Reading an MT5 Strategy Tester Report: What Each Field Means and What Misleads

When an MT5 Strategy Tester run finishes, the Backtest tab fills up with dozens of figures. Most people look at net profit and the graph, but the real risk sits in the other fields. This article goes through them one at a time: what each means, where the trap is, and what an acceptable value looks like (as our own rule of thumb, not an industry standard). Every field is computed from one example trade list, used throughout.

The example used throughout

The trade list below is made up to explain the method. It is not the result of a real system or market. The deposit is 1,000 USD and trades are opened one at a time. Trades 1–10 fall in 2025 and trades 11–20 in 2026. “Deepest” is the worst floating loss while the trade was open (a losing trade closes at its deepest point). We assume 200 USD of margin is used whenever a trade is open.

Trade	Side	Result	Deepest	Trade	Side	Result	Deepest
1	Buy	+42	−10	11	Buy	+65	−12
2	Sell	−25	−25	12	Sell	+18	−5
3	Buy	+38	−8	13	Buy	+47	−20
4	Buy	+55	−15	14	Sell	−26	−26
5	Sell	−48	−48	15	Buy	+88	−18
6	Buy	−45	−45	16	Sell	+48	−9
7	Sell	+20	−110	17	Buy	−31	−31
8	Buy	−32	−32	18	Sell	+36	−14
9	Sell	−27	−27	19	Buy	+40	−6
10	Buy	−30	−30	20	Buy	−29	−29

We computed every field with a script, following the report’s definitions, and laid them out the way the MT5 Backtest page does.

Test settings

History Quality tells you how complete the price history used by the chosen mode was. The trap is reading 100% as “as accurate as the real thing”. The figure does not say whether the ticks were real or generated. Rule of thumb: it should be close to 100%, and for an EA that is sensitive to price movement inside the bar, the Journal tab must also confirm that the tested period has real ticks.

Bars and Ticks are the number of bars on the tested timeframe and the number of ticks the tester modelled. The trap is ignoring them. If Ticks is close to Bars, the tester saw only a few prices per bar, and a tight trailing stop or pending orders will look better than they really are. More in MT5 Strategy Tester modes.

Initial Deposit is the starting balance. Every dollar figure depends on the deposit and the lot you set. The trap is comparing a 5,000 profit from one test with 3,000 from another that used a different deposit or lot. Compare in percent, or bring the lots to the same size first.

Profit

Total Net Profit is the final balance minus the deposit, already including the swap and commission the tester charged. In the example, 1,204 − 1,000 = 204. The trap is that the tester’s commission may not match your real account. Open the deals list and check that the Commission column shows what your account actually pays.

Gross Profit and Gross Loss are the sum of all winning trades and the sum of all losing trades: 497 and −293 in the example. They feed the next fields.

Profit Factor = Gross profit ÷ |Gross loss| = 497 ÷ 293 = 1.70. The trap is a high value from very few trades. A system that made 3.0 over 15 trades tells you nothing yet. Our rule of thumb: below about 1.1 is thin, because live slippage can easily flip it, and about 1.3 or more over several hundred trades and several years is more comfortable.

Expected Payoff = net profit ÷ number of trades = 204 ÷ 20 = 10.20 per trade, the average you make each time you open a trade. The trap is not comparing it with the cost per trade. If a trade makes only a few times the spread, costs the backtest can’t see, such as slippage, can eat all of it. In our own work on gold at 0.01 lot, the bar is at least 0.20 USD per trade in the weakest year, against an average spread we measured at about 0.27 USD per trade. More in Spread cost per trade.

Recovery Factor is net profit divided by the maximum drawdown in money. It tells you how many times the deepest hole the whole period’s profit covers. Divided by equity drawdown, the example gives 204 ÷ 203 = 1.00. Divided by balance drawdown, it gives 204 ÷ 162 = 1.26. The trap is not knowing which drawdown the figure in your report used. The safe habit is to divide by Equity Drawdown Maximal yourself every time. Rule of thumb: only compare versions over the same period, and if a multi-year test comes in below about 2, the profit for the whole period is only about two holes deep.

Sharpe Ratio is a ratio of return to volatility. Higher means smoother. We don’t confirm here exactly which period and reference rate MT5 uses to compute it, so the example leaves it out. The trap is comparing it with a fund’s Sharpe ratio or one from other software. It is only comparable between MT5 tests with the same settings, and it works as a tie-breaker, not as the main criterion.

Drawdown and Margin level

Drawdown is how far the account falls from its previous peak. The report splits it into Balance (closed trades only) and Equity (including open trades), each in three forms.

Field	Meaning	Balance	Equity
Absolute	Largest fall below the deposit	1,000 − 948 = 52	1,000 − 907 = 93
Maximal	Largest fall from a peak, in money	162 (14.59%)	203 (18.29%)
Relative	Largest fall from a peak, in percent	14.59% (162)	18.29% (203)

In the example, Maximal and Relative come from the same hole, so they match. When an account grows a lot, the two can come from different periods. The details are in Drawdown: equity vs balance.

Look at trade 7. Before it opens, the balance is 1,017. While it is open, it falls as far as −110, so equity drops to 1,017 − 110 = 907. From the 1,110 peak, that is 203 gone, or 203 ÷ 1,110 = 18.29%. But the trade closes at +20, so the balance never sees this hole. The deepest hole in the balance is only 1,110 down to 948, after trade 10.

Always size lots from Equity Drawdown Maximal, because it is the hole your account money actually went through. Balance drawdown only measures the losses you have already closed.

The gap between the two can be huge. In our backtest on gold, a basket system that holds trades until they reach a small target, with no Stop loss per trade, reported a balance drawdown of only 0.31% in 2024 (ticks generated from 1-minute bars), while its real equity drawdown was 44.19%, about 142 times larger. Backtest results do not predict future results.

Another trap is daily drawdown. Some tools, and the rules of prop-style accounts, measure the fall within a single day, which is always less than or equal to the equity drawdown over the whole test. In our own work we once nearly drew the wrong conclusion by dividing profit by daily drawdown, which made profit-to-risk look far better than it was. Don’t treat these three as interchangeable.

Our rule of thumb: decide first how far you will let the account fall, then make sure the worst equity drawdown in the test is no more than about half of that limit, because the future can go deeper than what you saw. And remember that an 18.29% hole needs a 203 ÷ 907 = 22.38% gain just to get back to where it was. Try other numbers with the recovery calculator.

Margin Level in the report is the lowest value during the test = equity ÷ margin used × 100. In the example, the low comes during trade 7: 907 ÷ 200 × 100 = 453.50%. The trap is that it depends on the leverage set in the tester, which may not match your real account, and the Stop out level varies by broker. Check the contract specification or your account terms. Rule of thumb: the lowest value should stay several times above the Stop out level. More in Leverage, margin and Stop out.

Trade counts and per-trade statistics

Total Trades is the number of closed trades (Total Deals counts entries and exits separately, so it is roughly twice as large). The trap is trusting every field when there are few trades, and comparing drawdown between versions with very different trade counts. A system that trades less often tends to show a lower drawdown without actually being safer. Rule of thumb: below about 100 trades, every field still swings a lot. The 20-trade example here only illustrates the method.

Short Trades (won %) and Long Trades (won %) are the number of Sell and Buy trades and the share of each that won: 8 Sells winning 50.00% and 12 Buys winning 58.33% in the example. The trap is one side looking great because the test period trended one way. Rule of thumb: if the two sides differ a lot, split by year and check whether the strong side still holds in years when the market went the other way.

Profit Trades and Loss Trades (% of total) are the counts of winning and losing trades: 11 of 20 won, or 55.00%. The trap is reading the win rate on its own. Compare it with the break-even win rate = 1 ÷ (1 + RR). Here RR = 45.18 ÷ 32.56 = 1.39, so break-even is 41.88%, and a 55% win rate is above it. The full reasoning is in Win rate vs break-even.

Largest profit trade and Largest loss trade are the biggest win, 88, and the biggest loss, −48. The trap is that the whole profit comes from a few trades. Rule of thumb: remove the biggest one or two winners. If net profit turns negative, the system relies on rare events. Here 204 − 88 = 116, still positive. The largest loss should be close to the Stop loss you set. If it is much bigger, find out in the deals list what happened to that trade.

Average profit trade and Average loss trade are 497 ÷ 11 = 45.18 and −293 ÷ 9 = −32.56. Use them to find the system’s real RR, which already includes everything that happened, not the RR you intended.

Maximum consecutive wins ($) and Maximal consecutive profit (count) are often confused. The first picks the winning run that is longest by number of trades and shows its money in brackets: trades 11–13, three wins in a row, 130 in total. The second picks the run that made the most money and shows the trade count in brackets: trades 15–16, 136 from 2 trades. The loss side reads the same way: the longest losing run is 3 trades (trades 8–10, −89 in total), and the most expensive run is −93 from 2 trades (trades 5–6). The trap is assuming the longest losing run in the past is the ceiling for the future. Rule of thumb: leave room, in your head and in your account, for a longer losing run than the report shows. See Losing streaks are normal.

The equity graph and the deals list

The Graph tab plots balance and equity over time. Look for three things: how much of the profit comes from one short stretch, how often equity dips below balance (a sign the system holds losing trades), and whether there are long periods where the line is flat or falling. A line that rises too straight for a system with no Stop loss often hides a hole in equity.

The deals list in the Backtest tab shows every entry and exit with time, direction, lot size, price, commission, swap and profit. Use it to check that the lot is what you intended, that commission is really charged, whether trades are held over weekends, and whether trades open at times when the spread is unusually wide. Better still, export it and do your own sums, such as splitting by year or finding profit per trade for each time window.

Split by year, and compare with raising the lot

The report adds good years and bad years together. In the example, net profit is +204, but split by year, 2025 loses −52 and all the profit comes from 2026 at +256. If a system only makes money in one kind of market, it can have another year like 2025 whenever the market changes. See When the market regime changes.

The gold basket system above is a good example. By balance drawdown, the worst year was 2022 at 10.04%. By equity drawdown, the worst year was 2024 at 44.19%, with the other years between 8.45% and 16.25% (2026 on real ticks, 2022–2025 on ticks generated from 1-minute bars). Only by splitting by year and reading the right field at the same time do you see which year sets the ceiling for your lot size. Backtest results do not predict future results.

Another step people often skip is comparing with simply raising the lot. If the deposit is large enough and margin isn’t a limit, raising the lot by some factor raises profit and drawdown in money by about the same factor (rounded to the lot step), so the Recovery factor stays the same. Say version B adds a rule that opens extra trades and makes 300 at an equity drawdown of 320. It looks better than the original’s 204. But raise the original’s lot by 320 ÷ 203 = 1.576 times and it makes about 204 × 1.576 = 321.6 at the same drawdown. Version B loses, despite the higher profit.

Every change to a system has to beat simply raising the lot at the same drawdown. If it doesn’t, you paid for extra complexity and cost for nothing.

If you plan to raise the lot as the account grows, read Compounding: growing position size with the account.

Common mistakes

• Looking only at net profit and the graph, without opening the Drawdown section
• Sizing lots from balance drawdown or daily drawdown instead of Equity Drawdown Maximal
• Trusting a History Quality of 100% without checking the Journal for whether the period ran on real ticks
• Comparing Profit factor or drawdown between tests with very different trade counts, deposits or lots
• Reading the win rate without comparing it with the break-even from Average profit and Average loss
• Looking at a multi-year total without splitting by year, and missing that the profit came from one year
• Adding a new rule and celebrating higher profit without comparing it with raising the lot at the same drawdown
• Forgetting that a backtest has no slippage from real order execution, so live results are usually lower. More in Why live results usually trail the backtest