Every trader I know who blew up a month did it on one bad day. Not a slow bleed across four weeks, one afternoon where a losing trade turned into a revenge trade turned into three more, and by the close the account looked like it had been through a wood chipper. Most of these people had a daily loss limit written down somewhere. They just also had the ability to ignore it, and on the exact day it mattered, they did. So the interesting question is not what number to pick. It is how to make the number bind when you are the least likely to want it to.
Size it off your own volatility, not a round number
The mistake is picking a limit that feels tidy. Two percent, five hundred dollars, whatever sounds serious. The problem is that a limit disconnected from how you actually trade will be either so tight you trip it on a normal Tuesday, or so loose it never triggers until the damage is already done. Both get overridden, because a limit that fires on nothing teaches you to ignore it, and a limit that only fires on catastrophe is not really preventing anything.
Start from the size of your typical losing trade instead. If you trade with a defined risk per position, call that 1R, one unit of risk. A common and pretty sane heuristic is to set your daily stop at three losing trades, or 3R. The logic is that being wrong three times in a row is a normal, survivable event that happens to good traders in a normal month. Being wrong three times and then continuing to trade is where the normal event turns into the account-ending one. Three strikes, you are done for the day, come back tomorrow.
If you do not trade with a clean fixed R, do the same thing empirically. Pull your daily P&L for the last few months and look at the spread. Your typical down day, the kind that happens roughly once a week, is your baseline. Set the daily limit somewhere around the edge of that normal range, so it catches the genuinely abnormal day without punishing an ordinary red one. You want it to fire maybe a handful of times a quarter, not every week and not once a year.
Put weekly and monthly tiers behind it
A daily limit alone has a gap. You can hit your daily stop on Monday, again Tuesday, again Wednesday, and technically follow the rules every single day while still having the worst week of your life. So you tier it.
- Daily: roughly 3R, or the edge of a normal down day. Hit it, you are flat and offline until tomorrow.
- Weekly: roughly two to three daily limits stacked, so about 6R to 9R. Hit it, you are done for the week. This is the one that catches the losing streak that individual daily stops would let slide.
- Monthly: the number that means something is actually wrong, not with the market but with you or your process. Historically I would set this around two weekly limits. Hitting it is a signal to stop and review, not to keep pressing.
The tiers matter because tilt does not always show up as one explosive day. Sometimes it is a slow, disciplined-looking slide where you obey the daily rule and still hemorrhage, because each day you tell yourself yesterday was the fluke. The weekly tier is the adult in the room that does not care about your story.
A limit you can override is a suggestion
Here is the part everyone skips. Knowing your number does nothing if hitting it just pops a message you click through. On a tilt day, your judgment is the broken instrument. You cannot use the broken instrument to decide whether to keep trading. The whole point is to make the decision earlier, when you are calm, and to make it stick when you are not.
There are roughly three levels of enforcement, and they get harder to defeat as you go down the list.
The softest is a sub-account with a capped balance. You only fund the trading account with what you are willing to lose that day or that week, and the rest sits somewhere with enough friction that moving it takes real minutes. When the tradeable balance is gone, there is nothing to trade with. This is crude and it works, because the tilt version of you wants to act now, and a transfer that takes a day is enough to break the loop. The catch is discipline in the funding step, and leverage can blow past the cap faster than the balance implies, so it is a floor, not a ceiling.
The middle level is an API-level kill switch. If you trade programmatically, or through anything that exposes an API, you write a small watcher that tracks realized plus unrealized P&L for the day and, when it crosses your limit, cancels open orders and flattens positions, then pulls or disables the trading key. The key detail is that the watcher runs somewhere you are not tempted to touch, and it revokes access rather than just alerting. An alert is a suggestion. A revoked key is a wall. Some venues also expose native controls, position or loss limits you set at the account level, and those are worth turning on because they live on the exchange side where the you-of-2pm cannot reach.
The hardest level is a platform lockout you cannot self-reverse. A few brokers and prop setups let you set a daily loss limit that locks the account until the next session, and crucially, only support can lift it early. If your platform offers that, use it. If it does not, you can approximate it by handing the ability to re-enable to something outside yourself, a cooling-off period, a second person, a time lock. The design goal is simple. The decision to stop is made once, in advance, by the calm version of you, and the tilt version has no button to undo it.
A setup you can stand up this week
If you want something concrete to actually build, here is the order I would do it in.
- Compute your R and pull a few months of daily P&L. Set daily at about 3R, weekly at about 6R to 9R, monthly at about two weeklies.
- Move most of your capital out of the tradeable account. Fund it to roughly your weekly limit so a bad streak physically runs out of ammunition.
- Turn on every native exchange-side loss and position limit the venue offers. Free, and it lives where you cannot override it in the moment.
- If you trade via API, write the watcher that flattens and revokes the key at the daily number, and run it off your trading machine.
- Write down, before the session, the one sentence that ends the day. Something like: at 3R down, I close the terminal, and I do not get to renegotiate that from inside a losing position.
None of this makes you a better trader on your good days. It is not supposed to. It exists for the afternoon when you are down three trades, certain the next one comes back, and completely wrong about your own state. The best limit is the one you set once and then cannot reach. Everything else is a note you will ignore at the worst possible time.