MT5 Strategy Tester modelling controls which historical price events a native MQL5 Expert Advisor receives during a test. The modes are Every tick based on real ticks, Every tick, 1 minute OHLC, Open prices only and Math calculations. They differ in recorded versus generated data, intrabar detail and event timing. Execution delay and the historical Forward split are separate settings.
MT5 backtest modes at a glance
As of 25 September 2026, MetaQuotes documents the following modes. The detailed tick-generation page supplies qualifications that the shorter settings descriptions omit.[1][2]
Scroll horizontally to read every column.
| Mode | What the EA receives | Main limitation |
|---|---|---|
| Every tick based on real ticks | Recorded ticks accumulated by the broker, where available. | Missing tick minutes can be generated from minute bars. |
| Every tick | Generated OHLC and intermediate ticks derived from minute data. | The intraminute path is constructed, not a record of every original quote. |
| 1 minute OHLC | Up to four price events for each minute: open, high, low and close. | Intermediate movements are absent; event times are generated. |
| Open prices only | The EA's OnTick handler runs at the selected timeframe's bar opening. | Intrabar EA reactions are omitted; timeframe access is restricted. |
| Math calculations | Initialisation, tester and deinitialisation callbacks. | No historical prices or ticks are loaded; this is not a trade replay. |
The modes answer a modelling question inside the backtesting process. They do not select a strategy, remove research bias or validate an external alert route.
What does Every tick based on real ticks actually use?
Real ticks are quote or trade events accumulated by a broker from exchanges and liquidity providers. They preserve recorded detail that a reconstruction from minute bars cannot supply. MetaQuotes also notes that the spread can change within a minute when real ticks are used.[2]
The mode name is not proof of complete recorded coverage. When a minute bar exists without corresponding tick data, MT5 generates ticks for that minute using the Every tick method. When ticks exist but the corresponding minute bar does not, MetaQuotes says those ticks are ignored.[2]
Record the broker server, symbol, requested dates and available history. A dataset mixing recorded and reconstructed minutes cannot be described simply as an untouched quote-by-quote record. Read the detailed coverage and tester messages before making that claim.
Recorded ticks still feed a simulator. They do not recreate an actual broker order queue or establish the outcome of a later trade request. Separate the price history's detail from the assumptions used to turn an EA request into a simulated execution.
How do Every tick and 1 minute OHLC differ?
Every tick builds a path from cached one-minute records, including intermediate ticks. The algorithm uses the minute's open, high, low, close and tick volume. More generated points add detail to the model, but do not recover an unknown historical sequence.[2]
For generated ticks, MetaQuotes uses the spread recorded for the corresponding minute bar. The ask is calculated from bid plus that spread. Unlike real ticks, the generated spread therefore does not vary inside that minute.[2] The bid-ask spread guide explains why quote side matters for entries and stops.
1 minute OHLC keeps only the minute's OHLC events. When a minute has tick volume above four, four prices are generated; lower-volume bars follow the documented reduced formation rules. The prices come from the history, while their arrival times are generated.[2]
Choosing an H1 testing timeframe does not turn this mode into four events per hour. The selected timeframe remains H1, while the price-event stream comes from the one-minute bars inside it. Keep the chart period and the modelling mode in separate fields in your test record.
1 minute OHLC calls to OnTick = 4 × N, when all N minute bars have tick volume above four
N is the number of complete minute bars in the interval being counted. OnTick is the EA handler called for a new tick event. This count describes the stated tick model, not every possible EA callback or internal tester action.[2]
Worked example: the same hour, different event streams
Illustrative numbers, not a recommendation or a strategy result. Consider the hour from 10:00 inclusive to 11:00 exclusive, with 60 complete minute bars. Assume every minute's tick volume exceeds four and the tester's selected timeframe is H1.
- 1 minute OHLC: 60 × 4 = 240 OnTick calls.
- Open prices only: one OnTick call, at the 10:00 H1 opening.
- Every tick: a generated stream whose count depends on the source minute records and generation rules.
- Real ticks: recorded events where available, with the documented minute-data fallback.
The difference is not only the count. Suppose one illustrative minute has open 100, high 104, low 98 and close 102. Both 100 → 104 → 98 → 102 and 100 → 98 → 104 → 102 share those four summary values.
An EA reacting first to 103 or 99 would encounter those levels in a different order on the two paths. The OHLC record alone cannot establish which event came first. An exact calculation over an assumed path remains conditional on that path.
What does Open prices only leave out?
Open prices only calls the EA's OnTick handler at the opening of each bar on the selected testing timeframe. MetaQuotes still describes OHLC generation inside the tester. Do not translate the mode name into a claim that every internal price check has disappeared.[2]
The distinction matters if the EA modifies a stop, reacts to an intermediate quote or updates a trailing rule during a bar. A stop or pending order can also trigger without another call to the EA's OnTick handler. MetaQuotes warns that trigger prices can differ from the specified levels in this mode.[2]
- Lower timeframes: the EA cannot access data below the testing timeframe.
- Higher timeframes: accessed periods must be multiples of the testing timeframe. MetaQuotes gives M20 to H1 as permitted, but M20 to M30 as unavailable.
- Other symbols: similar restrictions apply, based on the first timeframe accessed for each symbol.
- Delay: Random Delay cannot be used with Open prices only.
- Long periods: W1 and MN1 are exceptions to the ordinary opening schedule; bars are generated daily.[2]
An entry rule evaluated once per bar does not establish that the entire EA is insensitive to intrabar events. Exit logic, pending orders and other symbols may still depend on information this mode omits. Compare individual events when testing that assumption.
What is Math calculations mode for?
Math calculations uses the tester for mathematical optimisation. It does not download historical data or symbol information and does not generate ticks. Only OnInit(), OnTester() and OnDeinit() are called in sequence.[2]
MetaQuotes presents the mode for mathematical calculations that require parameter selection. Calling this a price-history backtest would misdescribe the experiment. A parameter-search result in this mode establishes nothing about market fills unless those assumptions are separately supplied and evaluated.
What do MT5 tester delays simulate?
The Execution setting inserts a simulated delay between an EA's trade request and its execution. Prices can change in that interval. The documented choices are No Delay, Random Delay and Fixed Delay, with preset or custom values for the fixed option.[1]
No Delay supplies an idealised case: the documentation describes requested-price execution without requotes. Under Instant Execution, delayed requests can receive a requote if the price difference exceeds the permitted deviation. These are model behaviours, not measurements of your current connection.
Placing a pending order and activating it are different events. MetaQuotes applies the simulated delay to EA operations such as order placement and stop modification. It does not add that same network delay when an already-placed pending order activates on the simulated server.[1]
The delay setting is not a reconstruction of TradingView calculation, alert triggering, webhook delivery and PineConnector processing. Those stages precede the EA's broker request in an alert-driven workflow. A native-EA test cannot establish their observed behaviour.
Is the Forward option live forward testing?
MT5's Forward option reserves the later part of the selected historical date range. The documented choices include a half, a third, a quarter or a custom start date. The later period remains historical data; no waiting for new market observations is implied.[1]
During optimisation, MT5 searches parameters on the earlier period and evaluates selected candidates on the later one.[3] Choosing the final candidate by inspecting Forward Results uses that period for selection. Preserve an independent evaluation for the resulting choice, as the out-of-sample guide explains.
For repeated development and later evaluation windows, see walk-forward analysis. Changing the schedule after inspecting its combined results is itself another research choice.
Can MT5 test a TradingView strategy sent through PineConnector?
A native MQL5 EA is tested in MT5. A Pine Script strategy runs its historical simulation in TradingView's broker emulator, described in the TradingView strategy report guide.[1][4] An MT5 modelling mode does not turn Pine Script into a native EA.
Historical replay of the complete PineConnector webhook route in the MT5 Strategy Tester is not documented; test on demo. PineConnector's recommendations separate historical chart checks from realtime alert observation.[5]
For execution evidence, follow the documented setup test: match the alert log, processing result and intended broker trade.[6] Keep condition, alert, webhook, PineConnector processing, EA request, broker acceptance, deal and position distinct. MT5's order definitions distinguish the final stages.[7]
What belongs in a reproducible MT5 test record?
- Implementation: EA version, input parameters and required symbols or timeframes.
- History: broker server, exact symbols, requested dates and any shifted start date noted in the tester Journal.
- Price model: modelling mode, recorded tick coverage and spread treatment.
- Execution: delay setting, execution mode, order types and relevant volume or stop constraints.
- Account assumptions: deposit currency, leverage, netting or hedging, commission and swap treatment.
- Research boundaries: development dates, Forward dates and every comparison that influenced selection.
MetaQuotes allows custom account and symbol settings, so the connected broker's name alone does not identify the full test configuration. Its help also states that profit calculation in pips removes commission, swap and margin control.[1] Record that option explicitly if used.
More detailed ticks cannot repair look-ahead or survivorship bias. Nor do matching aggregate results prove that two modes generated the same orders. Diagnose the first differing decision or fill before comparing report totals.
Frequently asked questions
What does Every tick based on real ticks mean in MT5?
Every tick based on real ticks uses historical ticks accumulated by the broker where available. Recorded spreads can vary within a minute. If a minute bar exists without tick data, MT5 generates ticks for that minute. The mode therefore requires a coverage check and remains a simulation of EA execution.
Does 1 minute OHLC force an MT5 test onto M1?
No. The selected testing timeframe and the modelling mode are separate. An H1 test in 1 minute OHLC mode receives OHLC events from each minute inside the hour. With 60 minute bars and tick volume above four in each, the illustrative count is 240 calls to the EA's OnTick handler.
Can Open prices only test intrabar trailing stops?
Open prices only does not reproduce an EA's reactions to every intrabar quote because its OnTick handler runs at the selected bar's opening. The tester can still process stops and pending orders internally. An EA that updates trailing levels during a bar depends on behaviour the reduced event stream omits.
Do MT5 tester delays reproduce TradingView webhook delivery?
MT5 tester delays model the interval between an EA's trade request and simulated execution. They do not reconstruct TradingView alert creation, triggering, webhook delivery or PineConnector processing. Pending-order placement receives the configured delay, but activation of an already-placed pending order does not receive the same added network delay.
Is MT5 Forward testing a live demo test?
MT5's Strategy Tester Forward option uses the later portion of a selected historical range. It evaluates an EA's parameters beyond the earlier development period. A live demo test observes incoming data and actual requests in a broker demo account. These checks answer different questions and require separate records.
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
- How to backtest a trading strategy
- Why TradingView and MT5 backtest results differ
- Out-of-sample testing and holdouts
- MT5 Strategy Tester report explained
Sources
- MetaQuotes – Strategy Testing: settings, delays, Forward and account assumptions, accessed 25 September 2026.
- MetaQuotes – Real and Generated Ticks, accessed 25 September 2026.
- MetaQuotes – Strategy Optimization: historical Forward evaluation, accessed 25 September 2026.
- TradingView – Strategies: the broker emulator, accessed 25 September 2026.
- PineConnector – Recommendations: backtests and repaints, accessed 25 September 2026.
- PineConnector – Test your setup, accessed 25 September 2026.
- MetaQuotes – Basic Principles: orders, deals and positions, 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.
