Risk Management for Traders: The 1% Rule, R Multiples and Expectancy

Lose 30 percent. You need 42.9 percent to get back to where you started. Everything else on this page exists because of that sentence. The gain is calculated on a smaller base than the loss was, so the hole gets harder to climb out of at exactly the rate it gets deeper, and no amount of skill at picking entries changes the shape of that curve.

Fix the loss before you look at the chart

One percent of the account is the standard setting. On $20,000 that is $20,000 x 0.01 = $200.

That is the loss if the trade reaches your exit. It has no connection to how much stock you own. You might own $2,000 of it or $10,000 depending on where the exit sits.

People starting out ask whether 1 percent is too timid. Look at what the larger settings do to a run of six losses. Every method on earth produces one sooner or later.

Risk per trade	$20,000 after 6 straight losses	Drawdown
1%	$18,829	5.9%
2%	$17,717	11.4%
5%	$14,702	26.5%
10%	$10,629	46.9%

At 10 percent a trade, six losses in a row, which is unremarkable for a method that wins 40 percent of the time, have taken almost half the account, and the recovery table further down shows why the bottom row is effectively terminal.

The share count is an output

Two inputs you already have produce it, the dollar risk and the distance to the stop.

shares = dollar risk / stop distance

Entry $50, stop $48, distance $2. With a $200 budget, $200 / $2 = 100 shares. That is a $5,000 position. A quarter of a $20,000 account, and a $200 loss when the stop trades.

Entry	Stop	Distance	Shares	Position value	Loss if stopped
$50	$49	$1	200	$10,000	$200
$50	$48	$2	100	$5,000	$200
$50	$45	$5	40	$2,000	$200
$120	$114	$6	33	$3,960	$198

Every row loses the same. The volatile name that needs a wide stop gets the smallest position. That is the reverse of what almost everyone does by instinct.

Two errors follow from running this backwards. Pick the share count first and place the stop where the loss feels bearable, and your stop now sits at a price the market has no reason to defend. Use margin to carry more than the calculation supports and you have changed the loss without changing the plan; the margin calculator shows what borrowed money does to the downside.

R

R is the amount risked. Express every result as a multiple of it and trades at different prices become comparable.

Risk $200 and make $600: a 3R win. Risk $200 and hit the stop: a 1R loss. Risk $200 and bail early for $50: a 0.25R win. A record kept in R shows you things a record kept in dollars hides, because a $400 gain on a trade that risked $400 is mediocre while the same $400 on a trade that risked $100 is excellent, and your statement prints them identically.

Put an R column in the journal from trade one. Everything below is calculated from it.

Expectancy

expectancy = (win rate x average win) - (loss rate x average loss)

Worked: 40 percent of trades win, the average win is $300, the average loss is $150.

(0.40 x $300) - (0.60 x $150) = $120 - $90 = $30 a trade, or $6,000 over 200 trades before a dollar of costs comes out.

Now move one input at a time, leaving the other two where they are.

Win rate	Average win	Average loss	Expectancy	200 trades
40%	$300	$150	$30	$6,000
40%	$200	$150	minus $10	minus $2,000
30%	$300	$150	minus $15	minus $3,000
40%	$300	$250	minus $30	minus $6,000
55%	$200	$200	$20	$4,000

The second row is the one you will live. Same win rate, same losses, winners cut a third shorter. A profitable method is now losing $2,000 a year. The fourth row is the other common ending, the stop that got moved. Notice that the win rate is unchanged in both, which is all you need to know about anybody who describes a method to you using a win rate and nothing else.

What the expectancy has to beat

Two hundred round trips a year. One hundred shares, a $50 price, a two cent spread. Crossing the spread costs 100 x $0.02 = $2 a side, so $4 a round trip, or 200 x $4 = $800 a year. A broker charging $1 a side adds 200 x 2 x $1 = $400. Total $1,200, which on a $20,000 account is 6 percent.

Against the $30 expectancy, 200 trades gross $6,000 and net $4,800 while the second row of the table loses $2,000 and pays another $1,200 in costs.

Cost per round trip is also why a small edge cannot be traded often. Halve the average win to $150 and expectancy becomes (0.40 x $150) - (0.60 x $150) = -$30. No amount of discipline recovers that, and no amount of frequency.

The recovery curve

recovery gain = (1 / (1 - drawdown)) - 1

Drawdown	Gain required
10%	11.1%
20%	25.0%
30%	42.9%
40%	66.7%
50%	100.0%
60%	150.0%

Check the third row: 0.70 x 1.429 = 1.00. Past 40 percent the requirement outruns what most methods produce in a strong year, which is the practical case for keeping risk per trade at 1 or 2 percent. Run your own figures through the drawdown recovery calculator.

Heat

Add the risk across every open position. Five at 1 percent is 5 percent of heat, and the assumption that those are five independent risks collapses the moment they share a sector, a theme, or a sensitivity to the same rate move.

How bumpy the resulting equity curve is takes you into return per unit of risk, which the Sharpe ratio calculator computes and measuring portfolio risk explains properly.

The rules that stop you

Set a monthly loss limit alongside the per trade limit. A common version ends the month’s trading at a 6 percent account loss. That caps the damage from a bad stretch and, more usefully, from the decisions a bad stretch produces, which are the subject of trading psychology and discipline.

Write the stopping rules while you are flat. Nobody has ever invented a sensible loss limit while sitting in a loss.

Kelly, and why you use a fraction of it

Kelly answers a different question from the 1 percent rule. Given an edge, what fraction of capital maximises long run growth? For a fixed reward to risk ratio,

f = W - ((1 - W) / R)

where W is the win rate and R is the ratio of average win to average loss. At a 40 percent win rate and 2 to 1: 0.40 - (0.60 / 2) = 0.40 - 0.30 = 0.10. That is 10 percent of capital at risk per trade.

Nobody sensible trades full Kelly. The formula assumes you know W and R. You estimated them from a few dozen of your own trades, which is a sample small enough that a ten point overstatement of the win rate is ordinary, and the formula responds to that overstatement by telling you to bet far more than the edge supports. A quarter of the figure, 0.10 / 4 = 2.5%, is nearer to what practitioners use. It lands close to where the drawdown table already pointed.

Work your numbers with the Kelly criterion calculator and read the derivation and its failure modes in the Kelly criterion and position sizing.

When this arithmetic is automatic, test it against the risk management and position sizing quiz, then apply it to a timeframe using how to start trading stocks.