Swing Trading Playbook

1. What you'll learn and why it matters

Chapters 2 to 4 gave you the parts: candlesticks, support and resistance, trends, chart patterns, moving averages, RSI, ATR and volume. This chapter puts the parts together into a playbook: a written process you can follow every evening, in about 30 to 60 minutes, without watching the screen during market hours.

By the end you will be able to:

• Decide, from three numbers, whether the market is friendly to new long trades (the regime).
• Pick strong sectors and strong, liquid stocks (relative strength).
• Trade four concrete setups with exact rules for entry, stop, size and exit.
• Size every position so that one loss costs about 1% of your account.
• Manage trades with partial profits, breakeven stops, trailing stops and time stops.
• Place the orders with GTT (Good Till Triggered) so you don't need to watch prices all day.
• Keep a trade journal, run a weekly review and judge a setup by its numbers, not by memory.

2. What swing trading is (and isn't)

Swing trading means holding a position for a few days to a few weeks to capture one "swing": one leg of a trend, or a bounce from a level. Varsity's Module 2 finale also calls this positional trading and recommends it to beginners over day trading, because you decide with end-of-day (EOD) data, calmly, after the close.

The practical frame (from Varsity, adapted)

• Timeframe: daily candles (EOD). The higher the timeframe, the more reliable a signal tends to be.
• Lookback: look at 6 to 12 months of daily candles when judging a setup. Extend to about 2 years when you mark support and resistance.
• Opportunity universe: don't scan 2,000 stocks. Build a fixed list of liquid names you know. Varsity suggests the Nifty 50 as a starting universe, or a minimum volume of around 5 lakh shares a day. This course uses a 37-stock list of large, liquid NSE names.
• Time commitment: one focused evening session (Section 17) plus a weekly review on the weekend (Section 16).

What it costs, and how it is taxed (briefly)

Swing trades are delivery trades (product type CNC on Zerodha): you buy, the shares reach your demat account and you sell days later. As of Zerodha's published charges (checked October 2026), equity delivery has zero brokerage, but statutory charges still apply: STT (Securities Transaction Tax) 0.1% on both buy and sell, NSE transaction charges 0.00307%, SEBI fee Rs 10 per crore, GST 18% on (transaction + SEBI charges), stamp duty 0.015% on the buy side, and a DP charge of Rs 15.34 per stock per day on which you sell. On a Rs 1.1-1.25 lakh position, the worked examples in this chapter come to 0.23-0.26% of the buy value for the round trip: STT alone is 0.2%, and two partial sells mean two DP charges. The backtests use a flat 0.24%. Section 9 works out one example line by line.

Gains on listed shares held 12 months or less are short-term capital gains, taxed at 20% plus cess for FY 2026-27 (unchanged since 23 July 2024). Very frequent trading can also raise the question of whether your income is business income. Chapter 13 covers costs and tax in detail. In this chapter, costs are included in every R-multiple (the profit or loss of a trade measured in units of its initial risk, defined in Section 7). Taxes are not.

3. The top-down process in one picture

A professional swing trader doesn't start with a hot stock. They start with the market and narrow down. Every trade must pass eight gates, in order:

The logic: most stocks move with the market most of the time. Varsity's Dow Theory chapters make the same point with the primary trend: fighting it lowers your odds. So the first question is "is the tide coming in?" Then "which boats are rising fastest?" (sectors and relative strength). Only then "is there a low-risk place to step on board?" (the setup, trigger and stop). Finally, "how big, and how do I get off?" (size and management).

4. Gate 1: market regime

The regime answers one question: how much risk should I take on new long trades today? We use three measures, all computed from end-of-day data.

4.1 Nifty 50 versus its 200-day moving average

Plain language: the 200-day simple moving average (200-DMA) is the average closing price of the last ~10 months. If the index is above it, and the average itself is rising, the long-term trend is up.

The logic: the 200-DMA is a slow, widely watched line. When price is above it, the average investor who bought in the last 10 months is in profit, so dips tend to get bought. Below it, rallies tend to get sold by people who want to get out at breakeven. It does not predict tomorrow. It describes the "weather".

4.2 Breadth: how many stocks are participating?

An index can be held up by a handful of giants while most stocks fall. Breadth counts how many stocks in your universe are above their own moving averages.

On 30 Sep 2026 only 7 of 37 stocks (18.9%) were above their 200-DMA, and only 3 of 37 (8.1%) were above their 50-DMA. That is a very weak tape. When fewer than 1 in 5 stocks are in a long-term uptrend, buying setups have little tailwind.

4.3 India VIX: the market's fear gauge

India VIX is computed by NSE from Nifty option prices. It estimates the annualised volatility the market expects over the next 30 days. Roughly, a VIX of 13 implies typical daily Nifty moves of about \(13/\sqrt{252} \approx 0.8\%\). On 30 Sep 2026 VIX was 13.49, close to its 1-year median of 12.6. So the current weakness is a slow grind, not a panic. In March 2026 VIX hit 27.89, and daily swings and overnight gaps were much larger.

How we use VIX: not as a buy or sell signal, but as a size dial. Higher VIX means wider stops and more gap risk. Because our stops are ATR-based, a volatile market automatically gives smaller share counts. On top of that, many traders halve their per-trade risk when VIX is above ~18.

4.4 Turning the three numbers into a rule

Does the regime filter actually help? Section 13 tests it on 2,192 historical trades. In short: trend-following setups (S1 pullback, S2 breakout) earned a small positive average when Nifty was above its 200-DMA and lost money when it was below. Setup 3, a mean-reversion bounce, behaved the opposite way. Nifty was above its 200-DMA on 76.1% of trading days from 2017 to 2026, so the filter keeps you trading most of the time and steps you aside in the worst stretches.

5. Gate 2: sector strength

Plain language: money rotates. In any quarter some sectors lead (more buyers) and others lag. A good setup in a leading sector has two tailwinds. The same setup in a lagging sector is swimming upstream.

The logic: institutions allocate by sector and their buying happens over weeks, which creates persistent trends at the sector level. Varsity's Module 10 momentum chapter builds on the same observation: what has outperformed over the past months tends, on average, to keep outperforming for a while.

How to use it: rank sectors by 3-month RS every weekend. Prefer longs in the top third and avoid the bottom third. Where 3-month and 6-month RS disagree (IT here), the shorter number is often a bounce and the longer one is the trend. Your universe may not map neatly to NSE sectors (our 37 stocks include just one pharma name), so also look at the stocks themselves (next gate).

6. Gate 3: stock selection (liquidity and relative strength)

6.1 Liquidity filter

You must be able to get in and, more importantly, out at a fair price. Varsity suggests a tight bid-ask spread or a minimum daily volume (about 5 lakh shares). A cleaner test is turnover, because 5 lakh shares of a Rs 150 stock and of a Rs 12,000 stock are very different amounts of money.

On 30 Sep 2026 all 37 stocks passed. The most liquid was HDFCBANK at about Rs 2,069 crore a day. The least liquid was IRCTC at about Rs 28 crore. For a Rs 5 lakh account, liquidity is never the constraint in these names. It starts to matter with mid- and small-caps, where a Rs 1 lakh order can move the price.

6.2 Relative strength (RS) of the stock

We want the stocks that are stronger than the market: they fall less in declines and rise more in rallies. We score each stock on its 3-month (63 sessions) and 6-month (126 sessions) return minus Nifty's over the same window, then average the two.

How to use it: each weekend, rank your universe and keep a focus list of the top 10-12 names. During the week, look for setups mainly in the focus list. This mirrors Varsity's momentum-portfolio idea (rank by past return, keep the top dozen) at a swing-trading scale.

7. Gates 4-7: setup, trigger, stop and position size

7.1 Setup versus trigger

A setup is a chart condition that makes a move likely enough to watch, for example "price has pulled back to the 20 EMA in an uptrend". A trigger is the price event that says "the move has started, act now", for example "price trades above yesterday's high". Separating the two stops you buying falling knives: if the bounce never comes, the trigger never fires and you never enter.

7.2 The initial stop: where am I wrong?

The stop goes where your idea is proven wrong, not where the loss "feels" acceptable. Below the pullback low, below the breakout bar, below support. We add a small buffer of a fraction of ATR(14) so that ordinary noise doesn't take you out. ATR is the average true range from Chapter 4.

\(R\) is the unit everything else is measured in. A trade that makes \(2\times\) the per-share risk is a +2R trade. A full stop-out is −1R. Thinking in R lets you compare a Rs 300 stock with a Rs 14,000 stock.

7.3 Position size: the most important calculation in this course

Varsity's Module 9 opens its position-sizing chapters with a poker story: the author lost his buy-in three times in 25 minutes by betting too big after losses. The lesson is that position sizing, not stock picking, decides whether you survive. Losses compound against you:

The fix is fixed-fractional risk: decide how much of the account you will lose if the stop is hit (we use 1%), then let the stop distance set the number of shares. Varsity calls this family "percentage risk" or "percentage volatility" sizing, following Van Tharp.

The second term is an exposure cap: no single stock may be more than 25% of the account. Without it, a very tight stop would tell you to put the whole account in one stock, and one overnight gap would then cost far more than 1%.

7.4 Which "capital"? Varsity's three equity models

When positions are already open, what number do you multiply by 1%? Varsity's Module 9 (after Van Tharp) lists three models:

• Core equity: starting capital minus money already committed to open trades. Conservative and simple.
• Total equity: cash plus the value of open positions including unrealised P&L. Aggressive: it counts paper profits before they are banked.
• Reduced total equity: core equity plus only the profit that is locked in by a trailing stop. Varsity's author prefers this one, and so do we. It rewards you for trailing stops.

Varsity also describes sizing by a fixed number of shares or lots per lakh (simple, but ignores risk), by a percentage of margin (useful for futures), and by percentage volatility: shares = (capital × x%) ÷ ATR, so each position contributes the same daily rupee swing. For ETERNAL on 30 Sep 2026, ATR(14) was 8.15. At a 1% volatility budget that gives 5,000 ÷ 8.15 = 613 shares. Compare that with 358 shares from our stop-based method in Section 17. Both are reasonable. Pick one method and use it every time.

7.5 Portfolio heat

Each trade risks ≤1%, but trades are correlated: in a sell-off, all your longs fall together. So cap the total open risk (the sum of (entry − current stop) × qty over all positions) at about 5-6% of capital, hold at most 2 positions per sector, and count a position whose stop is at breakeven as zero heat.

8. Gate 8: trade management and exits

Varsity's advice is blunt: once a well-checked trade is placed, do nothing until the target or the stop is hit, except trail the stop, which is "a healthy practice". We turn that into four mechanical rules that apply to every setup (Setup 3 uses a fixed target instead of rules 1-2):

1. Partial profit at +2R: sell half the position when price reaches Entry + 2R. This banks a gain that already pays for about two losers.
2. Breakeven stop: after the partial, move the stop on the remaining half to the entry price. The rest of the trade is now a free ride, apart from gap risk.
3. Trailing exit for the rest: for daily setups, exit at the next open after a daily close below the 20 EMA. For the weekly setup, exit after a weekly close below the 10-week EMA. Using closes instead of intraday lows avoids being shaken out by one bad hour.
4. Time stop: if price has not reached +1R within 10 sessions (15 for Setups 3 and 4), exit at the close. A swing trade that goes nowhere ties up capital and attention. A maximum hold (60 sessions for daily setups, 120 for the weekly one) forces a fresh look.

9. Setup 1: trend pullback to the 20 EMA

The idea: in an established uptrend, buy a temporary dip to a rising short-term average rather than chasing strength. You get a better price and a nearby, logical stop.

The logic: trends move in waves. After a push higher, early buyers take profits and price drifts back toward its "fair" short-term level, the 20 EMA. Buyers who missed the first leg wait there. If the trend is healthy, they step in and price turns back up. If the dip goes much deeper (below the pullback low), the trend may be ending, and we want out.

Exact rules (S1)

Setup 1 statistics (37 stocks, 2017-Sep 2026, net of costs): 906 trades, win rate 38.3%, average +0.080R, profit factor 1.13 (total R won ÷ total R lost; above 1 means the winners outweighed the losers), average hold 13.8 sessions. With Nifty above its 200-DMA: +0.124R (767 trades). Below: −0.166R (139 trades). In 2025-26 alone: 121 trades, win rate 24.0%, −0.399R. A choppy market is the pullback trader's worst enemy, because every "dip to the EMA" keeps going.

10. Setup 2: breakout to a 52-week high on volume

The idea: when a stock closes above the highest price of the past year, on heavy volume, after a period of sideways movement (a "base"), buy the next morning.

The logic: the 52-week high is a resistance level everyone can see (Chapter 3). Sellers who bought near it have been waiting to "get out at cost". A close above it on unusual volume means demand has absorbed all of that supply. With nobody left holding at a loss above, there is little overhead resistance. Volume, as Varsity's volume chapter stresses, is the confirmation that big participants are involved.

Exact rules (S2)

Setup 2 statistics: 418 trades, win rate 37.8%, average only +0.012R (gross +0.073R before costs), profit factor 1.02. It was positive in 2021-23 (+0.176R) and clearly negative in 2024-26 (−0.268R, win rate 29.6%). On our large-cap list, plain 52-week-high breakouts were roughly break-even after costs. That doesn't make breakouts useless. It means the edge, if any, comes from selectivity (base quality, sector strength, volume quality) and from the occasional MARUTI-style runner.

11. Setup 3: support bounce with a bullish candle (the Varsity checklist)

This setup is a direct translation of the trading checklist in Varsity Module 2's finale chapter. In Varsity's words, the checklist is "fairly stringent", so few trades qualify, but conviction is high when one does. Here are its items, and how we made each one exact:

Entry: buy-stop above the pattern high, valid for 2 sessions. Exit: the full position at the target or at the stop. The time stop applies if price has not reached +1R within 15 sessions, and the maximum hold is 40 sessions. There is no partial exit or trailing here, because the target is a real resistance level.

The logic: support is a price where buyers showed up before. A bullish candle there, on decent volume, shows they are showing up again. The stop sits just under the pattern, so it is close and the reward to the next resistance can be several times the risk. The weakness: you are buying in a short-term downtrend, so when support breaks, it can break hard.

Setup 3 statistics: 172 trades (the checklist is strict!), win rate 40.7%, average +0.150R, profit factor 1.23, average hold 6.5 sessions. Opposite to the trend setups: with Nifty below its 200-DMA it averaged +0.509R (58 trades, 51.7% wins), and above it −0.033R (114 trades). 2025-26: 48 trades, +0.081R. Adding a "stock above its 200-SMA" filter made it worse (74 trades, +0.099R). The bounces that worked best came in washed-out markets. Treat these splits with care: 58 trades is a small sample.

12. Setup 4: weekly trend + daily trigger

The idea: use the weekly chart to decide direction (only stocks in a weekly uptrend) and the daily chart to time the entry after a short pullback. This is the multiple-timeframe approach Varsity hints at when it says higher timeframes give more reliable signals. It is also the most "positional" of the four setups.

The logic: weekly bars filter out daily noise, so a weekly uptrend reflects the bigger players' sustained buying. Inside it, short pullbacks (daily RSI dipping below 45) are pauses. A daily close above the prior two highs shows buyers are back. Because the trailing exit is weekly, the trade can last for months, which is where large R-multiples come from.

Exact rules (S4)

Setup 4 statistics: 696 trades, win rate 41.4%, average +0.293R, profit factor 1.53, average hold 27.5 sessions, best trade +21.57R. It was the best of the four, and its edge came from a few long-held trends (the median trade was −0.68R). Regime: +0.323R above the 200-DMA (588 trades) vs +0.132R below (108). 2025-26: 104 trades, 30.8% wins, −0.162R.

13. How the four setups did, 2017-2026 (honest numbers)

Varsity's Module 10 opens by defining a trading system: an approach you can write down as a process, with inputs, rules and outputs, that can be measured and backtested. Anything you can't define as a process isn't a system. Here we measure our four.

Method: the rules above, applied mechanically to 37 NSE stocks with daily data from Oct 2016 to 30 Sep 2026 (trades start in 2017 once the 200-day averages exist). One open trade per stock per setup. Costs of 0.24% of the entry value are deducted from every trade. When a bar touches both the stop and the target, we assume the stop was hit first. Gaps through stops fill at the open.

Reading the numbers: expectancy

For S4: \(0.414 \times 2.06 + 0.586 \times (-0.95) = 0.85 - 0.56 = +0.29R\) per trade. Risking Rs 5,000 per trade (full 1%), that's about Rs 1,465 per trade on average, before tax, across hundreds of trades. Any single month can be far worse. S1 and S4 both had 16 losses in a row at some point. Varsity calls the belief that "after 6 losses the 7th must win" the gambler's fallacy: each trade has the same odds as the first. The antidote is fixed position sizing, not bigger bets after losses.

Caveats: why these numbers are probably too good

• Survivorship bias. Our 37 stocks are today's large, successful names. Stocks that collapsed and left the index over 2016-2026 aren't in the list. Long-only trend setups on today's winners look better than they would have looked in real time.
• Correlated trades. The 2,192 trades are not independent. Many happen in the same weeks, so the true range of outcomes is wider than the trade count suggests.
• Fills. We assumed fills exactly at trigger prices (except gaps). Real GTT limit orders may fill a little worse or not at all. Costs are included, but taxes and slippage beyond 0.24% are not.
• Data quirks. Yahoo daily data. We back-adjusted the Tata Motors demerger (Oct 2025), a Trent price rescale (Jan 2026) and early CDSL data. The 1 Feb 2026 Budget Sunday session is missing. Zero-volume holiday rows were dropped.
• Regime dependence. 2021-23 was excellent and 2024-26 poor. A system that "worked" over ten years can lose for two straight years. That is normal, and it's the hardest part to live through.
• Not optimised (good), but sensitive to choices. Parameters were set before testing. As a robustness check, forcing every stop to be at least 1 × ATR away barely changed anything: S1 +0.073R, S2 +0.004R, S3 +0.099R, S4 +0.295R. That suggests the results don't hinge on one lucky parameter.

14. Event risk: results days, gaps and the Budget

Swing trades are held overnight, so they are exposed to news that arrives when the market is closed. The biggest scheduled source is quarterly results. Under SEBI's LODR Regulations, listed companies must publish quarterly results within 45 days of the quarter end (60 days for the annual results) and must announce the board-meeting date in advance. That gives you a calendar to plan around.

Rules for results and other events

• Know the date before you enter. Check the company's board-meeting announcement on the NSE corporate filings page. Your broker app also shows upcoming results.
• Don't open a new swing trade within ~5 sessions of results unless you accept a possible 2R-3R loss. Or enter at half size.
• Already in a trade? If it's at +2R with half booked and the stop at breakeven, holding through results is reasonable: the worst realistic case is a small loss on the remaining half. If it's still near entry, many traders exit before results, or cut to half.
• Market-wide events: the Union Budget, RBI policy, election results and global shocks can all gap every stock at once. ONGC (Section 10) is our example.

Upcoming results in our universe (estimated dates from Yahoo Finance; confirm on NSE): TCS ~8 Oct 2026, DMART ~10 Oct, HCLTECH ~12 Oct, ETERNAL / WIPRO / POLYCAB / NESTLEIND / HDFCAMC ~15 Oct, RELIANCE / DIXON ~16 Oct, the large private banks ~17 Oct, INFY ~23 Oct, and LT, ITC, MARUTI, ASIANPAINT, COALINDIA and others in late October. Any trade opened in early October 2026 is very likely to run into results.

15. Placing the orders: GTT mechanics

A normal stop-loss order on an Indian exchange is a day order: it expires at the close. For multi-day trades you would have to re-enter it every morning. Brokers solve this with GTT (Good Till Triggered) orders. Per Zerodha's GTT support documentation (accessed Oct 2026):

• A GTT is a trigger stored by the broker, not at the exchange. When the last traded price crosses your trigger during market hours, the broker sends an order to the exchange: either a limit order at your chosen limit price, or a market order with market protection (a market order that won't fill beyond a set band from the last price). For stocks in the F&O segment, GTTs on the stock trigger only until 3:15 pm.
• Single GTT: one trigger (buy or sell). OCO (one-cancels-other) sell GTT: a stop trigger and a target trigger on holdings. When one fires, the other is cancelled. Kite also offers an optional trailing stop-loss on GTTs (the stop trigger moves up by a fixed number of points as price rises). Our trailing rule is based on daily closes versus the 20 EMA, so we trail by hand instead.
• Valid for up to 365 days or until triggered. A GTT fires even if the stock opens beyond the trigger on a gap. A trigger fires only once: if the resulting order doesn't fill by the close, it is cancelled like any day order and you must place a fresh one.
• No funds are blocked when you create a GTT. Margin and funds are checked at trigger time. Sell GTTs on your holdings must be authorised with your CDSL TPIN (a PIN from the depository that approves shares leaving your demat account).
• GTTs are cancelled when the stock has a corporate action (bonus, split, rights issue, amalgamation) whose impact is more than 2% of the market value on the ex-date or record date. Re-place them afterwards.
• Placing a GTT is free, and Zerodha allows up to 500 active GTTs per account. Other brokers have similar orders under different names. Check yours.

16. Trade journal and weekly review

Varsity's chapter on trading biases names attribution bias: winners feel like skill, losers get blamed on the broker, the market or bad luck. Its cure is a written journal. Record why you entered and exited, then compare plans with outcomes. Over months the journal shows you your real tendencies. Here is a template, filled in with four trades from this chapter:

Also save a chart screenshot at entry and at exit, and add a line on how you felt (anxious, overconfident, bored). Over time those notes reveal patterns, for example "I break rules after 3 losses".

The weekly review (weekend, ~30 minutes)

1. Numbers: trades closed, win rate, average R, total R this week and over the last 20 trades (a rolling expectancy). Plot your equity curve. Varsity's equity-curve chapter normalises a portfolio to a starting value of 100 and tracks it daily.
2. Rule audit: how many trades broke a rule (entry without trigger, stop moved, size too big, results ignored)? Rule-breaks are the most useful number in the journal.
3. Open risk ("heat"): add up (entry − current stop) × qty across open trades. Is it under ~5% of capital?
4. Regime and RS refresh: recompute the regime table, the sector ranking and the top-12 focus list for next week.
5. Calendar: note the results dates, RBI policy, Budget and expiry-week events coming up.
6. One lesson: write one sentence you'll apply next week.

17. The end-of-day routine (worked on 30 Sep 2026)

After the close (data is final around 4-6 pm), set aside 30 to 60 minutes and follow the same steps in the same order every day:

1. Update data (2 min): refresh your EOD data (framework.data.update_ohlc).
2. Manage open trades first (5-10 min): for each position check stop, 2R, trailing condition, time stop and upcoming results. Adjust GTTs.
3. Regime (2 min): Nifty vs 200-DMA, breadth, VIX. This sets today's risk per trade.
4. Scan (5 min): run the scanner on the focus list / universe for S1-S4 (python -m framework.scan).
5. Evaluate each candidate (5-10 min each): the pre-trade checklist (Section 20). Most candidates fail.
6. Plan and place (5 min): size, then create the buy GTT and write the plan in the journal before the trade happens.
7. Close the laptop. No watching intraday. The orders do the work.

18. The mind behind the playbook

Varsity's Module 9 closes with trading biases. Each one maps onto a rule in this playbook:

Chapter 11 goes deeper into risk of ruin, drawdowns and the psychology of losing streaks.

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 reproduces a number from this chapter.

(a) The regime line and breadth

import pandas as pd, glob, os
DATA = "data/ohlc"
nifty = pd.read_csv(f"{DATA}/IDX_NSEI.csv", index_col=0, parse_dates=True)
nifty["SMA200"] = nifty.Close.rolling(200).mean()
last = nifty.iloc[-1]
print(f"Nifty {last.Close:.2f} vs 200-DMA {last.SMA200:.2f} -> {(last.Close/last.SMA200-1)*100:+.2f}%")

above = []
for f in glob.glob(f"{DATA}/*.csv"):
    sym = os.path.basename(f)[:-4]
    if sym.startswith("IDX_"):
        continue
    d = pd.read_csv(f, index_col=0, parse_dates=True)
    d = d[d.Volume > 0]                       # drop stale holiday rows
    above.append(d.Close.iloc[-1] > d.Close.rolling(200).mean().iloc[-1])
print(f"Breadth: {sum(above)} of {len(above)} stocks above their 200-DMA = {sum(above)/len(above)*100:.1f}%")
vix = pd.read_csv(f"{DATA}/IDX_INDIAVIX.csv", index_col=0, parse_dates=True)
print("India VIX:", round(vix.Close.iloc[-1], 2))
# -> Nifty 22620.45 vs 200-DMA 24372.85 -> -7.19%
# -> Breadth: 7 of 37 stocks above their 200-DMA = 18.9%
# -> India VIX: 13.49

(b) Position size with the exposure cap

def position_size(capital, risk_pct, entry, stop, max_pos_pct=0.25):
    risk_rs = capital * risk_pct                 # e.g. 5,00,000 x 1% = 5,000
    per_share = entry - stop                     # rupees lost per share if stopped
    qty_risk = int(risk_rs // per_share)         # shares allowed by the risk budget
    qty_cap = int(capital * max_pos_pct // entry)   # shares allowed by the 25% exposure cap
    qty = min(qty_risk, qty_cap)
    return qty, qty * entry, qty * per_share

print(position_size(500_000, 0.01, 3762.00, 3651.45))   # TITAN
# -> (33, 124146.0, 3648.15...)

(c) Find Setup-1 bars on TITAN

import pandas as pd
from framework.indicators import ema   # EMA seeded with the first 20-day SMA, as in Chapter 4
d = pd.read_csv("data/ohlc/TITAN.csv", index_col=0, parse_dates=True)
d = d[d.Volume > 0]
d["EMA20"] = ema(d.Close, 20)
d["SMA50"] = d.Close.rolling(50).mean()
d["SMA200"] = d.Close.rolling(200).mean()
trend = (d.Close > d.SMA50) & (d.SMA50 > d.SMA200) & (d.SMA200 > d.SMA200.shift(20)) & (d.EMA20 > d.SMA50)
recent_high = d.High.rolling(50).max().rolling(10).max() <= d.High.rolling(10).max()
touch = d.Low <= d.EMA20 * 1.005
print(d[trend & recent_high & touch].loc["2025-10":"2025-11", ["Open", "High", "Low", "Close", "EMA20"]].round(2))
# -> 2025-11-03 3730.0 3762.0 3667.3 3724.5 3660.53   (our example's setup bar)
# -> 2025-11-26 ... (a second touch while already in the trade)

(d) Expectancy of a setup from the trade log

import pandas as pd
t = pd.read_csv("data/extra/ch05_trades.csv")
s4 = t[t.setup == "S4"]
win = (s4.R_net > 0).mean()
avg_win = s4.R_net[s4.R_net > 0].mean()
avg_loss = s4.R_net[s4.R_net <= 0].mean()
print(f"trades {len(s4)}  win {win:.1%}  avg win {avg_win:.2f}R  avg loss {avg_loss:.2f}R  "
      f"expectancy {win*avg_win + (1-win)*avg_loss:+.3f}R")
print(s4.groupby("regime_up").R_net.agg(["count", "mean"]).round(3))
# -> trades 696  win 41.4%  avg win 2.06R  avg loss -0.95R  expectancy +0.293R

The full engine (all four setups, the simulator and every figure) is in course/scripts/ch05_lib.py and course/scripts/ch05_figures.py. Change a rule there, re-run, and see how the statistics move. That is the first step toward building your own system in Chapter 12 and Chapter 17.

20. Pre-trade checklist

Print this. Every box must be ticked before a buy GTT is placed.

• Regime checked today: risk-on / mixed / risk-off, and the matching risk per trade (1% / 0.5% / ≤0.5% or none). VIX overlay applied.
• The stock's sector is in the top half by 3-month RS (or you have a written reason why not).
• The stock is in the top third of the universe by RS score, and passes the liquidity filter (≥ Rs 20 crore median daily turnover).
• The setup matches one of S1-S4 exactly (no "almost").
• Trigger price written down. You enter only if the trigger trades.
• Stop placed at a structural level with an ATR buffer. Stop distance noted in % and in R.
• For S3: support within 4% of the stop and RRR ≥ 1.5 to the nearest resistance.
• Position size computed: min(risk-based qty, 25% cap qty). Rupee risk is within budget.
• Portfolio heat after this trade ≤ ~5-6%. No more than 2 positions in this sector.
• No results, Budget or policy event within ~5 sessions (or size halved deliberately).
• Exit plan written: 2R partial level, trailing rule, time-stop date.
• Journal entry created before the order is placed, with a chart screenshot.
• You can accept this trade losing 1R, or 2R on a gap, without changing your next trade.

21. Key takeaways

22. Quiz and exercises

23. Sources

Zerodha Varsity (concepts explained in our own words): Module 2, ch. 19 "The Finale – Helping you get started" (timeframe, lookback, opportunity universe, two-part scan, trading checklist) · Module 2, ch. 11 Support and Resistance · Module 2, ch. 12 Volumes · Module 2, ch. 13 Moving Averages · Module 2, ch. 17-18 Dow Theory · Module 9, ch. 11 Position Sizing for active traders · Module 9, ch. 12 (equity models) · Module 9, ch. 13 (unit, margin and volatility sizing) · Module 9, ch. 6 Equity Curve · Module 9, ch. 15 Trading Biases · Module 9, ch. 16 Trading Biases (Part 2) · Module 10, ch. 1 What to expect (definition of a trading system) · Module 10, Momentum Portfolios.
Rules and rates (checked Oct 2026): Zerodha charges (delivery STT 0.1% each side, NSE txn 0.00307%, stamp 0.015%, DP Rs 15.34) · Zerodha: Introducing GTT · Zerodha Support: What is a GTT order (limits, validity, corporate actions) · Value Research: STCG/LTCG FY 2026-27 · ClearTax: STT rates after Budget 2026 · NSE circular: live trading session on 1 Feb 2026 (Budget day) · SEBI: LODR timelines for financial results.
Data: daily OHLCV for 37 NSE stocks, Nifty 50 and India VIX (NSE via Yahoo Finance, to 30 Sep 2026); NSE ind_close_all index files for 30 Sep 2025, 30 Mar, 30 Jun, 31 Aug and 30 Sep 2026 (data/extra/); quarterly results dates from Yahoo Finance via yfinance (data/extra/ch05_earnings_dates.csv). All statistics, trade logs and figures: course/scripts/ch05_figures.py + ch05_lib.py, outputs in data/extra/ch05_trades.csv and data/extra/ch05_results.json.