GLOSSARY · 13 TERMS
MetaTrader terms
The platform terms you meet when running an Expert Advisor: the switches that let it trade, the tabs that say why it did not, and the tester settings that decide whether a backtest means anything.
Algo Trading button
Also called: AutoTrading button.
The Algo Trading button in the MetaTrader toolbar is the global switch that permits Expert Advisors to place orders, and nothing automated happens while it is off.
It is a single toolbar toggle, green when enabled and red when not, and it overrides every per-EA setting. An EA can be attached, loaded, running its logic and generating signals while this button is off, and it will place no orders at all.
The state does not always survive a restart, and some MetaTrader builds reset it after an update. This makes it the first thing to check whenever a previously working setup stops trading — far more often than not, that is the answer.
A related trap catches people who have enabled it: the EA also needs "Allow Algo Trading" ticked in its own properties dialog, and both must be true. See AutoTrading is disabled for the full checklist.
Expert Advisor
Also called: EA, Forex EA, Expert Adviser, Trading robot, MT5 Expert Advisor, What is an Expert Advisor.
An Expert Advisor (EA) is a program that runs inside MetaTrader 4 or 5 and trades for you: it reads prices, applies coded rules and places, changes or closes orders without anyone clicking. In forex, EA is the usual name for any MetaTrader trading robot.
Read the full entry: Expert Advisor →
Investor password
Also called: Read-only password.
The investor password is a second MetaTrader password that grants read-only access to an account — trades and history are visible, but no order can be placed and nothing can be modified.
Every MetaTrader account has two credentials: the master password, which can trade and withdraw, and the investor password, which can only look. Handing over the investor password lets someone audit the account without being able to touch it, which is what makes independent verification possible at all.
This is the mechanism behind Myfxbook and every similar service, and it is what a serious buyer should ask for. A vendor who will not supply read-only access to the account behind their claims is declining the one check that costs them nothing and proves the most.
The safety limits are worth knowing precisely. Read-only access cannot place, close or modify orders, and cannot withdraw. It can see full position and balance history, and on some brokers it is enough to identify the account holder. It is safe to share for verification and it is not anonymous.
Journal and Experts tabs
The Journal and Experts tabs in the MetaTrader terminal window are the logs that record what the platform and your Expert Advisors actually did, including every error.
The Experts tab shows output from EAs — their own messages plus initialisation and deinitialisation events. The Journal shows platform-level activity: connections, order requests, server responses and rejection reasons.
When an EA is not behaving, these two tabs contain the answer almost every time. A rejected order appears in the Journal with a specific reason code, and the difference between "invalid stops", "not enough money" and "trade disabled" points to three entirely different fixes.
The habit worth forming is checking the Journal before changing any setting. Most support requests about EAs not trading are resolved by reading a message that was already sitting there — our troubleshooting section is organised around the specific messages you will find.
Magic number
A magic number is an integer an EA stamps onto every order it places, allowing it to identify its own positions and ignore trades opened manually or by other robots.
Without one, an EA managing "all open positions on this symbol" would also manage your manual trades, and two EAs on the same symbol would fight over each other's positions. The magic number is what makes several systems able to share an account safely.
It must be unique per EA per account. Running two copies of the same EA on different symbols with the same magic number is a common misconfiguration, and the symptom is bizarre — positions being closed by an instance that did not open them.
Ours is 202401, and it is the mechanism by which the engine detects which of its own trades have closed and records the result to the performance tracker. That tracker is what drives strategy selection, so a magic number collision would corrupt the tournament rather than merely confusing the trade list.
Market Watch
Market Watch is the MetaTrader panel listing available instruments with live bid and ask prices, and the place where each symbol's full contract specification is found.
By default it shows only a subset of what the broker offers. Right-clicking and choosing "Show All" reveals the complete list, which is frequently necessary because gold is often hidden behind a metals or CFD grouping rather than sitting with the FX pairs.
The specification dialog for each symbol is the authoritative source for contract size, minimum volume, stop level, swap rates and trading hours. These vary by broker and account type, and every one of them can break an automated system that assumed a different value.
A symbol also has to be visible in Market Watch for an EA to reference it reliably. This is a common cause of a bot working on one terminal and not another with apparently identical settings.
MetaTrader 4
Also called: MT4.
MetaTrader 4 is the older MetaQuotes platform, still widely used for forex despite being superseded, with its own EA language and a much larger legacy library of robots.
Released in 2005, MT4 remains popular largely because of inertia and the enormous existing library of indicators and EAs written for it. MetaQuotes stopped meaningful development years ago and no longer licenses it to new brokers.
Technically it is more limited: fewer timeframes, no depth of market, a weaker single-threaded strategy tester and no native support for exchange-traded instruments. Its backtesting in particular is known to produce optimistic results relative to MT5's multi-threaded, real-tick tester.
For a new automated system there is no good reason to choose MT4, and our engine targets MT5 only. Traders arriving from MT4 should expect the account model difference to matter most — MT5 defaults to netting on many brokers where MT4 was always hedging, which changes how multiple positions on one symbol behave.
MetaTrader 5
Also called: MT5.
MetaTrader 5 is the trading platform most retail forex and CFD brokers offer, and the environment in which Expert Advisors — automated trading programs — actually run.
MT5 combines charting, order entry, a strategy tester and a programming environment in one desktop application. Its significance for automation is that the EA runs inside the terminal, on your machine, rather than on the broker's servers — which means the terminal must be open, logged in and connected for anything automated to happen.
It is not simply a newer MetaTrader 4. The two use incompatible programming languages, incompatible account models and separate marketplaces, so an MT4 EA cannot be run on MT5 without being rewritten. See MT4 vs MT5.
The desktop terminal is Windows-only in practice, and that single fact determines the architecture of every serious MT5 bot. It cannot run in a browser, on iOS or on Android, so continuous operation requires either a machine left on permanently or a VPS. Our Mobile + PC plan exists specifically because of this constraint — see can I run a trading bot on my phone.
Modelling quality
Modelling quality is the Strategy Tester's estimate of how faithfully it reconstructed price movement during a backtest, expressed as a percentage.
A high figure means the tester had sufficient data to model intrabar movement rather than guessing it. A low figure means large parts of the test were interpolated, and any result from such a run should be treated as indicative at best.
The percentage is not a quality score for the strategy and is frequently misread as one. It describes the data, not the edge — a 99% modelling quality backtest of a bad strategy is an accurate simulation of losing money.
The deeper point is that modelling quality only measures fidelity to historical prices. It says nothing about spread variation, slippage, requotes, or whether your broker would actually have filled those orders. Those gaps are why live results underperform even a well-modelled backtest.
MQL5
MQL5 is the C-like programming language used to write Expert Advisors, indicators and scripts for MetaTrader 5, and also the name of the associated marketplace.
The language is close enough to C++ to be familiar to anyone who has written either, with built-in access to price series, indicator handles and trading functions. Programs are compiled to a binary EX5 file, which is what is distributed to users — source is not required to run one.
That compilation step is why buying an EA rarely means seeing its logic. You get a binary that trades your account, and its actual behaviour can only be inferred from documentation and observed trades. This is a genuine reason to prefer vendors who publish specifics about what their system does and does not do.
MQL5 is also the marketplace where most retail EAs are sold, with a rating system and a public track record on demo and live accounts. For a buyer it is one of the few places where a vendor's claims are attached to a verifiable history, which makes it worth checking even for products sold elsewhere.
Strategy Tester
The Strategy Tester is MetaTrader's built-in backtesting and optimisation engine, which replays historical data through an EA to estimate how it would have performed.
MT5's tester supports several modelling modes, from open prices only through to real ticks. The differences are enormous — a strategy that looks profitable on open prices frequently collapses under real tick modelling, because the cheaper modes never test what happened inside the bar.
Optimisation, which sweeps parameter ranges to find the best-performing combination, is the tester's most dangerous feature. Running thousands of combinations guarantees that some will look excellent on the sample by chance, and selecting the best is a highly efficient way to produce an overfitted system.
Our own testing does not use the Strategy Tester at all. We replay real broker bars through the same signal code the live engine runs, with the real tracker driving strategy selection, so there is no separate backtest implementation that could diverge from live behaviour. Reimplementing the strategy for testing is a common and invisible source of false results.
Symbol suffix
A symbol suffix is the broker-specific text appended to an instrument name — XAUUSDm, XAUUSD.raw, XAUUSDpro — which makes hard-coded symbol names fail across brokers.
Brokers use suffixes to distinguish account types and feeds. Exness appends "m" for its standard and cent accounts, giving XAUUSDm; others use a dot notation or a suffix indicating raw pricing. The underlying instrument is the same, the string is not.
For an EA this is a hard failure rather than a degradation. Code looking for "XAUUSD" on a terminal that only offers "XAUUSDm" finds no symbol, and depending on how it is written either does nothing at all or throws an unknown-symbol error on every tick.
The correct handling is to resolve the symbol from the chart the EA is attached to, or to search Market Watch for a match. If you are configuring a bot manually, copy the symbol name exactly as it appears in Market Watch including case — see wrong symbol.
Tick data
Tick data records every individual price change rather than summarising a period into a bar, and it is the highest-resolution history available for testing.
A one-minute bar reduces potentially hundreds of price changes into four numbers. That summary discards the order in which the high and low were reached, which is exactly the information needed to know whether a stop or a target was hit first.
Testing on bar data therefore requires an assumption about intrabar sequence, and the assumption chosen determines the result for every trade where both levels were touched. Optimistic assumptions inflate backtests substantially, and the effect is largest on short-target strategies.
Real tick data is not a complete solution either — it is broker-specific, often incomplete, and heavy to store. What matters is knowing which you used. Our replay runs on real broker bars with an explicit and conservative intrabar rule, and two of our gold specialist strategies cannot be evaluated at all in replay because they need a tick stream the engine does not provide.
Other glossary sections
Further reading
Educational information only, not financial advice. Trading leveraged products carries substantial risk of loss. Last updated 2026-09-24.