Someone asked me a while back why their position sizes always felt wrong, and after twenty minutes of digging it turned out to be a thing I see constantly. They had read that you should never risk more than 1 percent of your account on a single trade, applied it faithfully, and the rule was quietly producing sizes that made no sense for their account or their strategy. The rule itself is fine. The problem is that it gets taught as a constant when it is really a variable, and almost nobody explains what it is a variable of.
The formula, because most explanations skip it
The mechanics first, since half the people who quote the rule get this part wrong. The common misreading is that you put 1 percent of your account into the trade. The actual rule sizes the position so that if your stop loss gets hit, the realized loss equals 1 percent of your equity. Those are very different numbers, and the difference is the whole point.
Three inputs. Account equity, risk percentage, and stop distance. Multiply equity by the risk percentage to get a risk budget in currency terms. Then take the distance between your entry price and your stop price. Divide the budget by that distance and you have your position size in units. Written out, position size equals equity times risk percent, divided by entry minus stop.
Concrete version. You have 10,000 in the account and you risk 1 percent, so the budget is 100. You want to buy at 50 with a stop at 46, a stop distance of 4. Position size is 100 divided by 4, so 25 units, which is 1,250 of notional. The notional is 12.5 percent of the account even though the risk is 1 percent. Tighten the stop to 49 and the same 100 of risk buys 100 units, 5,000 of notional. Stop distance drives the size while the risk stays constant, which is why traders with tight, well defined stops can run much larger positions at the same risk. On perps the formula is identical. Leverage changes how much margin you post, and it changes nothing about how much you lose when the stop fires.
One caveat that bites people in crypto specifically. The formula assumes the stop fills at the stop price. In a fast market it will not, so the real loss at the stop is usually somewhat worse than the theoretical one. If you trade thin pairs, or you hold perps through news, pad the stop distance in the calculation or shave the percentage, because your true risk per trade is 1 percent plus slippage plus fees.
Three places the rule quietly breaks
Small accounts first. On a 500 account, 1 percent is 5. Fees on entry and exit eat a meaningful chunk of that before the trade does anything, and on plenty of venues the computed position size lands below the minimum order size entirely. So you either do not trade at all, or you round up to a size that silently means 3 or 4 percent risk while you keep telling yourself it is 1. I would rather see someone with a small account consciously choose 2 or 3 percent, take fewer and more selective trades, and know exactly what they signed up for, than pretend a rule calibrated for five and six figure accounts fits a three figure one.
High frequency strategies fail in the opposite direction. If your system fires ten signals a day and positions stay open for a couple of days, you are routinely carrying fifteen or twenty open positions. At 1 percent each, that is 15 to 20 percent of the account exposed to stops at the same time, and one ugly session can hit most of them together. The per trade number looks conservative while the portfolio number is anything but. For anything that trades frequently, the figure that matters is total open risk, meaning the sum of what you lose if every stop triggers today. I cap that first, somewhere around 4 to 6 percent depending on the strategy, then divide by the typical number of concurrent positions to get the per trade figure. That often lands at 0.25 or 0.5 percent per trade, which looks timid until you remember how many trades are live at once.
Correlated entries are the sneakiest failure. Long BTC, long ETH, long SOL, plus a couple of alt perps in the same direction reads as five separate 1 percent trades. In a genuine risk-off move, crypto correlations shoot toward 1 and all five stops trigger within the same hour, so what you were actually holding was a single 5 percent position wearing five tickers. The same thing happens with a basket of semiconductor names, or three currency pairs that share a dollar leg. My fix is to size by bucket rather than by ticker. Everything that would plausibly get stopped out by the same move shares one risk budget, and adding a position to a full bucket means shrinking the others or skipping the trade.
Picking your own number
The honest way to set the percentage is to work backward from the drawdown that would make you abandon your own system, because that is the real point of failure. Most systems die when the trader stops following them mid drawdown, well before the account is in actual danger. Here is the sequence I use.
- Estimate your realistic losing streak. A strategy that wins roughly half its trades will produce runs of eight or more consecutive losses more often than intuition suggests. If you have backtest results, read the worst streak off them and assume live trading will be somewhat worse.
- Decide the maximum drawdown you can sit through without changing your behavior. Be pessimistic here, since nearly everyone overestimates this number.
- Divide that drawdown by the streak length, adjusted upward for however many positions you typically hold at once. The result is your per trade ceiling.
- Bucket correlated assets and apply one shared risk budget per bucket.
- Check that the computed position size clears your venue minimums and that fees are a small fraction of the risk budget. If either test fails, the account is too small for that percentage, and pretending otherwise just hides the real number.
- Recompute off current equity on a schedule, weekly is plenty, so sizes shrink in drawdowns and grow after gains without an emotional decision each time.
Run through that once and you will usually land somewhere between 0.25 and 2 percent depending on trade frequency and correlation, which explains why 1 percent survives as the default. It sits in the middle of the sensible range, and defaults exist for people who have not yet done the math on their own strategy. For the streak and correlation inputs I pull numbers from the backtesting side of Blockcircle, since worst losing streak and how positions historically moved together are exactly the stats this exercise needs, but a spreadsheet and a few months of honest trade logs will get you most of the way there.
The piece worth memorizing is the formula, equity times risk percentage divided by stop distance, because it converts whatever percentage you choose into an exact size in about ten seconds. The percentage itself deserves an hour of real thought once, and a review whenever your strategy or account changes shape. Most traders spend that hour on entries instead, which is roughly the wrong order.