Position Sizing and Stop Losses

Two numbers decide whether a losing trade is an inconvenience or a problem. How much you put in, and where you get out. Most investors settle the first one by feel and never settle the second one at all.

Here are the rules I use, the arithmetic behind them, and the distinction that causes more confusion than anything else in this subject.

Invested and at risk are not the same number

Start here, because mixing these two up is the error underneath most of the others.

Take a $100,000 portfolio. You open a 3% position, so $3,000 goes in. The stock is at $50, which buys 60 shares. You place your stop at $45.

The trade
Portfolio	$100,000
Position size, 3%	$3,000
Entry price	$50.00
Shares	60
Stop price	$45.00
Risk per share	$5.00
Planned risk to the stop	$300

$3,000 is invested. $300 is at risk, which is 0.30% of the portfolio. Those are wildly different numbers and they answer different questions.

So when someone asks how much you have on a trade, including when the person asking is you, know which one you mean. An investor who thinks a 3% position means 3% of the portfolio is in danger will size far too small. One who thinks a stop makes the position free will size far too large.

Two caps, and the smaller one wins

There are two ways to size a position and you need both running at once.

The first works backward from risk. Decide what you're willing to lose if the stop triggers, measure the distance from entry to stop, and divide. That tells you how many shares.

The second ignores the stop entirely and just caps how much of the portfolio any single name can occupy. My rule is 3% of capital on opening a position, and 5% as the absolute maximum after a winner has run.

When the two methods disagree, and a risk calculation says you could put 6% into one name because the stop is tight, the value cap wins. Take the smaller number.

A tight stop does not make a large position safe. It makes a large position cheap to be wrong about, right up until the stock gaps past the stop overnight.

Those caps come from four decades of watching what happens to concentrated portfolios when they turn out to be wrong, however carefully the position was sized going in.

Where the stop goes

Not at a round number. Not at a percentage you picked because it sounded tolerable. The initial stop goes at the price that proves the trade was wrong.

That's a technical level: a break of the support the setup was built on, a failed pattern, the low of the pullback you bought. Whatever defines the premise, the stop sits just the other side of it. If price gets there, the reason you bought is gone, and the position should be too.

This ordering matters more than it looks. The stop comes first and the position size falls out of it. Decide the stop, measure the distance, then calculate how many shares that distance allows. Doing it the other way around, picking a position size and then hunting for a stop that fits, produces stops at levels that mean nothing and get hit by ordinary noise.

Three layers, not one

In my own process there are three stops operating at different stages, and they do different jobs.

The initial technical stop. Placed at entry, at the level that invalidates the setup. This is the one that determines position size.

The 8% trailing stop. Layered in once the position is showing a profit, trailing 8% below the peak. Its job is protecting gains as a trade develops, not defining the original risk.

The 20% to 25% hard stop. Sitting underneath both as a backstop against a thesis that broke without the chart telling you first. This is not a working exit. If you are regularly hitting it, something upstream in your process has failed.

One caution on the trailing stop. In a retail brokerage account you can usually place a trailing stop order and let the platform maintain it. Not every system supports them, and in those cases it becomes a manual discipline you have to actually perform. A rule you don't execute isn't a rule.

Reward to risk, and why it's a filter rather than a law

Before entering, compare what you stand to make against what you stand to lose. My screen is 3:1, risking a dollar to make three.

That number is a discipline, not a guarantee, and I want to be straight about why. No ratio produces positive expectancy on its own. A 1.5:1 setup with a well-tested win rate can beat a 3:1 setup with a poor one. Expectancy is what actually decides whether a method makes money, and expectancy needs your win rate and your average win and loss, not just the ratio you hoped for at entry.

What the 3:1 screen does is keep me out of trades whose economics are already marginal before costs and before the win rate is accounted for. Setups that fail the screen aren't necessarily bad. They just don't get my capital.

Portfolio heat

Here's the one that catches careful people.

Every position individually sized within the rules can still add up to a bad day. Five positions, each risking 0.30% to its stop, is 1.5% of the portfolio at risk. Manageable. But if all five are in the same sector and the sector drops together, those stops don't trigger independently. They trigger at once.

Add up the risk to stop across every open position. That total is your portfolio heat. Cap it the way you cap individual positions, and watch the correlation between the names, not just the count.

Does the edge survive the cost of trading it

A setup that looks profitable on a chart can evaporate once you pay for it. Bid-ask spread on the way in and the way out. Slippage when the fill isn't the price you saw. Commissions if your broker still charges them. Margin interest if you're borrowing. Short-term capital gains tax on the winners.

Run the numbers with costs included from the start, not as a correction afterward. The frequency of a strategy matters enormously here: a method that trades twenty times a year survives friction that would destroy the same method trading twenty times a month.

What to do with this

Take a position you currently hold and answer three questions in writing. How much is invested. How much is at risk to your stop, in dollars and as a percentage of the portfolio. And what specific thing would have to happen for you to conclude the position was a mistake.

If you can't answer the second question because there is no stop, that's the finding. If you can't answer the third, the stop you do have is a number rather than a reason, and it will get hit by noise and feel like bad luck.