The MT5 Strategy Tester report summarises an Expert Advisor's simulated trades, returns and drawdowns over a chosen historical test. Read its data and settings before its profit figures, then separate cash results, trade averages and the path of balance or equity. A report describes that test; it does not establish future returns or verify a TradingView alert's execution at a broker.[1][2]
MT5 Strategy Tester report at a glance
Scroll horizontally to read every column.
| Question | Fields to inspect | What to check next |
|---|---|---|
| What was tested? | History Quality, Bars, Ticks, Symbols, Initial Deposit | EA version, inputs, dates and modelling settings |
| What were the cash results? | Total Net profit, Gross Profit, Gross Loss | Costs, trade count and concentration |
| What happened per trade? | Expected Payoff, average profit/loss, AHPR, GHPR | Cash versus proportional returns |
| What happened between endpoints? | Balance and Equity Drawdown, recovery, sequences | Open losses, peak basis and ordering |
As of 25 September 2026, MetaQuotes' help locates the detailed report on the tester's Results tab. Use the trading metrics library for the definitions behind the fields.[1]
What must match before comparing two reports?
Record the EA version and parameters, broker history, symbols, date window, initial deposit, leverage, costs and execution assumptions. MT5 tests an EA on historical data. Changing the modelling mode changes the available price path, so an identical strategy name does not establish an identical experiment.[2]
History Quality describes the quality of one-minute price data, including incorrect bars and gaps. It is not a score for the strategy or proof that every simulated fill was attainable. Bars, Ticks and Symbols describe the test's data workload, not its number of completed trades.[1]
Total Trades counts trades that fixed a profit or loss; Total Deals counts deals. One order can produce multiple deals, and a position can involve opening, increasing and partial-closing deals. Do not use deals interchangeably with closed trades when reconstructing averages.[1][3]
How do profit, payoff and recovery fit together?
| MT5 field | Meaning | Definition guide |
|---|---|---|
| Total Net profit | Financial result of the test's trades | Net result and average trade |
| Gross Profit / Gross Loss | Sum of profitable / losing trade results; use loss magnitude in ratios | Profit factor inputs |
| Profit Factor | Gross profit divided by absolute gross loss | Profit factor |
| Expected Payoff | Average cash result per trade in the test | Expectancy and payoff ratio |
| Recovery Factor | Net profit divided by maximal cash drawdown | Recovery factor |
| Average profit / loss trade | Mean result within the winning / losing subset | Average win, loss and payoff ratio |
These definitions follow MetaQuotes' testing-report help.[1] “Gross” does not justify subtracting a cost twice. Reconcile the trade results and the test's cost settings before adding an external cost adjustment.
Net profit = gross profit − absolute gross loss
Expected Payoff = net profit / number of closed trades
Profit factor = gross profit / absolute gross loss
Recovery factor = net profit / maximal drawdown in currency
Profit factor has no finite value when gross loss is zero. Recovery factor is undefined when its drawdown denominator is zero. MetaQuotes' help calls the denominator the maximum drawdown without naming the balance or equity series.[1] Record which drawdown series an exported or independently calculated ratio uses; do not substitute a percentage drawdown into a cash denominator.
Worked example: why a zero-result trade still matters
Illustrative, not a recommendation or strategy results. Construct ten closed trades: four gains of 150 USD, five losses of 80 USD and one result of zero after the chosen cost treatment. Gross profit is 600 USD, gross loss magnitude is 400 USD and net profit is 200 USD.
Expected Payoff = 200 / 10 = 20 USD
Profit factor = 600 / 400 = 1.5
Average win / average loss = 150 / 80 = 1.875
The win frequency is 4/10 = 40%, the loss frequency is 5/10 = 50%, and breakevens are 1/10 = 10%. Therefore cash expectancy is (0.40 × 150) − (0.50 × 80) = 20 USD. Using 60% as the loss frequency would incorrectly turn the breakeven into a loss.
Expected Payoff is a sample average, not a promise about the next trade. It is also different from expectancy in R. When initial cash risk varies, calculate each trade's result divided by its own initial risk, then average those R-multiples.
What do absolute, maximal and relative drawdown mean?
MT5 reports each drawdown family for balance and equity. Balance tracks realised account changes; equity also reflects open-position results. A balance report can therefore miss a deep floating loss visible in equity.[1][2]
| Field | Measurement | Common misreading |
|---|---|---|
| Absolute | Initial deposit minus the lowest value below that deposit | Not a decline from a later peak |
| Maximal | Largest cash decline from a local peak to its subsequent trough; associated percentage in brackets | The bracketed percentage need not be the largest percentage decline |
| Relative | Largest percentage peak-to-trough decline; associated cash amount in brackets | Its episode need not have the largest cash loss |
Drawdown at time t = (running peak − current value) / running peak
Maximum drawdown % = maximum of those fractions × 100
Use one named balance or equity series throughout. MetaQuotes also documents special treatment of EA withdrawals: calculations restart from current balance and equity while retaining the largest earlier drawdown. A cash-flow-free example does not reproduce that case.[1]
In the diagram's illustrative sequence, the initial deposit is 10,000 USD. Balance falls to 8,000, later rises to 20,000 and then falls to 17,000, with no intervening lower troughs or cash flows. Absolute drawdown is 2,000 USD; maximal drawdown is 3,000 USD (15%); relative drawdown is 20% (2,000 USD). MetaQuotes' own drawdown calculation example shows the same split: its largest cash drop is 2,500 (31.25%), while its largest percentage drop is 33.3% (2,000).[7]
Read the drawdown guide before comparing those fields across platforms. Matching the name alone is insufficient.
How should you read Sharpe, AHPR, GHPR and LR?
Sharpe Ratio relates return to variability. MetaQuotes specifies a zero risk-free rate in its tester. The general definition uses the mean differential return divided by the standard deviation of that same differential-return series.[1][4]
Dₜ = returnₜ − benchmark returnₜ
Sharpe = mean(Dₜ) / standard deviation(Dₜ)
Do not compare a tester value with a value rescaled to a longer period, or calculated elsewhere, without matching sampling and calculation conventions. See the Sharpe ratio guide. No single Sharpe value establishes the probability of the next trade losing.
AHPR is the arithmetic mean of trade holding-period factors; GHPR is their geometric mean. For a factor gᵢ = 1 + proportional trade return, AHPR = Σgᵢ/N and GHPR = (∏gᵢ)^(1/N). The bracketed percentage expresses the factor minus one.[1]
Illustratively, successive factors of 1.20 and 0.80 give AHPR = 1.00 (0%), but GHPR = √0.96 ≈ 0.979796 (−2.0204%). Their compound factor is 0.96, a 4% reduction. Averaging proportional changes and compounding them answer different questions; neither is cash Expected Payoff.
LR Correlation measures the relationship between the balance graph and its fitted regression line. LR Standard Error measures deviation from that line in money; MetaQuotes cautions that comparison requires similar starting conditions. Smoothness is not evidence that a rule will generalise.[1] Use the metrics library to compare these path measures with outcome measures.
What do Z-score and consecutive-trade fields add?
Z-Score examines runs of wins and losses. MetaQuotes describes positive values as associated with alternation and negative values with clustering. Treat the field as a sequence diagnostic, not as an instruction to predict or resize the next trade. Dependence, small samples and repeated testing can undermine an interpretation.[1]
- Maximum consecutive wins/losses ($): longest winning or losing run, with that run's cash total.
- Maximal consecutive profit/loss (count): largest cash total across same-sign runs, with that run's trade count.
- Average consecutive wins/losses: average length of the corresponding runs.
The longest run need not be the most expensive. Link sequence questions to the losing-streak guide; the observed maximum is not a future ceiling. The report also includes MAE/MFE relationships, holding times and minimum Margin Level. See trade excursion and margin level for those distinct measurements.[1]
Does this report test a PineConnector workflow?
A TradingView strategy uses TradingView's broker emulator.[6] Routing its alerts through PineConnector does not transfer its backtest into MT5's EA tester. Read the TradingView strategy metrics guide for that simulation.
The PineConnector converter documentation distinguishes simulated fills from broker outcomes. Its adapter does not resynchronise their positions, and broker stops are separate from strategy stops.[5] Broker acceptance, a deal and the resulting position require separate evidence. A stop activating is not proof that the position closed.[3]
What are the limits of an MT5 tester report?
- One experiment: the figures describe one EA version, input set, symbol set, date window and tick-generation mode. Every tick, 1 minute OHLC and Open prices only give the tester different price paths.[2]
- Balance can hide open losses: balance values change when a position is closed, so read the equity drawdown fields as well.[2]
- Trade counts depend on deals: partial closes and volume increases add deals to one position, so rebuild averages from closed trades, not deals.[1][3]
- Predictive wording: MetaQuotes says Expected Payoff "is considered to display the expected return of the next trade". This article treats it as a sample average only.[1]
- No execution evidence: a tester run does not show a TradingView alert, webhook delivery, PineConnector processing, an EA order request or a broker fill.
Frequently asked questions
What is Expected Payoff in MT5?
Expected Payoff in the MT5 Strategy Tester is the average cash result per trade in the historical test. Reconstruct it as net profit divided by the relevant closed-trade count, including zero-result trades. The statistic describes the sample and does not establish the next trade's result.
What is MT5 Recovery Factor?
MT5 Recovery Factor relates net profit to maximal drawdown in currency. A ratio calculated with percentage drawdown has different units and is not equivalent. Record the drawdown convention when comparing exports or platforms; a positive ratio does not prove that equity has regained its latest peak.
What does MT5 Z-score tell you?
MT5 Z-score examines the sequence of winning and losing trades for clustering or alternation. MetaQuotes describes negative values in terms of clustering and positive values in terms of alternation. A historical sequence statistic does not supply a reliable forecast for the next trade or a sizing instruction.
What is the difference between AHPR and GHPR?
AHPR averages trade holding-period factors arithmetically, while GHPR averages them geometrically. GHPR reflects multiplication of proportional changes. An illustrative gain of 20% followed by a loss of 20% has a zero arithmetic average but leaves capital 4% lower, before any additional costs.
Reviewed 25 September 2026. Facts were checked against the linked sources on that date. Nothing in this article was tested on a trading account and no code was compiled.
Related reading
- Trading metrics library
- TradingView strategy report metrics
- MT5 Strategy Tester modelling modes
- How to backtest a trading strategy
Sources
- MetaQuotes – Testing Report, accessed 25 September 2026.
- MetaQuotes – Strategy Testing, accessed 25 September 2026.
- MetaQuotes – Basic Principles, accessed 25 September 2026.
- William F. Sharpe – The Sharpe Ratio, accessed 25 September 2026.
- PineConnector – Pine Script converter, accessed 25 September 2026.
- TradingView – Strategies (Pine Script v6 User Manual), accessed 25 September 2026.
- MetaQuotes – Trading Report: drawdown calculation example, accessed 25 September 2026.
PineConnector executes the instructions you send it. It does not select trades, manage money, or hold funds. Trading carries risk, and past performance of any strategy does not indicate future results.
