Fusion EA Master: An Advanced Guide to Its Gold Entry Filters and Risk Engine

Fusion EA Master opens a trade on Gold only when a trend filter, a momentum filter and a minimum signal score all agree. This guide explains each layer of that decision, the risk engine around it, and how to test and monitor the EA before you trust it with real funds.

It is written for traders who already know MetaTrader 5 and the Strategy Tester, and who want to understand the logic rather than run a black box. Every feature described here is part of version 4.2. Where a setting depends on your broker, the text says so.

The article does not show profit figures. Backtests depend on data, spread and settings, and no test predicts future results. The goal is to give you enough understanding to run your own tests and draw your own conclusions.

Why Gold calls for a filter-first approach

XAUUSD has four traits that punish a system which trades every signal.

• Wide intraday ranges. Gold can move several hundred points in a session. A stop that is safe on a quiet day can be hit by normal noise on a busy one.
• Session behaviour. Activity usually rises during the London and New York sessions and falls during the late Asian hours. Signals in thin markets tend to be less reliable.
• Variable spread. Spreads on Gold widen around news, rollover and low-liquidity periods. A setup that looks valid on the chart can be expensive to enter.
• A link to the US dollar. Gold is quoted in dollars, so changes in dollar strength often show up in its price.

A filter-first design answers these traits by saying no more often than yes. Each layer removes a class of weak setups. What remains is a smaller number of entries where trend, momentum and price location point the same way.

How a trade decision is made

A setup must pass every check below before an order is sent. If any one of them fails, no order is opened.

1. Trend filter: is price on the trade side of the H1 EMA50?
2. Momentum: does the M15 RSI confirm the direction?
3. Optional filters, if enabled: ADX, DXY correlation, round-number levels, tick imbalance.
4. Signal score: is the score at or above the minimum?
5. Execution checks: spread, broker stop level, margin, no open position on the symbol and magic number.
6. Daily limits: are the daily loss and the daily trade count still within limits?

If all six pass, the order is opened with its stop loss and take profit, and a lot size calculated from your risk percentage. The first four checks decide whether the setup is good enough. The last two decide whether the account can take it now. The sections below follow the same order.

Layer 1: the H1 EMA50 trend filter

The first layer decides which direction is allowed. By default, the EA reads a 50-period exponential moving average on the H1 chart. Buy setups are considered only when price is above the EMA, and sell setups only when it is below.

An EMA gives more weight to recent prices than a simple average, so it turns sooner after a change in direction. On H1, a 50-period EMA covers roughly the last two trading days. That is long enough to ignore most intraday noise and short enough to follow a new swing.

The filter works on a higher timeframe than the entry on purpose. A trader who enters on M15 against the H1 trend is often trading a pullback that becomes a reversal. Requiring agreement with H1 removes those counter-trend entries.

The trend timeframe can be changed in the inputs. A higher timeframe makes the filter slower and more selective. A lower one makes it react faster but lets more noise through.

Layer 2: M15 RSI momentum confirmation

The second layer checks that momentum supports the direction chosen by the trend. By default, the EA reads the Relative Strength Index on the M15 chart before a setup is accepted.

The RSI compares the size of recent up moves with recent down moves on a scale from 0 to 100. Readings above the middle of the scale show that buyers have had the upper hand recently. Readings below it show the opposite.

Together, the two layers form a classic multi-timeframe structure. H1 answers "which way is the market going?" and M15 answers "is it moving that way now?". An uptrend on H1 with fading momentum on M15 is skipped, because the move may be pausing or turning.

The momentum timeframe can also be changed in the inputs. Keep it lower than the trend timeframe, or the two layers will measure the same thing.

Layer 3: the signal score

A setup that passes the trend and momentum checks is not traded automatically. It first receives a score, and only setups at or above the minimum score set in the inputs are traded.

The score combines three measurements:

1. Distance from the EMA. It tells the EA how stretched price is relative to the trend line. An entry far from the average carries more risk of a pullback against it.
2. The RSI reading. It indicates how strong the current momentum is, not only its direction.
3. Nearby support or resistance. It shows whether price is entering into open space or straight into a level that may stop it.

The minimum score is the main dial for the EA's character. A high minimum produces fewer trades, each with stronger agreement between the measurements. A low minimum produces more trades and accepts weaker setups. Neither is right in general: the best value depends on your symbol data, spread and how often you want the EA to act.

When you change the minimum score, change nothing else in the same test. That way any difference in results comes from the score alone.

Optional filters

Four additional filters can be switched on in the inputs. Each one adds a condition, so each one you enable reduces the number of trades.

Add optional filters one at a time in testing. Enabling all four at once can reduce the trade count so much that the test has too few trades to be meaningful.

The DXY filter requires the Dollar Index symbol to be available from your broker. Check its exact name in your Market Watch before you enable it.

Pre-trade execution checks

A valid signal still has to pass four execution checks before an order is sent.

1. Spread. The current spread is compared with the maximum you set. If Gold's spread has widened, the entry is skipped rather than paid for.
2. Broker stop level. The stop loss and take profit must respect the broker's minimum distance from price. An order that would be rejected is not sent.
3. Available margin. The EA checks that the account can carry the position before opening it.
4. One position per symbol and magic number. If a position is already open, no second one is added. There is no martingale, no grid and no averaging into a losing trade.

The fourth rule matters most for risk. Exposure on the symbol is never more than one position, so the planned risk of a trade is the risk of the account on that symbol.

Position sizing with AutoLot

With AutoLot enabled, the lot size is calculated so that a stop-loss hit costs a fixed share of the account. The calculation uses three inputs: your risk percentage, the stop-loss distance and the symbol's tick value.

Lots = (Balance x Risk %) / (SL distance in ticks x Tick value per lot)

A worked example, with round numbers chosen for clarity rather than taken from any broker:

• Balance: 10,000 USD, risk per trade: 1 %, so the amount at risk is 100 USD.
• Stop-loss distance: 500 ticks.
• Tick value: 0.10 USD per tick for 0.01 lot, so 10 USD per tick for 1 lot.
• Lots = 100 / (500 x 10) = 0.02.

Any volume sent to the broker must also respect its lot step and its minimum and maximum volume, so check these values in the symbol specification.

Two practical points follow. First, with ATR-based stops, the stop distance changes with volatility, so the lot size shrinks when Gold is volatile and grows when it is calm. The money at risk stays the same. Second, tick value and contract size differ between brokers, which is why you should always check the first trades on a demo account.

Exits: stop loss, take profit, breakeven and trailing

Every trade has a stop loss and a take profit from the moment it opens. Both can be set in one of two ways.

• ATR-based. The distances are multiples of the Average True Range, so they adapt to current volatility. On a busy Gold day the stop sits further away, on a quiet day closer.
• Fixed in points. The distances stay the same in all conditions. This is easier to reason about but does not adapt to volatility.

Once a position is open, two optional tools manage it.

• Breakeven moves the stop loss to the entry price after price has moved a set distance in your favour. It reduces the risk on that trade but does not remove it: a gap or slippage can still close the trade at a loss.
• Trailing stop follows price at a set distance as the trade moves in your favour, locking in part of an open gain.

A tight trailing stop on Gold is often closed by normal intraday swings. Test the trailing distance against the typical size of a pullback on your chosen timeframe.

Daily risk governance

Three rules work above the level of a single trade and limit what a bad day can cost.

The lot reduction is the opposite of a martingale. A martingale increases size after losses to recover them. Fusion EA Master reduces size, so a losing streak costs less as it continues.

Set the daily loss limit from your own tolerance, not from a backtest. A useful check: multiply your risk per trade by the maximum trades per day. If that worst case is more than you would accept in one day, lower one of the two.

A testing method you can trust

The free demo runs in the MetaTrader 5 Strategy Tester. How you run it decides whether the result means anything.

Use real ticks and your broker's data. Choose "Every tick based on real ticks" mode on XAUUSD from the broker you plan to trade with. Gold's spread and tick flow differ between brokers, and the generated-tick modes cannot reproduce spread spikes.

Split the history in two. Pick settings on one period (in-sample), then run them unchanged on a later period you did not look at (out-of-sample). Settings that only work on the first period were fitted to its noise.

Judge stability, not the best number. When you vary one input, look at a range of nearby values. If results collapse when the minimum score moves by one step, the setting is fragile. A plateau of similar results is a better sign than one sharp peak.

Look past net profit. For each test, record:

• the number of trades, since fewer than about 100 says little;
• the maximum drawdown, in money and percent;
• the profit factor, gross profit divided by gross loss;
• the longest losing streak, which tells you what you must be ready to sit through.

Forward-test on a demo account. Run the chosen settings on a demo account for several weeks before any live use. The forward test checks what no backtest can: live spreads, slippage and execution at your broker.

Keep a short log of every test: date range, mode, set file and the four figures above. It turns testing into evidence you can compare.

Configuration principles

Two set files are attached in the product comments: the default H1/M15 configuration and an alternative for the M1 timeframe. A set file only changes input values. The trading logic and risk controls stay the same.

Beyond the set files, most configurations fall between two profiles:

Start selective, then loosen one setting at a time. Lower the risk per trade before you raise the number of trades: a new configuration should first prove itself at small size.

Adjust three values to your broker before anything else: the maximum spread, the SL/TP distances and the lot or risk settings. Point values for Gold differ between brokers, so a distance in points that suits one broker may be too tight at another.

Monitoring: dashboard and CSV journal

The on-chart dashboard shows the account's state at a glance: balance, equity, daily profit and loss, risk status, trade count, the current signal, open positions and a cumulative realized profit chart. Check the risk status line first each day: it tells you whether the daily limits still allow new entries.

Every closed deal is also written to a CSV trade journal, together with a summary file. Open it in a spreadsheet each week and compute a few figures yourself:

• Win rate and average win versus average loss. Together they give the expectancy per trade.
• Results by hour or session. If losses cluster in one session, consider restricting trading hours in testing.
• Live versus backtest. Compare the live trade count and average result with the backtest over the same weeks. A large gap usually points to spread, slippage or a broker difference.

Review the journal on a fixed day and change settings only after enough trades. Reacting to every losing trade turns a tested system into an untested one.

Limitations and risk

Fusion EA Master is a rules-based tool, not a forecast. Its limits should be clear before you use it.

• Filters reduce bad trades, they do not remove losses. Losing trades and losing streaks are a normal part of any trend-following system.
• Trend filters lag. At a sharp reversal, the EMA turns after the price does. The first trades after a change in direction can be stopped out.
• Ranges are difficult. In a sideways market, trend signals appear and fail often. The ADX filter helps but does not solve it.
• Actual losses can exceed planned risk. Slippage, spread widening and price gaps, for example at the weekly open or around news, can close a trade beyond its stop.
• Results depend on the broker. Spread, commission, tick value and execution all change the outcome.

Trading leveraged instruments such as Gold involves substantial risk of loss. Backtests and past results do not guarantee future performance. Test on a demo account first, and only trade with capital you can afford to lose.

Conclusion

Fusion EA Master trades Gold only when trend, momentum and price location agree strongly enough to pass a minimum score. Around that decision sit execution checks, risk-based sizing and daily limits that cap what one trade and one day can cost.

The most useful next step is your own test: real ticks, your broker's data, an out-of-sample period and a few weeks on a demo account. If you have questions about the logic or the settings, ask in the product comments or send me a private message on MQL5.

Fusion EA Master on the MQL5 Market, with the free demo and the set files: https://www.mql5.com/en/market/product/145303