Risk: the only part you control

By the end of this lesson

The position-sizing formula, and why it must run from the stop backwards
R multiples and expectancy — the arithmetic that decides whether a method is viable
Why the mathematics of drawdown is unforgiving and asymmetric

The one thing

Every element of technical analysis is probabilistic. Risk management is not — it is arithmetic, entirely under your control, and it determines whether you survive long enough for the probabilistic parts to matter.

The sizing formula

Position size

Shares = (Account × risk% ) ÷ (Entry − Stop)

Account × risk%: the rupee amount you are prepared to lose on this trade
Entry − Stop: the risk per share, in rupees

The order matters. The stop is decided by the chart; the size is then computed from it. Deciding the size first and placing the stop wherever it happens to be affordable is the reversal that ends most accounts.

Worked exampleSizing three trades to identical risk

A ₹10,00,000 account risking 1% — ₹10,000 — per trade. Note how the position value varies by a factor of five while the risk stays constant.

Trade A — entry / stop: ₹420 / ₹396₹24 risk
Trade B — entry / stop: ₹1,180 / ₹1,090₹90 risk
Trade C — entry / stop: ₹86 / ₹79₹7 risk

Trade A: 10,000 / 24 = 417 shares → position ₹1,75,000

Trade B: 10,000 / 90 = 111 shares → position ₹1,31,000

Trade C: 10,000 / 7 = 1,429 shares → position ₹1,23,000

Three very different position sizes. Identical risk: ₹10,000 each.

A fixed ₹2,00,000 position in all three would have risked

₹11,400 / ₹15,300 / ₹16,300 — silently unequal, and largest

in exactly the trade with the widest stop.

So → Equal rupee positions do not mean equal risk. Sizing from the stop is what makes each position genuinely comparable, and it is the only way a win rate or an expectancy figure means anything.

R: the unit that makes everything comparable

Define 1R as the amount risked on a trade. A trade that hits its stop loses 1R. One that runs to three times the initial risk gains 3R. Expressing every outcome in R makes results comparable across instruments, position sizes and account values.

Expectancy

Expectancy (in R) = (win rate × average win in R) − (loss rate × average loss in R)

This single number determines whether a method makes money. A positive expectancy with a low win rate is completely normal — and is what most trend-following approaches actually look like.

Worked exampleTwo methods, and why the intuitive one is worse

Method A — win rate: 35%
Method A — average win / loss: +3.2R / −1.0R
Method B — win rate: 72%
Method B — average win / loss: +0.6R / −1.9R

Method A: (0.35 × 3.2) − (0.65 × 1.0) = 1.12 − 0.65 = +0.47R per trade

Method B: (0.72 × 0.6) − (0.28 × 1.9) = 0.43 − 0.53 = −0.10R per trade

Method A is wrong about two trades in three and makes money.

Method B is right about nearly three in four and loses.

So → Win rate on its own is not a measure of anything. Method B is the psychologically comfortable one — frequent small wins, occasional large losses — and it is the profile of an account that dies slowly. Method A requires tolerating long strings of losses, which is why most people cannot run it despite its being better.

Where the stop goes

Stop placement methods

Structural
Where: Below the most recent swing low, or below the base/spring low
When appropriate: The default. It's the price at which the thesis is false.
Volatility (ATR)
Where: 2–3 ATR below entry
When appropriate: When structure is far away, or on volatile instruments
Level-based
Where: Just below a strong support band
When appropriate: When a specific level is the reason for the trade
Time-based
Where: Exit if the move hasn't started within n sessions
When appropriate: Underused — capital tied up in a stalled setup has a real cost
Percentage
Where: A fixed % below entry
When appropriate: Weakest. Unrelated to the chart or the instrument's volatility.

Method	Where	When appropriate
Structural	Below the most recent swing low, or below the base/spring low	The default. It's the price at which the thesis is false.
Volatility (ATR)	2–3 ATR below entry	When structure is far away, or on volatile instruments
Level-based	Just below a strong support band	When a specific level is the reason for the trade
Time-based	Exit if the move hasn't started within n sessions	Underused — capital tied up in a stalled setup has a real cost
Percentage	A fixed % below entry	Weakest. Unrelated to the chart or the instrument's volatility.

The best stops satisfy two conditions simultaneously: they sit beyond ordinary noise (check against ATR) and at a price that falsifies the thesis (check against structure). If those two disagree, the setup is poorly located and the honest response is to skip it.

The drawdown arithmetic

Losses and the gains needed to recover them are not symmetric, and the asymmetry accelerates viciously. This table is worth memorising because it is the reason risk limits exist.

What it takes to get back to even

−10%
Gain required to recover: +11%
−20%
Gain required to recover: +25%
−33%
Gain required to recover: +50%
−50%
Gain required to recover: +100%
−75%
Gain required to recover: +300%
−90%
Gain required to recover: +900%

Drawdown	Gain required to recover
−10%	+11%
−20%	+25%
−33%	+50%
−50%	+100%
−75%	+300%
−90%	+900%

This is why professional risk limits look conservative to newcomers. Small losses are recoverable arithmetically; large ones are recoverable only in theory. A 50% drawdown does not require a good year to fix — it requires a doubling.

Worked exampleWhy 1% per trade, and not 5%

A losing streak of eight — unremarkable for a 35% win-rate method — at three different risk levels.

Risking 1% per trade: −7.7%recoverable
Risking 3% per trade: −21.6%painful
Risking 5% per trade: −33.7%needs +51% to recover
Risking 10% per trade: −57.0%structurally damaged

Eight consecutive losses at 35% win rate:

probability ≈ 0.65^8 ≈ 3.2% for any given run of 8 —

which means across a few hundred trades it WILL happen.

Position sizing must survive the streak you know is coming,

not the average outcome you hope for.

So → Size for the streak, not the average. The purpose of a small per-trade risk is not timidity — it is ensuring that a normal, expected sequence of losses leaves you able to take the next trade with a clear head.

Per-trade risk: commonly 0.5–1% of capital for a developing approach; 2% is aggressive.
Portfolio heat: the total risk across all open positions. Cap it — 6% is a common ceiling. Six trades at 1% each is your full allowance.
Correlation: five positions in the same sector is one position with five names on it. Sector-wide shocks — as September 2018 demonstrated for NBFCs — do not respect diversification by ticker.
Leverage: multiplies both the outcome and the gap risk. Everything above assumes unleveraged positions; with leverage the drawdown table arrives faster and the stop protects you less.

Where people lose money

Moving the stop because you are attached to the trade

The stop was placed when you were thinking clearly. It gets widened when price approaches it and you are not. The rationalisation is always available — "this is a shakeout", "the level is a zone, not a line", "the fundamentals haven't changed".

Occasionally that is right, which is exactly the problem: intermittent reinforcement makes the habit extremely durable. The rule that survives contact with reality: you may tighten a stop, never widen it. If you find yourself needing to widen it, your size was wrong at entry.

Before every position

☐Where is the stop, and what specifically does it falsify?
☐Is it beyond ordinary noise for this instrument? Check against ATR.
☐How many shares does 1% risk allow given that stop distance?
☐What is the realistic objective, and is it at least 2R away?
☐What is my total portfolio heat with this position added?
☐Is this correlated with anything I already hold?
☐If it gaps through the stop overnight, is the loss survivable?

In one screen

1Compute size from the stop, never the reverse.
2R makes outcomes comparable; expectancy, not win rate, decides viability.
3The best stops are both beyond noise and at the price that falsifies the thesis.
4Drawdown recovery is brutally asymmetric — size for the losing streak you know is coming.
5You may tighten a stop; you may never widen it.

Check yourself

Nothing is scored or stored — the explanation after each answer is the point.

1/2

Method A wins 35% of the time with +3.2R average wins and −1R losses. Method B wins 72% with +0.6R wins and −1.9R losses. Which is viable?

2/2

Why is per-trade risk usually capped around 1% rather than 5%?

Next lesson →

Entries, exits and timeframes

← Previous

Relative strength: what to own, not when

← All lessons