This is the dedicated demonstration-account evidence page for GHX-09 Higher Timeframe Bias v2.0. The record begins with the first eligible forward-test trade and will include each trade after it closes, whether it wins or loses, together with the rationale recorded before its outcome was known. Coming to forward testing. It is not yet available for account activation.
The approach behind the record: Direction comes from confirmed daily price structure alone — the sequence of turning points the daily chart has actually completed, each counted only once the candles that confirm it have closed. The 4-hour chart is then asked to agree. It can confirm or block the daily reading, but it can never supply a direction of its own, and an unclear daily picture is simply not traded rather than resolved by a faster chart. Anything neutral, conflicting, out of date or already given up is left alone. If the level the daily or 4-hour reading depends on has been traded through, permission is treated as lost and stays lost until the chart confirms a new one — a recovery in price does not restore it. Inside that permission it watches completed hourly candles for a measured setback after a run in the permitted direction, and judges only the first attempt to resume. Too shallow, too deep, too slow, too late, or lacking conviction in its own close, and the sequence is finished rather than waiting for a better-looking candle. The entry timeframe is the hourly candle and nothing faster. There is no 15-minute or 5-minute fine tuning, no moving average, no volume gate and no retracement drawing. Entry is valid only for a short window after that candle closes, at a fresh executable price, and only while the spread and the distance moved since the close are both inside what it accepts. Every trade carries a stop just beyond the setback it resumed from, and a target taken from the nearest daily or 4-hour level ahead that price has not already reached since that level was confirmed. If that level leaves too little room after costs the trade is skipped — it never reaches past it for a better-looking ratio and never invents a level of its own. One position at a time, and never any addition to it: no averaging down, no adding to a winner, no part exit, no breakeven jump and no reversal. Once the trade is a full unit of risk ahead on verified prices, protection follows newly confirmed hourly turns behind it and is only ever tightened. It closes if the daily or 4-hour reading stops supporting it, before a major United States announcement, ahead of the next verified trading break, and in any case within a fixed number of hours of the fill. This is an intraday expression of a daily reading rather than an overnight swing system. Agreement between two views of the same prices is not independent evidence, and a lower trade frequency is an intended distinction rather than a measured result.
What each published trade shows: Date and time the trade opened and closed, in UK time. Direction, entry price, exit price, and the distance to the stop and the target. Result in R — the outcome measured against the risk taken, net of spread, commission and swap, with the gross figure beside it. Win or loss, how long it was held, the session it traded, and the model version that took it. The plain-English reason the model took the trade, written before the outcome was known. Never published: lot size, account balance, ticket numbers, and anything at all about a customer account.
How the record can be checked: published entries remain visible, and corrections are added as new entries linked to the original rather than silently changing history. Each entry will carry a check value linked to the previous entry, and signed summaries will preserve the record on each publication day. The same decision is released to the demonstration account and to every trader running the model in the same cycle, with no preferential ordering.
Scope and limits: these are demonstration-account results and demonstration results do not represent live trading. Nothing here is investment advice or a forecast, and no return is promised — leveraged gold carries a high risk of rapid loss. Sample size is always shown beside any win rate, and the record is published from the first trade rather than once it looks good. Results are measured against the risk taken, net of spread, commission and swap, with the gross figure alongside; cash profit and loss is not published. Gold Hunts operates from the United Kingdom under the laws of England and Wales. A model acts only while a trader leaves it switched on, on an account they connected, inside limits they set.
How Higher Timeframe Bias decides is explained in full at https://goldhunts.com/gold-models/ghx-09-htf-bias.
GoldHunts — the AI-powered gold trading operating system.