green line and only moves up. When price closes below it, the line flips above price, turns red, and from then on only moves down. It is a trend-following system and a trailing stop in one.

Settings: Kite's default, which Varsity uses, is period 7, multiplier 3, and Varsity suggests keeping the multiplier between 3 and 4. TradingView's built-in default is period 10, multiplier 3. A larger multiplier means fewer flips and later exits. The examples here use (10, 3), so your Kite chart with (7, 3) will flip on slightly different days.

13. Volume indicators: volume average, OBV, VWAP

13.1 Volume vs its 10-day average (Varsity's method)

Volume is the number of shares traded. A trade where one person buys 100 shares from another adds 100 to volume, not 200. Raw volume means little on its own, so Varsity compares today's volume with the average of the previous 10 sessions. Volume above that average suggests large participants ("smart money") are involved.

13.2 On-Balance Volume (OBV)

Joseph Granville's OBV keeps a running total. Add the day's volume on up-closes, subtract it on down-closes, and ignore flat days.

The absolute level of OBV is meaningless. Only its direction matters. Rising OBV with rising price confirms the trend. Price making new highs while OBV does not is a volume divergence and a warning.

13.3 VWAP: intraday, rolling and anchored

Volume-Weighted Average Price is the average price paid per share, weighted by how many shares traded at each price. It answers the question "where is the average participant's cost?"

• Intraday VWAP (the version Varsity describes) restarts every day at 9:15 am. Price above VWAP means the intraday trend is up, and institutions judge their execution against it (a buy below VWAP is a "good fill"). It is an intraday tool. Since this course does not target intraday trading, you will mostly use the two variants below.
• Rolling VWAP (e.g., 20-day): the volume-weighted average over a moving window, roughly the average cost of everyone who bought in the last month. Price below it means the recent buyers are, on average, sitting on losses.
• Anchored VWAP (AVWAP): the cumulative VWAP starting from a chosen event, such as a major low, a breakout, or a results day. It shows the average cost of everyone who has bought since that event. In uptrends, pullbacks often hold near the AVWAP from the last major low.

14. Donchian channels

Richard Donchian's channel is the simplest trend tool of all: the highest high and lowest low of the last n days. A close above the prior 20-day high is a breakout, and the famous "Turtle" traders of the 1980s used 20- and 55-day versions. It connects directly to Chapter 3's support and resistance. The channel is the most recent range.

(We use the previous n days, so today's close can break the channel.) A classic swing rule is: buy on a close above the 20-day high, and exit on a close below the 10-day low.

15. Relative strength vs Nifty (RS line and Mansfield RS)

For stock selection, this may be the most useful tool in the chapter. The question is not "is this stock going up?" but "is it going up more than the market?" Leadership tends to persist for months (the momentum effect, documented in many markets, including India). Stocks that outperform in a weak market are often the first to run when the market turns.

The RS line rises when the stock outperforms Nifty, even if both are falling. Mansfield RS (named after the Mansfield chart service and popularised by Stan Weinstein, using weekly data) compares the RS line with its own 52-week average. Above zero means the stock is outperforming its normal relationship with the index. Crossing from negative to positive is a classic early leadership signal.

Over the same six months (31 Mar – 30 Sep 2026), Nifty returned +1.3%. Notice how the IT sector (WIPRO, INFY, TCS, HCLTECH) clusters at the bottom. RS ranking reveals sector rotation without reading a single news article. Also notice that ETERNAL, HAL and DIXON were rising in RS (from +2.4, +2.6 and −4.8 thirteen weeks earlier), while ADANIENT's was falling from +37.5. Both the level and the direction matter.

16. Other indicators in Varsity's list

Varsity's "Other indicators" chapter covers a few more tools. You should recognise them, but none adds a new kind of information beyond the families above.

17. Context test: the same RSI signal across 37 stocks

Books often state that "RSI below 30 is a buy in an uptrend but a trap in a downtrend". Instead of taking that on faith, we tested it on the 37 stocks in our data, 2017–2026.

• Signals: the day RSI(14) crosses below 30 ("oversold"), and the day it crosses above 70 ("overbought").
• Context: uptrend = close above the 200-day SMA and 50-day SMA above the 200-day SMA. Downtrend = both below.
• Outcome: the return over the next 10 and 20 trading days. Windows touching corporate-action artefacts (CDSL listing, TMPV demerger, TRENT bonus) were excluded. No costs.

18. Combining indicators without redundancy

18.1 Most oscillators measure the same thing

Beginners often pile RSI, Stochastic, CCI, Williams %R and MACD onto one chart and feel reassured when they "all agree". They agree because they are built from the same closes over similar windows. Here is the correlation of daily readings across all 37 stocks since 2018:

• RSI vs %B: 0.89. RSI vs Stochastic: 0.83. RSI vs distance from 50-SMA: 0.86. Three momentum readings that are close to one reading.
• ADX vs RSI: 0.15. ATR% vs RSI: −0.17. These answer genuinely different questions (trend strength, volatility).

18.2 One trend + one momentum + one volatility tool

Varsity puts indicators last in its decision hierarchy: (1) a recognisable candlestick pattern, (2) support/resistance for entry and stop, (3) volume above average, and only then (4) indicator confirmation. If the indicators agree, take the full position. If they don't, trade smaller or skip, but do not reject an otherwise valid setup just because of an indicator, and do not take a trade on an indicator alone.

19. Try it yourself

These snippets need a copy of the course repository and Python with pandas and numpy (setup steps are in its framework/README.md). Run them from the repository's root folder, for example with .venv/bin/python, because the file paths are relative to that folder. Each snippet is self-contained and reproduces numbers from this chapter.

19.1 SMA, EMA and Wilder RSI on SBIN

import pandas as pd

def load(sym):
    df = pd.read_csv(f"data/ohlc/{sym}.csv", index_col=0, parse_dates=True)
    return df[~((df.Volume == 0) & (df.High == df.Low))]   # drop holiday placeholder bars

def wilder(s, n):
    # seed with the simple mean of the first n values, then (prev*(n-1) + x) / n
    s = s.dropna(); out = pd.Series(index=s.index, dtype=float)
    out.iloc[n-1] = s.iloc[:n].mean()
    for i in range(n, len(s)):
        out.iloc[i] = (out.iloc[i-1]*(n-1) + s.iloc[i]) / n
    return out

df = load("SBIN")
c = df.Close
sma20 = c.rolling(20).mean()
ema20 = c.ewm(span=20, adjust=False).mean()     # alpha = 2/(20+1)
d = c.diff()
avg_gain = wilder(d.clip(lower=0), 14)
avg_loss = wilder((-d).clip(lower=0), 14)
rsi14 = 100 - 100 / (1 + avg_gain / avg_loss)
print(pd.DataFrame({"close": c, "sma20": sma20, "ema20": ema20, "rsi14": rsi14}).tail(3).round(2))
# 2026-09-30: close 959.5, sma20 992.47, ema20 992.55, rsi14 33.72

19.2 MACD, Bollinger %B, ATR and position size on HAL

import pandas as pd
df = pd.read_csv("data/ohlc/HAL.csv", index_col=0, parse_dates=True)
df = df[~((df.Volume == 0) & (df.High == df.Low))]
c = df.Close
macd = c.ewm(span=12, adjust=False).mean() - c.ewm(span=26, adjust=False).mean()
signal = macd.ewm(span=9, adjust=False).mean()
mid = c.rolling(20).mean(); sd = c.rolling(20).std(ddof=0)
upper, lower = mid + 2*sd, mid - 2*sd
pct_b = (c - lower) / (upper - lower)
prev = c.shift(1)
tr = pd.concat([df.High - df.Low, (df.High - prev).abs(), (df.Low - prev).abs()], axis=1).max(axis=1)
atr14 = tr.ewm(alpha=1/14, adjust=False).mean()          # Wilder smoothing
day = "2026-07-21"
print(round(pct_b[day], 3), round(atr14[day], 2))        # 1.097 95.92
entry, stop = c[day], c[day] - 2*atr14[day]
print("qty for 1% of Rs 5 lakh:", int(5000 // (entry - stop)))   # 26

19.3 Rank stocks by Mansfield relative strength

import pandas as pd, glob, os
nifty = pd.read_csv("data/ohlc/IDX_NSEI.csv", index_col=0, parse_dates=True).Close
rows = []
for f in glob.glob("data/ohlc/*.csv"):
    sym = os.path.basename(f)[:-4]
    if sym.startswith("IDX_") or sym in ("TRENT", "TMPV"):   # skip indices and adjustment artefacts
        continue
    c = pd.read_csv(f, index_col=0, parse_dates=True).Close
    rs = (c / nifty.reindex(c.index).ffill()).resample("W-FRI").last().dropna()
    mansfield = (rs / rs.rolling(52).mean() - 1) * 100
    rows.append((sym, round(mansfield.iloc[-1], 1)))
print(sorted(rows, key=lambda r: -r[1])[:5])
# [('ADANIENT', 23.8), ('ETERNAL', 23.1), ('TITAN', 15.7), ('HAL', 13.8), ('DIXON', 12.5)]

Exercises with code: (a) change the RSI lookback to 7 and to 21 and count how many days SBIN spent above 70 in 2025–26; (b) rerun the Section 17 test with a 2-day RSI below 10 (the full script is course/scripts/ch04_figures.py, function context_test); (c) compute the Supertrend on your own favourite stock with multipliers 2, 3 and 4 and count the flips.

20. Practical checklist

• I identified the regime first: price vs 200-SMA (and 50 vs 200), ADX above or below 20–25, and the Nifty's own regime.
• I am using a tool that suits that regime: trend tools (MA, Supertrend, Donchian, MACD) in trends, range tools (Bollinger fades, Stochastic, RSI 30/70) only when ADX is low and bands are flat.
• I use at most one indicator per job: trend, momentum, volatility (plus optional volume and RS). No three oscillators saying the same thing.
• In uptrends I read RSI in the 40–80 frame (40–45 = pullback zone). In downtrends, in the 20–60 frame.
• I treat divergences and squeezes as alerts, and I wait for price confirmation (a higher low, a breakout close, volume).
• Breakouts: volume is above the 10-day average (ideally 1.5× or more). If not, I trade smaller or skip.
• My stop is based on structure and at least 1–2 × ATR away from entry. My position size = risk budget ÷ (entry − stop).
• For futures, I checked that (stop distance × lot size) fits within my per-trade risk. If not, I don't trade.
• I shortlisted stocks with positive and rising relative strength vs Nifty.
• I used the weekly chart for the big picture and the daily chart for timing.
• I have not tuned settings to fit one chart. Any rule I rely on has been tested across many stocks and periods, net of costs (Chapter 12).
• I know my indicator values may differ slightly from my broker's chart, and that is fine.

21. Key takeaways

22. Quiz and exercises

23. Sources

Primary (concepts, explained here in our own words): Zerodha Varsity, Module 2, Ch 12: Volumes · Ch 13: Moving Averages · Ch 14: Indicators (Part 1) — momentum, RSI, leading vs lagging · Ch 15: Indicators (Part 2) — MACD, Bollinger Bands, avoiding indicator overload · Ch 20: Other indicators — ADX, Alligator, Aroon, ATR, ATR bands, Supertrend, VWAP · Ch 22: The Central Pivot Range · Module 2 index.
Original indicator definitions: J. Welles Wilder, New Concepts in Technical Trading Systems (1978) for RSI, ATR, ADX/DMI and Wilder smoothing; Gerald Appel (MACD); John Bollinger, Bollinger on Bollinger Bands (2001) for %B, bandwidth, squeeze and band walk; George Lane (Stochastic); Joseph Granville (OBV); Richard Donchian (channels); Stan Weinstein, Secrets for Profiting in Bull and Bear Markets (1988) for Mansfield relative strength. Supertrend defaults: Kite (7,3) as stated in Varsity Ch 20; TradingView built-in (10,3).
Data: NSE daily OHLCV via Yahoo Finance (course dataset, snapshot to 30 Sep 2026; zero-volume holiday placeholder bars removed); NSE F&O UDiFF bhavcopy 30 Sep 2026 for the HAL futures lot size (150). All figures and numbers are reproducible with course/scripts/ch04_figures.py (indicator code in course/scripts/ch04_indicators.py).