A Complete Smart Money Concepts Routine for Gold: Top-Down Analysis to Entry Model
Lesson 9 of 10 · Advanced
Pre-session preparation
A good SMC routine starts before the kill zone opens, not while price is already moving. The purpose of preparation is that once the market is active, you are only comparing price behaviour with scenarios written in advance, not analysing from scratch. A common checklist:
- Check the economic calendar and the exact times of major US releases.
- Mark the previous day's high and low (PDH and PDL) and the previous week's high and low (PWH and PWL).
- Mark the valid, untouched FVGs and OBs on the Daily and H4.
- Define the current dealing range and whether price is in premium or discount.
- Identify the draw on liquidity and write the day's bias.
- Write the primary scenario, the alternative scenario and the invalidation condition for each.
This takes about twenty minutes, and its most important effect is fewer spur-of-the-moment, emotional decisions.
Top-down analysis
Top-down multi-timeframe analysis in SMC usually has three layers. The exact combination of timeframes is a matter of taste, and the table below is only a common example:
| Layer | Example timeframes | Main question |
| Direction | Daily and H4 | What is the bias, and toward which liquidity is price heading? |
| Zone | H1 and M15 | Which OB, FVG or OTE zone is suitable for a reaction? |
| Execution | M5 and M1 | Did a sweep and a structure shift with displacement occur in the zone? |
The key rule is that the lower timeframe is only for timing and has no right to invalidate the higher-timeframe bias unless a pre-written invalidation condition occurs. Many newer traders reverse their whole daily analysis on a single 5-minute CHoCH and, as a result, lose in both directions.
The entry model: an educational description
SMC entry models come in several versions, but most follow this sequence in some form. What follows describes a framework, not a signal:
- In the direction of the bias, price reaches a higher-timeframe zone (a point of interest) in the correct half of the range; for a bullish setup, that means in discount.
- This preferably happens in a kill zone.
- Before or at the moment of the reaction, a nearby liquidity pool is swept.
- On the execution timeframe, a CHoCH or MSS with displacement forms.
- That move leaves an FVG or a small OB, and an entry on the return of price to that zone is considered.
If even one of these conditions is missing, many traders stand aside. That strictness is what keeps the number of trades low and the quality of decisions high. Of course, no combination of conditions guarantees an outcome.
Stop, target and position-size logic
The stop-loss should sit where, if price reaches it, your idea was wrong; in the model above, usually beyond the extreme of the sweep. In gold you also need to allow extra distance for the spread and the normal volatility of that session, especially near news when spreads widen.
The target in SMC is usually the opposite-side liquidity: first internal liquidity such as a nearby FVG or internal high, then external liquidity such as the PDH. Some traders close part of the position at the first target. Reward-to-risk is calculated before entry, and if the logical target does not offer enough reward relative to the risk, there is no trade.
Position size is set by risk, not by a feeling of confidence. At most brokers a standard lot of gold is 100 ounces, so each one-dollar move is worth about 100 dollars per lot (check your own broker's contract specifications). Teaching example: with a 10,000-dollar account and a maximum risk of 1 percent per trade, which is 100 dollars, and a stop distance of 8 dollars, position size is about 100 divided by (8 x 100), or roughly 0.12 lots. The right risk percentage differs from person to person.
A complete hypothetical example
All numbers are for teaching. Before the session: the XAUUSD daily bias is bullish, the draw on liquidity is the previous day's high at 4,112, and an H1 FVG between 4,071 and 4,077 sits in the discount of a 4,060 to 4,110 dealing range. In the New York kill zone, after the 08:30 data, price falls to 4,069.5, sweeps the internal low at 4,072 and returns into the FVG. On M5 the internal high at 4,080 is broken by a strong candle, leaving a small FVG between 4,076 and 4,078.5.
In this framework, the entry area under consideration is 4,076 to 4,078.5, the invalidation point is below 4,069.5 plus the spread allowance (say about 4,068), the first target is the next internal high above and the final target is 4,112. The distance from a hypothetical entry at 4,077 to the stop is about 9 dollars and to the final target about 35 dollars. The example shows how the components fit together; in a real market price may never return to the small FVG or may hit the stop first.
Journal, backtest and stopping rules
Without a journal, SMC turns into a collection of selective memories. For every trade, and every setup you did not take, record: date and session, bias, higher-timeframe zone, type of sweep, confirmation timeframe, entry, stop, target, result, and a chart screenshot before and after the trade.
Before live use, backtest the model on a meaningful number of historical samples and then forward test for a period on a demo account or at very small size. Manual backtest results are usually better than live results because of hindsight bias, so expect live performance to be weaker. Win rate and returns differ from person to person and period to period.
Finally, write stopping rules: for example, stop trading after two consecutive losing trades in a day or on reaching a daily loss limit. These rules preserve the most capital on days when the bias is wrong.
Common mistakes
- Starting the analysis while price is already moving instead of preparing before the session
- Reversing the daily bias on every CHoCH on the 1- or 5-minute chart
- Placing the stop without allowing for the spread and gold's normal volatility in that session
- Sizing positions by a feeling of confidence instead of a fixed risk percentage
- Trusting manual backtest results without forward testing and without allowing for hindsight bias
What does volume say here?
Order flow can be added at three points in this routine. Before the session: on GC futures data, mark the prior day's POC and value area, untouched POCs, and the important HVNs and LVNs, and translate them to XAUUSD using the day's basis. An OB or FVG sitting on an LVN often behaves differently from a zone sitting on an HVN, and knowing that before the session is valuable.
When price reaches the zone: instead of waiting only for a CHoCH, check whether the sweep came with a volume spike and aggressive delta, and whether absorption and a delta turn follow. That confirmation often appears earlier than the structure change in the candles. After the trade: review the footprint of the trade to see whether participants were actually active in the entry zone or whether price crossed an empty area. Over time this review brings your SMC definitions into line with how the market really behaves.
📊 See real gold futures volume behind these ideas: free order flow course · today's gold volume map
Frequently asked questions
What are the steps of a daily SMC routine for gold?
Pre-session preparation covering news, the prior day and week levels, higher-timeframe zones and bias; then top-down analysis, waiting for the entry model conditions in a kill zone, and finally journaling and review.
Where is the stop-loss placed in Smart Money Concepts?
In the usual logic, beyond the point where price reaching it means the idea is invalid, often behind the extreme of the sweep, plus allowance for spread and volatility. This is an educational description and does not replace your personal risk management.
How many timeframes are needed for SMC analysis?
Typically three layers: a direction timeframe such as Daily and H4, a zone timeframe such as H1 and M15, and an execution timeframe such as M5 and M1. Choose the exact combination from your style and your own testing.
Education only, not a recommendation to buy or sell; we never give signals. Price examples are illustrative. Test any method on a demo account first, with strict risk management.