FTMO daily drawdown: get the alert before the firm does

Most prop firm accounts are not lost to a bad strategy. They are lost on one afternoon, when a daily loss limit is crossed while nobody is looking at the terminal. The firm's system sees it instantly; the trader finds out from an email. This guide explains how the daily drawdown rule is actually calculated, why expert advisors breach it more often than manual traders, and how to set up monitoring that warns you before the firm does.

How the daily loss limit is really calculated

The detail that catches people is that the daily limit is almost never based on closed trades alone. At FTMO, and at most firms that copied its model, the maximum daily loss is a fixed percentage of the initial balance (5% on the 2-Step Challenge and 3% on the 1-Step, at the time of writing), and the calculation includes floating profit and loss, commissions and swaps. The reference point resets once a day at 00:00 CE(S)T (Prague time), not at your local midnight.

Three consequences follow:

• An open position can breach the rule without ever closing. A trade that dips 5.1% and recovers to profit has still violated the limit.
• The allowance moves with your day. If you closed the previous day up 2%, the new day's reference is that higher balance, so the distance to the limit is measured from there.
• The reset time matters. A grid or swing EA holding positions through the reset starts the new day already carrying floating loss.

Every firm words this slightly differently (balance-based, equity-based, trailing, static). Read your provider's current rule page and write the formula down before you configure any alert. The rest of this guide assumes the common "initial balance, equity-inclusive, server-midnight reset" model.

Why EAs breach it more often than humans

Typical ways an EA crosses the daily limit

Cause	What happens
Lot sizing tied to balance, not to remaining daily allowance	After two losses the third trade is sized as if the full 5% were still available.
Several EAs on one account	Each respects its own risk; together they exceed the account's limit.
Spread widening at rollover or news	Floating loss spikes for seconds. Equity-based rules count those seconds.
Recovery logic (grid, martingale, averaging)	Exposure grows exactly when the day is already negative.
Copying the same EA to many accounts	One bad session breaches all of them at the same minute.

None of these are strategy failures. They are supervision failures: the information needed to stop was available minutes earlier, but nothing was watching it.

What a useful drawdown monitor has to measure

1. Equity, not balance. Sample account equity at least once a minute. Balance only changes when trades close, which is too late.
2. Day-start reference at the firm's reset time. Store the balance (or equity, depending on the rule) at server midnight and compute today's loss against it.
3. Distance to limit as a percentage of the allowance. "You have used 62% of today's loss allowance" is actionable. "Equity is 98,140" is not.
4. Freshness of the data itself. If the terminal or VPS dies, the last known equity looks fine forever. A monitor must mark an account stale when reports stop, and treat stale as an alarm, not as calm.
5. Per-account and fleet view. With five funded accounts you need to see the worst one first.

Alert thresholds that leave you time to act

A single alert at 100% is a post-mortem. Use a ladder:

Suggested alert ladder for a 5% daily limit

Allowance used	Daily loss	Action
50%	−2.5%	Notification. No action; awareness only.
70%	−3.5%	Stop opening new positions. Disable recovery logic.
85%	−4.25%	Close or hedge open risk manually. Turn AutoTrading off.
Stale data > 5 min	unknown	Treat as 85%: log in and check the terminal now.

The exact numbers depend on how fast your strategy can lose. A scalper on gold needs wider gaps between steps than a swing EA on EURUSD. Calibrate on a demo or a small challenge first.

Why independent, read-only monitoring matters

You can code these checks inside the EA, and you should: a hard daily stop in MQL5 is the first line of defence. But an in-EA guard shares every failure mode of the EA. If the terminal freezes, the guard freezes. If the VPS reboots, both are gone. An external monitor that only receives telemetry keeps working when the terminal does not, and because it has no ability to place or modify orders it cannot make the situation worse.

There is a second benefit. When a firm disputes a day, or you dispute their calculation, a time-stamped, independent record of equity, lots and decisions is evidence. Screenshots are not.

This is the model Quantisentry follows: the EA reports its state through one MQL5 include, the platform validates and stores it, and dashboards are read-only. Routes that could alter trading state do not exist.

A ten-minute setup checklist

• Write down your firm's daily-loss formula and reset time.
• Add a hard daily stop inside each EA at 80–85% of the allowance.
• Make lot size depend on remaining daily allowance, not on balance.
• Report equity and open lots to an external monitor every minute.
• Configure the 50 / 70 / 85 ladder and a stale-data alarm.
• Decide in advance who acts on each alert, and how (phone, RDP, broker app).
• Test it: on demo, force a loss and confirm each alert fires.

Nothing here improves a strategy, and nothing here is investment advice. It only makes sure that when a day goes wrong, you find out while there is still something to do about it.