XAUUSD · METATRADER 5 · TESTED ON UNSEEN DATA

A gold trading bot that benches its own losing strategies

Most XAUUSD bots run one strategy and hope the market keeps cooperating. Tech Kick Bot was built by testing 34 of them — six built only for gold — scoring each on the lower bound of its measured win rate, and shipping only the four that earned it. Just the current leader touches your account, and when a strategy stops working it gets suspended automatically instead of quietly draining the balance.

30-day money-back guarantee · Hardware-locked licence · Cancel anytime

What it is, precisely

Tech Kick Bot technical specifications
InstrumentXAUUSD (Gold)Single-symbol by design
PlatformMetaTrader 5
Strategies tested34Only the proven ones ship enabled
Timeframes scannedM1 · M5 · M15
Trading hours06:00 – 21:00 UTC
Market regimesTrend · MixedChop ships disabled — it lost money
Default risk per trade1.5% of balanceUser-configurable
Default reward:risk1 : 2
News shieldHigh-impact calendarOn by default
Spread guardBlocks entries above threshold
Daily lockdownHard profit + loss stop
Minimum balance$100$700+ for % risk sizing to work; $1,000 ideal
Operating systemWindows 10 / 11 / VPS

How the bot decides what to trade

1. It reads the regime first

Before any signal is considered, the engine classifies the current market as trending, chopping, or mixed. This matters more on gold than on most instruments: a breakout strategy that prints money in a trend is exactly the strategy that bleeds in a range. Strategies are gated by regime, so mean-reversion logic simply is not eligible during a strong trend.

2. Every eligible strategy submits a signal

Across M1, M5 and M15, each enabled strategy independently reports whether it sees a setup. Smart-money logic (order blocks, liquidity sweeps, structure shifts) runs alongside classical logic (VWAP mean reversion, EMA pullbacks, Bollinger reversals), with anything that could not prove an edge switched off.

3. Signals are scored on confidence, not luck

Each signal is ranked using the lower bound of its strategy's win-rate confidence interval. A strategy that is 3-for-3 scores below one that is 60-for-100, because three samples prove nothing. This is the single most important design decision in the engine — it is what stops the bot chasing whichever strategy got lucky this morning.

4. Only the champion executes

Duplicate signal types are collapsed, the highest scorer wins, and position size is calculated as a percentage of your balance with stop loss and take profit attached immediately after fill. Everything else is discarded — the bot does not stack correlated entries just because several strategies agreed.

The parts that stop you blowing the account

Entry logic is the part everyone markets. These are the parts that actually decide whether an account survives twelve months.

Daily lockdown

Hard profit and loss thresholds. Once either is hit the bot stops for the rest of the server day — the mechanism that makes revenge-trading structurally impossible.

News shield

Pulls the high-impact economic calendar and refuses entries around CPI, NFP and FOMC, when gold gaps and spreads blow out.

Spread guard

Rejects any entry while the spread sits above a configured ceiling, so the trade does not start in a hole it cannot climb out of.

Loss-streak pause

After consecutive losses the engine pauses globally rather than continuing to fire into conditions it is clearly misreading.

Auto-suspension

A strategy that falls under the win-rate floor after a minimum sample is benched, then re-evaluated later. Underperformers are removed from rotation automatically.

Two-tier trailing stop

Stop moves to breakeven partway to target, then locks in a portion of open profit as the trade matures.

Who can actually run this

If you are trading from the United States

US residents cannot trade spot gold CFDs through a standard retail account, and the broker this bot is tuned for does not accept US clients. We would rather tell you that here than take a subscription you cannot use. The route that does work is a proprietary trading firm account — you trade the firm's simulated capital under their rulebook, on MetaTrader 5, with gold available. Check two things before you start: that the firm permits Expert Advisors, and that you have set the daily loss lock below their drawdown limit. We cover the setup in detail on the prop firm trading bot page.

Everywhere else

Any MetaTrader 5 broker offering XAUUSD works. The bot is pre-configured for Exness ECN and Raw Spread accounts, including their symbol naming, so that path needs no setup changes. Raw-spread or ECN account types are strongly recommended — the strategy mix is spread-sensitive and a wide standard-account spread will erode it.

Gold trading bot FAQ

What is a gold trading bot?

A gold trading bot — also called an auto trading bot or Expert Advisor — is software that opens, manages and closes XAUUSD trades automatically on your MetaTrader account using pre-coded rules, with no manual input once it is running. Tech Kick Bot is a gold trading bot for MetaTrader 5 built by testing 34 separate strategies, disabling the ones that could not prove an edge, and routing live execution only to whichever survivor currently has the strongest measured win-rate confidence.

Why does the bot only trade gold instead of multiple pairs?

Because narrow beats broad. Gold has a distinct volatility profile — it routinely moves several times the daily range of a major FX pair — and strategies tuned for that behaviour perform poorly elsewhere. We previously ran Bitcoin alongside gold and disabled it after it produced a net loss across 117 trades at a 28% win rate. Rather than hide that, we removed the symbol. The same principle governs the strategy list: of 34 strategies built and tested, only the ones that proved an edge on gold are enabled for live trading.

How does the strategy tournament actually choose a trade?

On every scan the engine collects valid signals from all eligible strategies. Each is scored using the lower bound of its measured win-rate confidence interval, not its raw win rate — so a strategy that went 3-for-3 does not outrank one that went 60-for-100. Base priority and timeframe weighting are added, signals are deduplicated by type, and only the highest scoring signal is executed. A strategy that drifts below the win-rate floor after a minimum sample is automatically suspended.

Can traders in the United States use this bot?

Not through a standard retail broker account. US residents cannot trade spot gold CFDs at retail, and the broker the bot is tuned for does not accept US clients. The route that does work for US-based traders is a proprietary trading firm account — you trade the firm's simulated capital under their rules, which is not restricted the same way. Confirm your firm permits Expert Advisors before running any bot, and check the daily loss lock is set inside their drawdown limit.

Does it work on prop firm challenge accounts?

It is compatible with MetaTrader 5 prop firm accounts, and the daily profit and loss lockdown maps directly onto the daily drawdown rule most firms enforce — once your configured loss threshold is hit, the bot stops trading for the rest of the server day rather than continuing into a breach. You must still set the threshold below your firm's limit yourself, and confirm the firm allows EAs. No bot can guarantee passing a challenge.

What happens during high-impact news?

The news shield pulls the high-impact economic calendar and refuses to open positions inside a window around each event. Gold reacts violently to CPI, NFP and FOMC releases, and spreads widen sharply — entries taken into those conditions are where automated systems most often take outsized losses. The shield is enabled by default.

Do I need to leave my computer on?

The bot needs MetaTrader 5 running to trade, so either your PC stays on or you run it on a Windows VPS. A VPS is the normal setup — it keeps the bot online through reboots and power cuts and sits closer to the broker's servers, which reduces execution latency.

How much money do I need to run it?

It runs from $100, and we would rather explain what changes at that size than pretend nothing does. The broker's smallest trade is 0.01 lots, which costs roughly $9 to $12.50 of risk on gold. On a $100 account that works out to about 10% risk per trade rather than the 1.5% you configured, because the trade cannot be made any smaller — position sizing simply cannot respond to risk at that balance. The bot itself will tell you this at startup if you run it under $700. Simulated from $100 across 2026, the shipped configuration still drew down 36%. From around $700 percentage sizing starts working as designed, and $1,000 is the balance the backtested figures on this site were measured at. Starting small is fine as long as you know you are trading a much more aggressive risk profile than the settings suggest.

What did it do in backtesting?

There are two measurements and they answer different questions. Run as one unbroken six-month stretch on a $1,000 account with live percentage sizing — the closest thing to an installed bot — it returned +3.5% in total across 42 trades, with a 7.2% maximum drawdown, a 38% win rate, and four of seven months green. That is the figure to plan around. Separately, the rules were validated across seven independent monthly windows, each restarting the learning from a blank slate at a fixed trade size: January to July 2026 came out positive at a 21.2% drawdown and 44.8% win rate, and a replay on August to December 2025 — data the configuration had never seen, run as a stock customer install — was also positive at a 20.5% drawdown and 36.5% win rate, with two of five months losing money. The windowed method shows the rules have an edge across separate periods; it is not what a continuously running bot reproduces, and its percentages should not be projected forward. None of these scale down to a smaller account, because the minimum lot size forces a much higher risk per trade there. Backtested, hypothetical results — not a live track record.

What does it cost and is there a refund?

The bot is $20 per month with a 30-day money-back guarantee, and the licence is hardware-locked to one machine or VPS. Contact support to move a licence to a new device.

Run it on demo first

Put it on a demo account, let it trade a full week including a news day, and read the decision log — every entry it took and every one it refused. If it does not convince you, the 30-day guarantee is there for exactly that reason.

Every screen in the application

What you actually get when you install it. Ten screens, each explained — including the ones that exist to stop the bot trading rather than to make it trade.

Tech Kick Bot desktop dashboard showing the live log, open trades and the strategy performance tracker

Dashboard

The control surface. Start and stop the bot, see open positions, and read the live log as it works — every scan prints the regime it detected, the bias, the spread it measured and whether it acted. Most bots are a black box that either trades or does not; this one tells you which gate stopped it, which is the difference between "it is broken" and "there was no setup".

Bot configuration screen with MT5 credentials, risk per trade, reward-to-risk ratio and market regime selection

Settings and configuration

Where the account is connected and the risk is set. Risk per trade and reward-to-risk are the two numbers that decide survival, so they are sliders rather than buried config. The regime checkboxes at the bottom control which market conditions the bot is allowed to trade at all — chop ships switched off, because it did not earn money in testing.

Daily trading protection screen with configurable maximum daily profit and loss limits

Daily protection limits

A hard daily profit and loss lockdown. Once either limit is reached the bot stops trading until the broker's server midnight — it cannot be argued with in the moment, which is exactly when the urge to trade a loss back is strongest. This is also the control that maps onto a prop firm's daily drawdown rule; set it below the firm's limit, never equal to it.

Market regime screen showing seven market state dimensions and per-strategy fit scores

Market regime detection

A seven-dimension read of current conditions — direction, speed, volatility, momentum, liquidity, event risk and higher-timeframe alignment — resolved into a single bucket. It exists because no strategy works in all conditions: trend systems bleed in ranges and mean-reversion bleeds in trends. The fit score beneath decides which strategies are even eligible to compete for the next trade.

Strategy Lab table showing per-regime win rate, Wilson lower bound and flip state for each strategy

Strategy Lab

Per-strategy, per-regime performance with the ability to disable or flip any cell. Win rate alone is misleading on small samples, so the WLB column carries the Wilson lower bound — the pessimistic end of the confidence interval — which is what the engine actually ranks on. A strategy that won its only two trades does not get to outrank one with a long record.

Daily market news screen listing high-impact economic events with currency and impact rating

News shield

The high-impact economic calendar, filtered to the currencies that move your symbol. Gold reacts violently to US inflation, employment and rate decisions, and during those releases spreads widen and price moves discontinuously — a technical setup is not doing what it was measured doing. The bot pauses new entries around these events rather than trading through them.

Decisions screen showing analysis locked until five qualifying days of data are collected

Decisions and suggestions

Analysis that refuses to run until it has enough data. It requires five days of real uptime before it will suggest a single change, and it says so plainly rather than producing a confident recommendation from two days of noise. A tool that always has an opinion is a tool that will eventually give you a bad one.

Shadow trades screen comparing shadow-tracked signals against real executed trades

Shadow trades

Every strategy that is switched off still records what it would have done. That is how a strategy earns its way back to live status — on an accumulated real-conditions record rather than on an argument. It is also how the disabled ones stay disabled: the shadow log keeps proving they were right to be cut.

Pattern intelligence screen showing tracked setup fingerprints and a warm-up counter

Pattern intelligence

Fingerprints each setup by strategy, direction, regime and bias, then tracks how that exact combination has performed. Repeatedly good combinations get boosted and repeatedly bad ones get penalised. Note the warm-up counter — it will not act on a pattern until it has seen enough of them, for the same reason the Decisions tab stays locked.

Backtest screen with date range, balance, regime and per-strategy selection before running a replay

Built-in backtester

Replays real MT5 history through the same signal code that trades live — not a separate reimplementation, which is where most backtests quietly diverge from reality. Pick a date range, choose which strategies and regimes to include, and it applies the full gate chain exactly as the live engine does. No real money and no live positions are touched.

Account numbers, balances and profit figures are blurred in every screenshot. Account number plus server name is most of an MT5 login, and a screenshot of one account over a handful of trades is not evidence of anything — the validated backtest figures and the full strategy record are published separately, with their sample sizes attached.