Risk: position size and stop

Risk per trade is the amount you lose if the trade turns out wrong and the stop is hit. Not “how much I could make”, not “how much money is on the exchange”, but exactly the loss in the bad outcome. It is named before entry, not after.

Risk is best measured as a percentage of the account, not in dollars. The account changes: after a good month $100 is a smaller share, after a bad one a larger share. A percentage keeps the measure constant. A $10,000 account at 1% risk means $100 per trade; a $25,000 account at 0.5% means $125.

Which percentage to pick is your decision, and the course does not make it for you. People often take a fraction of a percent or about one percent — not because that is “correct”, but because of the arithmetic we go through below and in detail in section 17.

Why risk is usually small. Even good trading loses several times in a row — this is not an exception but the norm (we compute the probabilities in section 17). If you lose 1% of the current account each time, ten losses in a row leave 90.4% of the account: 0.99 to the tenth power. If you lose 10%, 34.9% of the account remains: 0.9 to the tenth power. The same ten trades, the same luck — but in the first case it is an unpleasant week, in the second it is the end of trading.

Note: ten losses of 10% are not minus 100% but minus 65.1%, because each next percent is taken from an already smaller account. That does not make it easier: to get back from 34.9% to 100% you would have to nearly triple what is left.

Beginners often mix up three different numbers. Let us separate them right away.

• Risk — how many dollars you lose when the stop is hit. For example, $100.
• Position size — how many dollars' worth of the coin you hold (the position's value). For example, $5,000. Profit, loss and fees are all computed from it.
• Margin — the collateral the exchange locks for the position. At ×10 leverage that is $500 of the same $5,000.

With a $10,000 account, a $5,000 position and a stop 2% from entry, the loss at the stop is $5,000 × 2% = $100, i.e. 1% of the account. The $500 margin plays no part in this calculation: it says how much money is locked, not how much can be lost.

A stop is the price at which the trade's idea has turned out wrong. Not the price where it “hurts too much”, and not the price that gives a nice round loss. If you buy because a liquidity wall sits below and is not being eaten, the idea is wrong once the wall is eaten and price goes lower. If you buy off the low of a move, the idea is wrong below that low. The stop goes just beyond such a point: where the market has already proven you wrong and where ordinary noise does not reach.

A common mistake is the reverse order. “I want a $10,000 position and can lose $100, so the stop is at 1%.” Here the stop is derived from a sum, and the market knows nothing about your sum. If the nearest place where the idea breaks is 2.5% from entry, a 1% stop sits in the middle of normal fluctuations: noise will take it out while the idea may still be right.

The right order has four steps:

1. The idea and the place where it breaks → the stop price.
2. The distance to the stop in percent: |entry − stop| / entry.
3. The acceptable loss in dollars: account × risk per trade.
4. Position size = acceptable loss / distance to the stop.

A $10,000 account, 1% risk = $100. A stop 1% from entry gives a position of $100 / 0.01 = $10,000. A stop at 2.5% gives $100 / 0.025 = $4,000. The loss at the stop is $100 in both cases. A far stop does not make the trade “riskier” — it makes the position smaller. A close stop does not make it “safer” — it makes the position bigger.

A stop is not a price guarantee. Once triggered, a stop order becomes a market order and fills against whatever is in the book (section 2). In a calm market the difference is a fraction of a percent. In a sharp one — especially when other people's stops and liquidations fire nearby (sections 12–13) — the fill can land well beyond the stop. So the planned loss is computed with room for slippage, and the stop is not placed right on a round number or an obvious level where everyone's stops sit.

One more thing: a stop you keep “in your head” rather than on the exchange is not a stop. At the moment you need it you will have to decide under pressure, and the decision will almost always be “I'll wait a bit more”.

There is one position size formula, but it is handy to compute it two ways — in dollars and in coins.

In dollars: size = acceptable loss / stop distance as a fraction. A $20,000 account, 0.5% risk = $100. A long entry at 2,000, stop at 1,960: the distance is 40 / 2,000 = 2%. Size = $100 / 0.02 = $5,000.

In coins: size = acceptable loss / stop distance in dollars per coin. 100 / 40 = 2.5 coins. Check: 2.5 × 2,000 = $5,000 — it matches.

For a short everything is the same, only the stop is above entry. A short at 80,000, stop at 81,200: the distance is 1,200 / 80,000 = 1.5%. With an acceptable loss of $150 the size = 150 / 0.015 = $10,000, or 150 / 1,200 = 0.125 BTC.

The percentage trap. The exchange shows a position's return as a percentage of margin, not of the position and not of the account. At ×10 leverage a 1% price move is 10% “on screen”.

Example. A $10,000 account, a $5,000 position, ×10 leverage, $500 margin. Price moves 2% against you. The loss is $5,000 × 2% = $100. Now three percentages about the same event:

• 2% — the price change;
• 20% — the loss relative to margin ($100 / $500), the number the exchange shows;
• 1% — the loss relative to the account ($100 / $10,000), which is the actual risk.

Hence a frequent confusion: “my stop is −20%” in the exchange interface at ×10 is only 2% of price. And “a stop 2% of the position” at ×10 is −20% on screen. Before comparing percentages, ask: percent of what?

What leverage does. If the position size is computed from the stop, leverage does not affect the loss at the stop. A $5,000 position with a 2% stop loses $100 both at ×2 ($2,500 margin) and at ×20 ($250 margin). Leverage only decides how much margin is locked — and, second, where liquidation is.

Roughly: with isolated margin and ×N leverage, the exchange closes the position when the loss has almost eaten the margin — about 1/N from the entry price, and a little closer once maintenance margin and fees are counted (the exact mechanics are in section 12). At ×10 liquidation is a bit closer than 10%, at ×30 a bit closer than 3.3%, at ×50 closer than 2%. If your stop is 4% from entry and leverage is ×30, the exchange closes the position before price reaches the stop: in effect there is no stop, and the whole margin is lost.

The check is simple: liquidation must be noticeably farther than the stop. If it is closer, lower the leverage, not the stop.

Sometimes the formula gives a position bigger than the account. A $2,000 account, 1% risk = $20, a 0.4% stop — the position is $20 / 0.004 = $5,000, 2.5 times the account. That is fine: such a position needs leverage of ×2.5 or more, and liquidation at that leverage (about 40%) is far beyond the stop. A 0.1% stop, though, would give a $20,000 position — ten times the account. The formula still works, but fees come into play: they are the next lesson, and there you will see why a very close stop is almost always more expensive than it looks.

Fees are charged on the position size, not on margin, and are paid twice — on entry and on exit. If the stop fills as a market order, the exit almost always pays the taker rate.

A $5,000 position, a 0.05% taker fee each side: $2.50 + $2.50 = $5. At $100 risk that is 5% of the risk — tolerable. Now a 0.25% stop: to keep risk at $100 the position = $100 / 0.0025 = $40,000. Fees: 0.1% × $40,000 = $40 — already 40% of the risk. With a 0.1% stop the position is $100,000 and fees are $100 — as much as the risk itself.

The general rule: fees as a share of risk = (entry fee + exit fee) / stop distance. 0.1% / 0.25% = 0.4; 0.1% / 1% = 0.1; 0.1% / 0.1% = 1. The closer the stop, the larger the share of risk eaten by fees — and slippage comes on top.

How to put fees inside the risk. The real loss at the stop = size × (stop distance + entry fee + exit fee). To keep it equal to the acceptable loss, size is computed like this:

size = acceptable loss / (stop distance + round-trip fees).

A $10,000 account, 1% risk = $100, a 1% stop, a 0.05% fee each side. Ignoring fees the size would be $10,000 and the loss at the stop $100 + $10 = $110. With fees: $100 / (0.01 + 0.001) = $9,090.91. Check: 9,090.91 × 1.1% = $100. The difference looks small, but over a hundred trades an extra 10% of risk on each adds up to a noticeable sum.

R — the unit of results. 1R is the planned loss at the stop of one trade. A trade's result in R = profit or loss / R. Lost exactly the stop — −1R. Made $250 with R = $100 — +2.5R. Moved the stop and lost $300 — −3R.

For a trade without partial exits, the result in R before fees can be computed straight from prices: (exit − entry) / (entry − stop) for a long. Entry 100, stop 98, exit 105: 5 / 2 = +2.5R. For a short the signs swap: a short at 3,000, stop 3,060, exit 2,880 — (3,000 − 2,880) / (3,060 − 3,000) = 120 / 60 = +2R. Fees are then subtracted in the same R.

Why bother. Trades come in different sizes, the account grows and shrinks, but in R any trades are comparable: +2R on a $5,000 account and +2R on a $50,000 account are equally good results. Everything that follows rests on R: the average per trade, streaks, judging your own trading (section 17), and EVA counts results in R too (section 18).