REPLAY · EXNESS XAUUSD M1 · 1% RISK PER TRADE
How big a daily drawdown should you expect at 1% risk per trade?
Short answer
About 1% for every trade your bot can take that day, and that full-loss day is common. On a year of Exness gold prices, a 1:2 bot with no edge at 1% risk and three trades a day lost all three on 28.8% of days, a 2.97% drop, and the deepest day in 200 replays was 3.02%. The bigger number is the month: the median month fell 12.35% from its high.
The worst day is the trade count times the risk
With one trade open at a time and a stop on every trade, a day can lose only as many stops as it has trades. At 1% of the balance per trade, three stops in a row cost 2.97%, because each one is 1% of a slightly smaller balance.
| Trades a day (at most) | Days down a full stop per trade | Median day, low point | Worst day of the year | Deepest day in 200 replays | Median month, high to low |
|---|---|---|---|---|---|
| 1 | 66.2% of days (1.00%) | 1.00% | 1.00% | 1.05% | 6.74% |
| 2 | 43.5% of days (1.99%) | 1.66% | 1.99% | 2.03% | 9.88% |
| 3 | 28.8% of days (2.97%) | 1.74% | 2.97% | 3.02% | 12.35% |
| 5 | 12.7% of days (4.90%) | 2.02% | 4.90% | 4.96% | 16.39% |
| 8 | 3.8% of days (7.73%) | 2.59% | 7.73% | 7.83% | 20.95% |
All figures are at 1% risk and measured on equity, open trades included, from the balance at the start of the day. "Worst day of the year" is the median across the 200 replays.
At three trades a day the account fell a full three stops on 28.8% of days, about one day in 3. That worst case is an ordinary day: in the median replay the worst day of the year was simply the three-stop day, 2.97%. At eight trades a day the full loss came less often, 3.8% of days, but cost 7.73%.
The losses did not bunch up. On days with all three trades taken, the share on which every trade lost money was 30.4%, a little above the 28.8% because a trade closed at 4:55 pm ET can lose less than a full stop. Treating each trade as a coin flip that loses 67.4% of the time gives 30.6%. So for your own bot the chance of a full losing day is its loss rate multiplied by itself once per trade, as long as its trades are as independent as these were.
A day went past its trade count only when price jumped through a stop. At three trades a day 185 of 102,487 stops (0.18%) filled beyond their level, the worst at −1.19R, and the deepest day in all 200 replays was 3.02%. At eight trades a day one stop filled at −2.52R, on 28 November 2025.
How we replayed it
- Prices: Exness XAUUSDm one-minute bars, 356,223 of them from 1 August 2025 to 4 August 2026, 260 gold trading days. A gold day runs from 6 pm to 5 pm ET (5 pm to 4 pm CT) [23:00 to 22:00 GMT in US winter, 22:00 to 21:00 GMT in summer].
- Entries: the bot drew its entry minutes at random between 7:00 pm and 3:59 pm ET (6:00 pm to 2:59 pm CT) [00:00 to 20:59 GMT in US winter, 23:00 to 19:59 GMT in summer], skipping the first hour after the reopen. One trade at a time: an entry waited for the open trade to close. Direction was a coin flip.
- Each trade: a stop $5 from the fill and a target $10 away (1:2), with the spread Exness recorded on that minute (median $0.16). A minute that touched both levels counted as a stop. Anything open at 4:55 pm ET (3:55 pm CT) [21:55 GMT in winter, 20:55 in summer] was closed, so no trade held over the daily break or the weekend.
- Size: 1% of the balance at the moment of entry, with 0.5% and 2% worked out from the same trades. Lots were not rounded to gold’s 0.01 step.
- 200 replays of the year for each trade count, each with its own entry minutes and directions: 154,962 trades at three a day. The replays share one year of prices: they show what a bot could have had in that year, not in other years.
Equity dips deeper than the closed balance
A daily limit measured on equity counts the open trade. On the median day at three trades and 1% risk, equity fell 1.74% below the day’s opening balance, against 1.00% on closed trades alone. On 29.0% of days the equity low was at least half a stop deeper than the closed low.
Only 26.4% of days ended in profit, and they dipped too: their median low point was 0.64% below the opening balance, and 30.7% of them were down a full stop before they finished green. At 1:2 a day of one win and two losses still ends a little red once the spread is paid.
Daily drawdown vs max drawdown
The daily number has a ceiling. The drawdown across days does not. At 1% risk and three trades a day the worst day of the year was 2.97%, but the median calendar month fell 12.35% from its high to its low, and one month in ten fell more than 20.10%. A month’s drawdown came to about 4.2 full losing days, built mostly from ordinary red days.
| Risk per trade | Three stops in a day | Median day, low point | Days past a 5% limit | Median month | One month in ten |
|---|---|---|---|---|---|
| 0.5% | 1.49% | 0.87% | 0.0% | 6.31% | 10.55% |
| 1% | 2.97% | 1.74% | 0.0% | 12.35% | 20.10% |
| 2% | 5.88% | 3.47% | 35.4% | 23.54% | 36.57% |
Three trades a day, same entries in every row. Halving the risk roughly halves each drawdown. At 2% the three-stop day costs 5.88%, so a 5% daily limit broke on 35.4% of days: every full losing day plus the days that lost two stops and then dipped more than half a stop on the third trade.
Over the whole year the no-edge bot lost a median 30.6% at 1% risk, mostly to the spread, and its deepest fall from a high was a median 46.7%. A daily limit never sees a loss like that coming, because no single day breaks it. The drawdown recovery calculator shows what a fall of that size takes to win back.
What to set on a bot that risks 1% a trade
- Multiply the most trades the bot may take in a day by its risk. That is your normal bad day, and on gold it came about once every 3 days at three trades. If your daily limit cannot take it, cut the trades or the risk.
- Leave a full stop of room under a prop firm’s limit if the firm counts equity, because a stop can fill past its level. Our FTMO simulation counts the challenges that ended on the daily limit at each risk.
- Check what your bot’s daily limit counts. The daily limits in the Tech Kick Trading Bot count closed trades only and block new entries, so a trade already open can carry the day past them. On a prop firm account run it under Custom or plain Conservative, which leave out the clock rules, so the daily loss limit applies to every trade, and check the lot and the limit against the firm’s rules before the first trade.
- Judge a bot by the month. A 12.35% month was the median for a 1% bot with no edge at three trades a day, so a month like that does not show on its own that a bot is broken.
Common questions
Does a 1:1 bot have smaller daily drawdowns than a 1:2 bot?
Its worst day is the same: every trade of the day stopped out. It gets that day less often. A 1:1 bot needs about 50% winners to break even, and at 50% all three of a day's trades lose 12.5% of the time. At 30%, about what our 1:2 replay won, it is 34.3%. That arithmetic assumes each trade is independent of the last, which held on gold in the replay.
Should a bot risk 1% of the starting balance or of the current balance?
Within one day it barely matters. Sized from the current balance, three stops at 1% cost 2.97%; sized from the starting balance they cost 3%. On a small account rounding matters more. Gold’s smallest lot is 0.01, and in our tests on a $1,000 account percent sizing rounded the lot down so far that the real risk came out as much as 46% below the intended risk.
How much does the spread cost a bot that risks 1% a trade?
In the replay the median gold spread was $0.16, which is 0.032R on a $5 stop. With the minutes that touched both the stop and the target counted as losses, the average random trade lost −0.039R. At three trades a day that adds up: the median replay year ended 30.6% down. A wider stop makes the same spread a smaller share of the risk.
Does holding trades overnight change the daily drawdown?
The replay cannot say: it closed every trade by 4:55 pm ET, so nothing crossed the daily break or the weekend. A trade held through the reopen can open beyond its stop and lose more than its 1%. Our weekend gap study measures those gaps.
Further reading
- Maximum Drawdown: Meaning, Formula, Prop Firm RulesMaximum drawdown is the largest fall from a peak to the low that follows. How to…
- Minimum Account Size for a Gold Trading Bot01 minimum lot sets the smallest risk you can take. 5% risk, and a $3 stop needs…
- Prop Firm Drawdown CalculatorConvert a firm's daily and maximum drawdown limits into the consecutive losses t…
Educational information only, not financial advice. Trading leveraged products carries substantial risk of loss. Last updated 2026-11-17.