Gold testing

    GOLD STRATEGY CHECKLIST

    A rule set that has never been written down cannot be tested, and a rule set that has not been tested to a real sample size cannot be trusted with money. This page is the checklist that sits between "I think this works on gold" and the first live trade, organised so each item can be marked pass or fail rather than judged by feel.

    Before trading a gold strategy live, verify eight things: the rules are written and unambiguous, the sample size clears 100 trades across at least one full macro cycle, cost assumptions are checked against your own live feed, risk limits and session restrictions are explicit, a news blackout policy exists, and the journal is already running. A go/no-go table converts these into a single decision.

    WHY A CHECKLIST INSTEAD OF A FEELING

    Traders rarely go live on a bad strategy because they misjudged the edge. They go live on a good strategy that was never fully specified, tested against a realistic sample, or risk-managed for gold's actual volatility, and the losses that follow get blamed on the market rather than the process that skipped a step.

    A checklist works because it forces a pass/fail answer to questions that are otherwise easy to wave away with "it's close enough" — a sample size of 60 instead of 100, a cost assumption borrowed from an article instead of measured on your own account, a risk limit that exists in your head but not on paper. Each of those gaps is small individually and compounds badly in combination.

    This is not a one-time gate. Re-run the checklist whenever the strategy changes materially, whenever you change broker or account type, and at minimum once a quarter even for a strategy already live, because cost structures and your own execution discipline both drift over time.

    The rule this checklist enforces

    If any single item below fails, the strategy is not ready for live capital, regardless of how good the backtest or demo equity curve looks. A strategy with a genuine edge survives a delay to fix a gap; a strategy pushed live with a gap rarely survives the gap catching up with it.

    1. WRITTEN RULES

    The rule set must exist as a document, not as a feeling the trader can recognise on a chart. Every entry condition, every stop and target rule, every position sizing decision and every exception must be written in language specific enough that a different competent trader could execute it identically from the document alone.

    The most common failure here is a rule that reads correctly in English but is unmeasurable in practice — "enter after a strong rejection" is not a rule, because "strong" has no defined threshold. "Enter after a wick that exceeds 1.5x the average of the prior five candles' range, closing back inside the range within two candles" is a rule, because it can be verified from a chart by two different people and they will agree on the answer.

    Written rules should also state explicitly what is not traded: which sessions are excluded, which news windows are excluded, and what the maximum number of trades per day or per week is. A rule set that only specifies when to enter and never specifies when not to trade is incomplete.

    WRITTEN RULES CHECKLIST

    1. 01

      Every entry condition is measurable, not descriptive

      Pass test: Could two traders independently mark the same chart and agree on every entry?

    2. 02

      Stop and target rules are stated as dollars per ounce or a defined multiple, not "feel"

      Pass test: Is the stop distance derivable from the rules alone, with no chart-reading judgement required?

    3. 03

      Position sizing formula is explicit

      Pass test: Given a balance and a stop distance, does the document alone produce a lot size?

    4. 04

      Excluded sessions and news windows are named

      Pass test: Is there a written list of when this strategy is not traded?

    5. 05

      Maximum trades per day/week is stated

      Pass test: Is there a hard cap that prevents overtrading on a choppy day?

    2. SAMPLE SIZE

    A gold strategy needs a minimum of 100 trades before its statistics are worth acting on, and those 100 trades need to span at least one full macro cycle — a CPI release, an NFP release and an FOMC decision — because gold's behaviour around those events is different enough from its behaviour on an ordinary day that a sample missing them is not representative of what live trading will actually look like.

    Sample size requirements scale with trade frequency, not just with time elapsed. A scalping strategy generating five trades a day reaches 100 trades in a month; a swing strategy generating two trades a week needs closer to a year. Do not shortcut this by combining backtest and forward-test samples into one number unless the assumptions (spread, slippage, execution) were held identical across both, which they rarely are.

    Track the sample separately by regime if the strategy is directional in nature — a breakout or trend-following approach needs enough trades in both a trending period and a ranging period to know whether it is a genuine edge or a regime-dependent pattern being mistaken for one.

    Strategy typeApprox. trades/weekTime to reach 100 trades
    Intraday scalp15-254-7 weeks
    Intraday session setup (sweep, breakout)5-1010-20 weeks
    Swing / multi-day1-38-24 months
    News-momentum only2-4 (release days only)Several release cycles per year — often 12+ months

    SIZE THE SAMPLE HONESTLY

    3. COST ASSUMPTIONS VERIFIED ON YOUR OWN FEED

    Every backtest and every article, including this one, states cost assumptions as ranges — $0.15 to $0.35 spread in London and New York, $0.50 or more at rollover. Those ranges are not a substitute for measuring your own broker's actual spread, commission and swap on your own account type, because the difference between a raw ECN account and a standard commission-free account can change a marginal strategy's expectancy from positive to negative.

    The verification is mechanical: log the live spread at the moment of entry on at least 30 real or demo trades across different sessions, including at least a handful deliberately placed near the 21:00-23:00 UTC rollover window, and compare the average against the assumption used in the backtest. If the measured average is meaningfully wider than what was assumed, re-run the backtest or demo expectancy with the corrected number before going live.

    Commission and swap need the same treatment. Confirm the exact round-turn commission per lot from your account statement rather than from the broker's marketing page, and check whether swap on the side you trade most (long or short) is charged or credited, because on gold it is frequently a cost that quietly erodes a strategy that holds positions overnight.

    COST VERIFICATION CHECKLIST

    1. 01

      Measured average spread across 30+ trades, by session

      Pass test: Does your logged average match or beat the assumption used in testing?

    2. 02

      Confirmed commission per lot from an account statement

      Pass test: Is it the exact figure, not an estimate from a broker webpage?

    3. 03

      Checked swap direction and size on the dominant trade side

      Pass test: Have you priced a full week of overnight holds if the strategy holds overnight?

    4. 04

      Re-tested rollover-window spread specifically

      Pass test: Do you have logged spread data from inside 21:00-23:00 UTC, not just an assumption?

    MEASURE YOUR OWN COSTS

    4. RISK LIMITS

    Risk limits need three layers, each written as a hard number rather than a guideline: risk per trade, maximum daily loss, and maximum drawdown before the strategy is paused for review. On gold, risk per trade of 0.25% to 1% of account balance is the institutional range, with the lower end appropriate for strategies with wide stops or lower win rates.

    The daily loss limit exists to stop a bad session from compounding into a bad week — a common structure is two to three times the per-trade risk, so that a run of losses triggers a mandatory stop for the day well before the account has taken meaningful damage. The drawdown pause threshold is larger still, typically expressed in R (see the metrics page) rather than currency, and triggers a full review rather than a same-day stop.

    All three limits must be enforced mechanically — a broker-side daily loss limit, a trading platform lockout, or at minimum a written commitment checked before every session — because a limit that only exists as a mental note is the first thing abandoned during a losing streak, which is exactly when it is needed most.

    LimitTypical valueEnforcement
    Risk per trade0.25% - 1% of balancePosition size calculator, checked before every entry
    Daily loss limit2-3x per-trade riskPlatform lockout or a hard personal rule checked at each new trade
    Drawdown pause threshold8-12R cumulative, or per prop firm ruleFull strategy review before resuming; not a same-day decision

    SET LIMITS BEFORE YOU NEED THEM

    5. SESSION RESTRICTIONS

    A strategy tested across all 24 hours of the trading day but only found to be profitable in a subset of them must be restricted to that subset in its written rules before going live. Trading outside a proven window "because the setup looks the same" reintroduces exactly the noise the testing process was meant to remove.

    The restriction should name specific UTC windows, not vague labels — "London morning" is not as useful as "07:00-11:00 UTC" because session character can shift meaningfully within a single loosely-labelled block. If the strategy's edge was found specifically in the New York overlap (13:00-16:00 UTC), the rule set should say so and exclude the rest of the New York afternoon even though it is nominally the same session.

    Rollover (21:00-23:00 UTC) should be excluded by default for any strategy not specifically designed and tested for that window, given the spread widening and manufactured wicks discussed elsewhere in this cluster.

    6. NEWS BLACKOUT POLICY

    A written news blackout policy names the specific releases to avoid and the exact window around each one, rather than a general instruction to "be careful around news". US CPI and NFP are released at 13:30 UTC and typically warrant a blackout from around 13:00 to 14:30 UTC given the range expansion and spread widening that follows. FOMC decisions at 19:00 UTC warrant a wider window, often 18:45 to 20:30 UTC, because the initial reaction frequently reverses on the press conference.

    The policy also needs to define what counts as a blackout violation for an existing open position, not just for new entries — a position opened before the blackout window that would still be open when it starts should either be closed or have its stop and size explicitly reviewed for the wider spread and volatility that is about to arrive.

    Unscheduled events — a surprise geopolitical headline, an off-cycle Fed speaker — cannot be scheduled in advance, so the policy needs a rule for them too: typically, no new entries for a defined period (for example two hours) after any move exceeding a stated dollar threshold in under five minutes.

    KNOW WHEN THE NEXT ONE LANDS

    • GOLD NEWS COUNTDOWN

      Live countdown to the next CPI, NFP and FOMC, with what gold typically does around each release.

    • GOLD MACRO DRIVERS

      Real yields, the dollar, CPI, NFP and FOMC — the context that decides whether a level breaks or holds.

    7. JOURNAL READINESS

    The journal must already be running before the first live trade, not set up retroactively after a few weeks of live trading have already gone unrecorded. A strategy that goes live without a working journal has no mechanism to detect decay, no evidence to distinguish a valid loss from a rule violation, and no data to support the decision to keep going or to stop.

    Readiness means the minimum field set is defined and tested on at least the last stretch of demo trades — UTC session, setup tag, spread at entry, intended versus executed stop, macro tag, exit reason and a violation flag — and that the review cadence (weekly, monthly, quarterly) is scheduled, not aspirational.

    Confirm platform sync is actually working before relying on it. A MetaTrader 5 connection that silently stopped syncing three weeks ago produces a journal with a gap the trader will not notice until the monthly review, at which point the missing data cannot be recovered.

    SET UP THE JOURNAL BEFORE GOING LIVE

    • FREE GOLD TRADE JOURNAL

      Sync MetaTrader 5 and let the journal attribute each gold trade to its UTC session, setup type and R multiple automatically.

    8. THE GO/NO-GO DECISION TABLE

    This table converts every section above into a single pass/fail check. A strategy needs a "go" on every row before the first live trade — this is deliberately not a majority vote, because any single failed row (a missing risk limit, an unverified cost assumption) is capable of turning a genuine edge into a losing account on its own.

    Checklist itemGo criterionNo-go signal
    Written rulesEvery entry, exit and sizing rule is measurable and documentedAny rule requires "feel" or judgement not defined in writing
    Sample size100+ trades spanning a CPI, an NFP and an FOMC cycleUnder 100 trades, or the sample misses a macro cycle entirely
    Cost assumptionsSpread, commission and swap measured on your own live feedAssumptions borrowed from an article or a different account type
    Risk limitsPer-trade, daily and drawdown limits written and enforced mechanicallyLimits exist only as a mental commitment
    Session restrictionsTrading windows named in UTC, matched to where the edge was provenStrategy trades all hours despite edge concentrated in one window
    News blackoutCPI/NFP/FOMC windows and an unscheduled-headline rule are writtenNo defined policy, or policy covers only scheduled events
    Journal readinessMinimum fields defined, sync verified, review cadence scheduledJournal to be set up "once live trading starts"

    If every row reads "go", the strategy is ready for a small live allocation — not full size, because the purpose of the first funded month is to test the trader's behaviour under real risk, which no amount of demo testing can substitute for. If any row reads "no-go", fix that specific gap and re-run only that section of the checklist before proceeding; there is no need to repeat sections that already passed.

    RISK NOTE

    Leveraged gold carries a high risk of rapid loss. Everything on this page is a testing method, not financial advice, a signal, or a projection of returns.

    GOLD TRADING FAQ

    What is the single most commonly skipped item on this checklist?
    Cost verification on the trader's own live feed. Traders routinely reuse generic spread assumptions from an article instead of measuring their own broker's actual spread, commission and swap, which can turn a marginally profitable backtest into a losing live strategy.
    Can I go live if my sample size is under 100 trades but the results look strong?
    Strong early results on a small sample are more likely to be variance than edge, and an early sample that also misses a CPI, NFP or FOMC cycle is not representative of what live trading will look like. Wait for the full sample before committing meaningful size.
    Do I need a written news blackout policy if I already avoid news intuitively?
    Yes. An intuitive habit erodes under pressure or after a string of wins, while a written policy with specific UTC windows can be checked mechanically before every trade regardless of mood or recent results.
    What counts as a "go" on session restrictions if my strategy performed well across all hours in testing?
    If the sample genuinely showed even performance across all UTC hours with sufficient trades per hour, that itself is a valid go — the requirement is evidence-based restriction, not restriction for its own sake.
    Should the journal be set up before or after starting demo testing?
    Before. Journal readiness should be verified using real demo trades, not retrofitted once live trading begins, so that the field set, sync and review cadence are already proven working by the time real money is at risk.
    What size should the first live allocation be if every checklist item passes?
    Small relative to eventual target size, and typically for a defined trial period of one month or a set number of trades, because the first live allocation exists to test the trader's behaviour under real risk, not to re-test the strategy itself.

    JOURNAL YOUR GOLD TEST FREE

    Live XAUUSD signals on a public ledger, a free gold trade journal that syncs with MetaTrader 5, and the macro pulse before the London open.

    EXPLORE GOLDHUNTS

    • FREE GOLD SIGNALS

      Live XAUUSD calls with entry, invalidation and target — tracked to outcome.

    • FREE GOLD TRADE JOURNAL

      Connect MetaTrader 5 and every XAU trade is recorded and analysed.

    • GOLD MACRO PULSE

      Real yields, the dollar, central bank demand and the data that reprices gold.

    • LEARN GOLD TRADING

      The fundamentals pillar: sessions, volatility, lot sizes, spreads and costs.

    • XAU PAIRS

      XAUUSD, XAUAUD, XAUEUR, XAUGBP, XAUJPY, plus futures and physical gold.

    • GOLD STRATEGIES

      Session breakouts, liquidity sweeps, mean reversion and news momentum on gold.

    • TESTING GOLD STRATEGIES

      Backtesting, demo-to-live, sample size and why gold edges decay.

    • GOLD INTELLIGENCE

      How gold signals are generated, validated and kept under human oversight.

    • GOLD CALCULATORS

      Pip value, lot size from your stop, margin, and a live UTC session clock.

    • GOLD DATA API

      Free JSON endpoints, OpenAPI and MCP tools so agents can call the gold desk.

    • GOLD BROKERS & PLATFORMS

      XAU symbol names on MT5, gold on Trading 212, and spread comparisons.

    THE FULL GOLD TESTING CLUSTER