Most blown accounts do not die from a hundred small cuts. They die in one afternoon. The sequence repeats in every broker's data: a normal loss, a fast re-entry, size creeping upward, and a trader who was down $200 at lunch logging off down $1,100. Nothing about the market changed between noon and four. The trader changed. A daily loss limit exists for exactly one purpose: to make that afternoon structurally impossible.
Choosing the number
A practical default is three times your per-trade risk. On a $10,000 account risking 1 percent, that is 3R, or $300, about 3 percent of the account. The logic: three full stop-outs in a day is a perfectly normal bad day for any real system, so the limit does not punish ordinary variance. Beyond three, one of two things is true: the market is hostile to your setup today, or your execution has degraded. Under either explanation, the expectancy of the next trade is worse than your average, and continuing is paying to find out how much worse. Set the number too tight, one loss and done, and you train yourself to fear the first trade. Set it too wide, 6 or 8 percent, and it cannot save you from anything meaningful. Three R sits in the working range for most intraday styles. Rebase it as equity changes meaningfully, the same way per-trade risk rebases, so the limit is always 3R of the current account rather than a stale number from January.
Why it must be mechanical
A limit you can renegotiate is decoration. The entire reason it exists is that decision quality falls as losses stack up; the person deciding whether to keep trading at minus $290 is a worse decision-maker than the one who set the rule on Sunday. So the rule must execute without a meeting. Use the platform's daily-loss setting where one exists. Where it does not, the rule is: limit hit, positions flat, terminal closed, done, and the day's review happens away from the screen. Some traders add teeth: a breached limit means half size for the following day. What cannot work is treating the limit as advisory, because the only day it matters is precisely the day you will want to overrule it. Green days deserve a version too: many professionals pair the loss limit with a giveback rule, ending the session when a morning's peak profit has been half surrendered, since a fading day produces the same degraded decisions from the other direction.
The recovery math
Put numbers on what the limit buys. Suppose your system's expectancy is 0.27R, $27 per trade on this account. A capped bad day costs $300, which average trading repairs in about 11 trades, call it two or three sessions. An uncapped spiral to $800 needs about 30 trades of edge, a week and a half of good work spent refilling one afternoon's hole. And that arithmetic is generous, because it assumes the deeper hole gets repaired by the same trader who dug it, when in practice drawdown degrades judgment and the repair crew shows up impaired. The limit is not protecting today's $500. It is protecting the next two weeks.
Stopping is a skill
Walking away down $300 feels like failure, and that feeling is why so few traders manage it. Reframe the scoreboard. A career is a few thousand trading days; no single one of them matters, but the habit that caps all of them is worth more than any winning day. Track limit-respected days in your journal the way you track winners. A day stopped at the limit with rules intact is a successful day that happened to be red. Traders who internalize that stop flinching at the rule, and, oddly, start hitting it less, because the trades taken after minus 2R were mostly the bad ones anyway. It also helps to pre-decide what replaces the trading: the review happens immediately, notes on what the market was doing versus what your setup needs, then something physical and away from screens. A limit that leads somewhere specific is far easier to obey than one that leads to an empty afternoon of watching the move you are no longer allowed to trade.