Never Risk More Than 5% Per Trade: Position Sizing and Stop Math

Risk per trade has a short answer: decide first how much of your total capital one trade is allowed to lose (most disciplined traders use 1%–2%, and 5% is a ceiling, not a target), then work backward from your stop to the number of shares. The formula is shares = capital × risk% ÷ (entry − stop). Position size is not how much you feel like buying; it is the output of how much you can afford to lose.

Key takeaways

• Risk is not position size. Risk = position size × stop distance. Cap the risk first, then derive the size.
• Shares = capital × risk% ÷ (entry − stop). The wider the stop, the fewer shares you can buy.
• After 8 straight losses, 1% risk leaves 92% of capital, 2% leaves 85%, 5% leaves 66%. The difference only shows up in bad luck, which is exactly when it matters.
• Per-stock and per-sector caps are two more ceilings. Breach either and the per-trade rule is void.
• In Stock Compass, set the recommended max % per position and an exit rule on floating_loss_pct, so the scanner watches the numbers for you.

Decide the maximum loss first, then the size

The usual order is: like a stock, decide to put $50,000 in, think about the stop later. That puts the most important number last. Reverse it:

1. Set the per-trade risk cap as a percentage of capital. On $100,000 at 1%, one trade may lose at most $1,000; at 2%, $2,000.
2. Set the stop. The stop belongs at the price where your reason for buying is proven wrong, not at a round number picked afterward. See the stop-loss you can't execute for how to choose one.
3. Derive the share count from the two:

shares = capital × risk% ÷ (entry − stop)

The same thing from another angle: position size = risk amount ÷ stop distance. On $100,000 with a 2% cap ($2,000):

• Stop at -5% → max position = 2,000 ÷ 5% = $40,000
• Stop at -10% → max position = 2,000 ÷ 10% = $20,000

Wider stop, smaller position. Wherever the stop sits, hitting it costs the same fixed $2,000. You have turned "how much I lose" from a random variable into a constant; the only thing left uncertain is whether the stock works.

A worked example you can follow with a calculator (Hong Kong)

Capital: HKD 500,000. Per-trade risk cap: 1%, so at most HKD 5,000 per trade.

A stock trades at HKD 62.00. You decide that a break below the recent consolidation low at HKD 57.50 would invalidate the idea.

• Risk per share = 62.00 − 57.50 = HKD 4.50
• Theoretical shares = 5,000 ÷ 4.50 ≈ 1,111
• Hong Kong stocks trade in board lots; assume 500 shares per lot, so round down to 1,000 shares (2 lots)
• Position = 1,000 × 62.00 = HKD 62,000, or 12.4% of capital
• Loss if stopped = 1,000 × 4.50 = HKD 4,500, or 0.9% of capital

Note the last line: actual risk is slightly under 1% because you rounded down. Always round down; never round up to make a nicer number.

Same stock, tighter stop at 59.00: risk per share becomes 3.00, shares = 5,000 ÷ 3.00 ≈ 1,666 → 1,500 shares, position HKD 93,000, or 18.6% of capital. Tighter stop, bigger position, but a stop-out still costs about HKD 4,500. That is "constant risk, floating size."

Why 1%–2% survives what 5% does not: losing-streak math

Every strategy has losing streaks. With a 50% win rate, you will very likely see 5 straight losses somewhere in 100 trades, and 8 is not rare. The question is never "will I have a streak" but "what is left after it."

Capital remaining after consecutive losses, each a fixed percentage of the remaining balance:

Extend to 10 straight losses: 2% leaves about 82%, 5% about 60%, 10% about 35%.

Losses need a larger gain to recover: an 8% drawdown needs about 8.7% to get back to even, 15% needs about 17.6%, 34% needs about 51%. The 1% trader who just lost 8 in a row is one or two ordinary winners from flat. The 5% trader needs a string of big wins, at exactly the moment their nerves are least able to execute.

This is why "never risk more than 5% per trade" is a ceiling, not a recommendation. Five percent is the edge beyond which one bad run can take you out. Day to day, stay well below it.

The second ceiling: max % per position

The risk formula has a side effect: the tighter the stop, the larger the position it allows. Put the stop in the example at 61.00 and risk per share is only HKD 1.00. The formula says you may buy 5,000 shares, HKD 310,000, 62% of your capital.

The formula is not wrong; its assumption is: that you will actually get filled at the stop. Gaps, trading halts and post-earnings plunges all make the real loss much bigger than planned. Sixty percent of capital in one name hands the whole portfolio to a single gap.

So you need a second ceiling: maximum % of capital per position. Common values are 15%–25%, depending on account size and number of positions. Final size = the smaller of the risk-formula size and the per-position cap. The trade with the 61.00 stop is capped at 20% (about HKD 100,000, rounded down to 1,500 shares by lot), and its actual risk drops to 1,500 × 1.00 = HKD 1,500, about 0.3%. Risking less than planned is fine. Risking more than planned is the problem.

Sector concentration stacks single-name risk

The 1% rule quietly assumes your trades are independent. When five holdings are all in the same sector, they are not. Sector-wide bad news triggers five stops on the same day, and you lose 5%, not 1%.

Correlation also rises in sell-offs. Semiconductor names that seemed to move independently gap down together on an industry order cut, and the stops get jumped in unison. What looked like risk spread across five trades was five copies of the same risk.

Two simple lines:

• ≤ 20%–25% per stock, so any single gap has a bounded cost.
• ≤ 40% per sector. Stock Compass shows a concentration warning on the holdings page when one sector exceeds roughly 40% of your holdings. Forty is not a magic number; it is where your portfolio starts behaving like a sector ETF rather than a diversified book.

If your strategy deliberately rides sector momentum, accept the concentration and lower per-trade risk to compensate, say from 2% to 1%, so the product of the two stays manageable. Sector rotation and relative strength covers how to gauge where a sector stands.

How to set this up in Stock Compass

Stock Compass does not connect to a broker, place orders or recommend stocks; it turns your rules into a daily scan and alerts. Position sizing lands as two settings:

1. Max % per position. In settings, enter your second ceiling as the recommended max per position, say 20%. A holding whose cost basis exceeds that share of capital is flagged on the holdings page, which also warns when one sector passes roughly 40%. It will not stop you; it makes the breach visible.

2. An exit rule on floating_loss_pct. floating_loss_pct is the unrealized loss as a percentage of cost, positive when you are losing. If your stop is "out on a 7% drop below entry," the exit rule is:

{
  "name": "Exit rule for 1% risk per trade",
  "market": "hk",
  "rules": {
    "buy":  { "v": 2, "outerOp": "OR", "groups": [] },
    "add":  { "v": 2, "outerOp": "OR", "groups": [] },
    "trim": { "v": 2, "outerOp": "OR", "groups": [] },
    "exit": {
      "v": 2,
      "outerOp": "OR",
      "groups": [
        { "innerOp": "AND", "conditions": [ { "indicator": "floating_loss_pct", "operator": ">", "value": 7 } ] },
        { "innerOp": "AND", "conditions": [ { "indicator": "is_20d_breakdown", "operator": "==", "value": 1 }, { "indicator": "floating_loss_pct", "operator": ">", "value": 3 } ] }
      ]
    }
  }
}

The first group is the hard stop: an exit signal once the unrealized loss passes 7%. The second is a structural stop: price breaks the 20-day low while you are already down more than 3%. The groups are joined by OR; either fires the alert.

Now connect 7% and 1%: a 7% stop and 1% risk per trade means this kind of trade should be sized at 1% ÷ 7% ≈ 14% of capital. Risk per trade, stop distance and position size: fix two and the third is determined. Do not pick each separately. Position health shows the unrealized loss next to your exit rule so you do not have to compare them by hand.

Common mistakes

• Confusing position size with risk. Buying $50,000 of stock is not $50,000 of risk, nor 5% risk. Risk is size × stop distance; skip that multiplication and the rule is empty.
• Moving the stop lower after entry. Move it once and the formula's denominator changes; the share count you calculated is now oversized. The stop is a precondition of the size; changing it opens a bigger trade.
• Rounding the wrong way. Hong Kong trades in board lots, mainland China in lots of 100. Round down every time. Rounding up "just a few hundred shares" quietly turns a 1% rule into 1.2%.
• Watching the stock, not the sector. Five same-sector names at 15% each is 75% sector exposure. No per-trade rule survives five stops on the same day.
• Sizing up after a streak to win it back. Raising risk from 1% to 5% after five straight losses is the worst path in the table above. A streak is a signal to size down, not up. Log every trade in a trading journal to see where control slipped.

Summary

Position sizing has one order of operations: decide the maximum loss for this trade, place the stop, then compute the shares. Add two ceilings, no more than 20%–25% in one stock and roughly 40% in one sector, so concentration cannot void the per-trade rule. None of this makes you more money. It makes sure that after eight losses in a row, both your capital and your judgment are still intact for the next trade.