EA parameters: read the behaviour behind a short input name

An EA input may have a name such as Risk, Step or StartHour. The name fits in a small box; the behaviour can take a paragraph to explain. A value can be accepted by MT5 yet produce a different result from the one you intended. Understanding parameters means knowing what the program will do, when it will do it and what that setting does not control.

A label is a signpost, not the full rule

MQL5 permits an input's displayed name to differ from its code identifier. A friendly label is useful, but neither the label nor the default value fully describes its meaning. Match the program version, input label and parameter documentation before changing anything. The MQL5 input-variable reference explains how displayed names and identifiers work.

Read README.txt for installation and operating notes, then PARAMETERS.pdf for the relevant setting. The on-page parameter list is a useful map; the delivered documentation adds context. With an EX5 edition, configure the inputs the program exposes. With MQ5 source, you can also inspect how a value is used; editing logic requires compilation and a new test.

Answer six questions before changing a value

1. Purpose: does it change an entry signal, order size, exit or reporting?
2. Unit: is the value lots, points, price, seconds, bars or percent? If percent, percent of what?
3. Meaning and limits: what are the allowed values, and what does zero mean? Do not assume zero means disabled.
4. Dependencies: must a mode or switch be enabled first? A fixed-lot input may have no effect in an automatic sizing mode.
5. Scope and timing: does it affect new entries, existing positions, this symbol or the whole account? Is it checked on a tick, a new bar or only at initialization?
6. Evidence: what observable change would show that it is working? Write that down before testing.

Common names that need more explanation

The following are hypothetical examples, not promises about a particular product.

• Risk: it might be a balance-based lot multiplier or a stop-distance-based sizing rule. A value of 1 does not, by itself, establish a maximum loss of 1%.
• StopLoss / Step: 100 points, 100 pips and a price distance of 100 are different. Check the symbol's point size and the EA's conversion; do not reuse a forex assumption for gold or BTC.
• MaxSpread: a filter might block new entries while leaving existing-position management active. “No new trade” does not mean “the EA has stopped”.
• StartHour / EndHour: identify the clock used and how overnight intervals work. Broker server time, UTC and your computer's local time are not interchangeable.
• Trailing / Mode: a distance may apply only after a profit threshold and an enable switch. Changing an inactive distance will not demonstrate a failure.

Write one sentence that you can test

Use this worksheet: version → input name → unit → active mode → affected orders → expected observation → rollback value. Add “unknown” where the documentation is unclear; that is a question to resolve, not a setting to guess.

For example: “In this version, when the spread exceeds the documented threshold in points, no new entry should be placed; existing-position exits should still follow their own rules.” Use that sentence only if the product's documentation actually describes this behaviour. Record what the log and order history show at the boundary.

A practical learning and testing loop

1. Keep the delivered files unchanged. Save a baseline SET and record the EA version, symbol, timeframe and relevant account conditions.
2. Choose one question and change one setting, or the smallest group of settings that must work together. Record the old and new values.
3. Load the SET into the intended version and read the active values back. In a single Strategy Tester run, check Value; Start, Step and Stop describe an optimization range, not ordinary run-time switching.
4. Use the same history, costs and test conditions for the comparison. Inspect entries, lot sizes, exits and log messages that relate to your question—not only final profit.
5. If the intended condition never occurred, the test is inconclusive. Choose a period or controlled test that exercises it; do not label an untouched switch “verified”.
6. Save the observation with the SET. Try the understood configuration on a demo account and keep a clear path back to the baseline.

The MT5 testing guide distinguishes a single run from optimization. Continue with saving and checking SET files and a reproducible MT5 backtest.

When the meaning is still unclear

Send support the product version, exact input label, current value, relevant mode, expected behaviour and a short log excerpt. Do not send passwords. With source access, trace where the input is read and which branches use it. An AI explanation can help form questions, but it is not evidence that your running version behaves that way.

Do not load another person's “best settings” without understanding the units, dependencies and environment. A preset is a saved configuration, not a performance guarantee. Clear documentation makes a setting understandable; testing establishes what happened in the conditions you actually exercised.

References and next steps

• MT5 program properties and input settings
• Install an EA in MT5

Back to top