A prop account is not lost when the market falls. It is lost at one line the firm draws under your equity. An equity guard exists to close everything before that line. Whether it does depends less on the line it watches than on how often it looks, and on when it thinks the day begins.
How we measured
We replayed real ticks of 21 October 2025 in the Strategy Tester. Gold fell hard that day, from about 4,371 to 4,093 on our tester's data. The account was $100,000 with a 5% daily limit, so the firm's line sat at $95,000. The guard was set to act at 80% of the daily allowance, at $96,000: $1,000 of room between the guard and the firm. One gold position was sized so that a 0.5% fall would cost 4.5%. The guard ran on a EURUSD chart, the way many of us park a utility on whatever chart is open. Between runs we changed one thing only: how often the guard checks.
| How the guard checks | Acted on the gold tick after the line was crossed | Lost beyond the guard's line | Share of the $1,000 room |
|---|---|---|---|
| Every price change it can see: its chart's ticks, the account equity every 250 ms, trade events | 0.19 s later (3 gold ticks) | $126 | 13% |
| A 1-second timer only | 0.19 s later (3 gold ticks) | $126 | 13% |
| Its own chart's ticks only (EURUSD) | 1.9 s later (21 gold ticks) | $301 | 30% |
| A 1-minute timer only | 44 s later (451 gold ticks) | $794 | 79% |
| A 15-minute timer only | 14 min 44 s later (6,590 gold ticks) | $439 | 44% |
What the table says
A guard that listens only to its own chart sees gold only when EURUSD ticks. It acted almost two seconds late, with almost a third of the room gone.
The 1-minute timer acted 44 seconds late. By then 79% of the room was gone: one more dollar of gold, and the firm's line would have been next.
The 15-minute timer acted almost fifteen minutes late, yet it lost less than the 1-minute timer, because gold had bounced by the time it woke up. A slow check's result depends on where the price happens to be when the timer fires: luck, not protection.
The 1-second timer matched the fastest check here only because its next beat came a fraction of a second after the crossing. Its worst case is a full second later, and we did not measure that case, so we will not put a number on it. In the 1.9 seconds the chart-tick guard needed, the loss had already grown to $301.
Even an instant check cannot beat a jump
Speed has a floor, and the market sets it. On 30 January 2026 gold moved so fast that the very tick which crossed our guard's line was already $361 beyond it, with a position sized so that a 1% move costs 4.5%. The guard acted on that same tick; more than a third of the room went in one price step.
Gaps are worse. On our tester's data gold's last tick before the weekend of 26 and 27 September 2026 was 4,285.24 and the first after it 4,277.56. Over Christmas 2025 the market closed on 24 December at 20:00 server time and reopened on 26 December at 10:00, 0.46% higher. No check, however fast, can close inside a closed market: the first tick after the break is the first chance, and the whole gap is already in the loss. The defences are the size of the room and not holding through breaks your firm forbids.
Prices also fill worse around the releases that cause many of these moves. In our broker's real gold ticks around US CPI releases from 2021 to April 2024, the widest spread in the release minute was 6.1 times the usual in the median release and 14.6 times in one release out of ten.
The other timer: when the day starts
Many guards also reset their daily limit at a fixed server hour. A firm does not. FTMO, for example, starts the day at midnight in Prague. On a broker server that runs at New York time plus seven hours, that is 01:00 most of the year and 02:00 for about three weeks a year, when Europe and the United States change their clocks on different dates (in 2026 and 2027: 26 October to 1 November, and 15 to 28 March). A guard reset at 01:00 server then disagrees with the firm for an hour every night about which day a loss belongs to.
A checklist for any equity guard
- Does it see every price change of every symbol you hold, or only its own chart and a timer?
- How often does it check when nothing on its chart moves?
- How much room does it leave below the firm's line, and is that room larger than the gaps and spikes of what you trade?
- Does it compute the firm's day start from the firm's time zone, summer time included, or use a fixed server hour?
- Does it treat a position without a stop, or a pending order, as part of the worst case?
- What happens to your positions before a weekend or a holiday?
A guard is only as good as the moment it looks. Ask when that moment is before you trust it with a challenge.
