Multi-Timeframe Analysis: Using H4, H1 and M15 Top-Down Without Confusion

Multi-timeframe analysis means looking at the same market on several timeframes and giving each chart a different job. The most common approach is top-down: start on the higher timeframe and work down. The higher chart tells you which way the market is going; the lower chart tells you where to get in. The benefit is that you stop reacting to every minute-by-minute wiggle. The risk is that without rules, you keep opening charts until you find one that “agrees” with what you already wanted to do.

Top-down: three layers, three jobs

Layer	Example	Job	Question to answer
Higher	H4	Direction and structure	Up, down or ranging? Where are the major levels?
Middle	H1	Setup	Has price pulled back to the zone I planned?
Lower	M15	Trigger	Has the entry signal I wrote down in advance appeared?

Structure here means the sequence of highs and lows. An uptrend makes higher highs and higher lows; a downtrend makes lower ones. The zones used for setups are usually support and resistance or Demand/Supply zones. See Support, resistance and supply/demand.

The higher timeframe gives context, the lower timeframe gives timing. No layer tells you the future. Every layer is past price, just grouped differently.

One H1 candle is four M15 candles combined

Candles on every timeframe come from the same prices. A bigger candle does not contain more information; it just rolls several small candles into one. There are four rules:

• Open = the open of the first small candle
• High = the highest high of all the small candles
• Low = the lowest low of all the small candles
• Close = the close of the last small candle

Using the illustrative series in the figure below, hour 2 is made of M15 candles 5 to 8.

M15 candle	Open	High	Low	Close
5	2,406.1	2,406.5	2,402.9	2,404.4
6	2,404.4	2,405.0	2,401.2	2,401.7
7	2,401.7	2,404.1	2,401.3	2,402.6
8	2,402.6	2,403.4	2,399.5	2,399.8
H1 hour 2	2,406.1	2,406.5	2,399.5	2,399.8

• Open = open of candle 5 = 2,406.1
• High = highest of 2,406.5 / 2,405.0 / 2,404.1 / 2,403.4 = 2,406.5
• Low = lowest of 2,402.9 / 2,401.2 / 2,401.3 / 2,399.5 = 2,399.5
• Close = close of candle 8 = 2,399.8

Go one layer further and all 16 M15 candles become a single H4 candle: open 2,400.0, high 2,410.7, low 2,397.3, close 2,409.5. That is a green candle with a 9.5-dollar body and a lower wick of only 2.7 dollars.

In the figure, price drops from the 2,407.4 high (M15 candle 4) to 2,397.3 (candle 10), a fall of 10.1 dollars. Someone watching only M15 might feel the market has turned down. On H4 the same period is a single green candle. That is why you need to know what size of story each chart is telling.

Choose timeframes about 4–6× apart

A widely used guideline is to space the layers about 4 to 6 times apart. Too close, and two charts tell the same story. Too far, and they don’t connect. The idea is best known from Alexander Elder’s Triple Screen, which uses three charts roughly a factor of 5 apart.

Style	Higher	Middle	Lower	Spacing
Holding for weeks	W1	D1	H4	~5×, 6×
Holding for days	D1	H4	H1	6×, 4×
Intraday	H4	H1	M15	4×, 4×
Very short term	H1	M15	M5	4×, 3×

Counted in candles: one H4 = 4 H1 = 16 M15 = 240 M1. In a market open almost 24 hours, one D1 holds about 6 H4 candles.

The lower you go, the bigger the share of your Stop loss that the spread eats. Sets that end on M5 or M1 need extra care here. See Spread on small timeframes.

H4 candles differ between brokers

H4 and daily candles are cut by server time. If server time differs, candles start at different times, so the same prices produce differently shaped candles.

Server time	H4 candles start at (Thai time, UTC+7)
GMT+0	03:00 07:00 11:00 15:00 19:00 23:00
GMT+2	01:00 05:00 09:00 13:00 17:00 21:00
GMT+3	00:00 04:00 08:00 12:00 16:00 20:00

Some servers shift by an hour with daylight saving time. You can see your server time in Market Watch in your trading platform. Any rule based on the colour of an H4 candle therefore depends on which server you use.

Top-down, step by step

1. H4 for direction. Mark the last 2–3 important highs and lows. Rising means up, falling means down, alternating means unclear. Then mark the major levels near price.
2. H1 for the setup. Wait for price to pull back against the H4 direction into the zone you marked. Don’t chase once price has run away from the zone.
3. M15 for the trigger. Use one pre-written signal, for example an M15 close above the high of the previous 4 candles (for a buy).
4. Plan the exit before entering. Where is the Stop loss, where is the target, then work out the lot.

Real rules should be written so that someone else could read them and make the same call. “Important low”, for example, needs a precise definition, or every chart can be read to suit your mood.

When timeframes disagree

• H4 up, H1 down. The most normal case. This is a pullback, and it is what step 2 is waiting for. As long as H1 hasn’t broken an important H4 low, the bigger picture hasn’t changed.
• H4 up, but price is pressing a major H4 resistance. Direction is still up, but there is little room left. Targets beyond that level usually have to get through selling. Waiting for price to clear the level, or skipping the trade, are both reasonable choices.
• H4 ranging, H1 trending. The higher layer offers no direction. Some traders temporarily treat H1 as the higher layer, with targets no further than the edge of the H4 range.
• Every layer points a different way. No need to talk yourself into a choice. Not trading is a valid decision.

The key is that your conflict rules are written in advance, not decided while looking at price.

Which timeframe should set the Stop loss?

The Stop loss belongs where reaching it means your reason for entering was wrong. In top-down analysis the reason comes from the middle layer’s setup, so the Stop loss usually goes beyond the middle layer’s zone or low, not behind a single M15 candle.

Using the same illustrative series: buy at the close of M15 candle 13 = 2,404.2, because it closed above the highest high of candles 9–12 (2,402.3). The account is 10,000 dollars and you accept 1% risk = 100 dollars. For gold, 1 lot = 100 ounces, so a 1-dollar move = 100 dollars per lot (check the contract specification for your own account).

	Stop from M15	Stop from H1
Placed below	Low of candle 13 (2,399.9) − 0.5	Low of the pullback (2,397.3) − 0.5
Stop loss price	2,399.4	2,396.8
Distance	4.8 dollars	7.4 dollars
Lot = 100 ÷ (distance × 100)	0.208 → 0.20	0.135 → 0.13
Loss if hit	0.20 × 4.8 × 100 = 96	0.13 × 7.4 × 100 = 96.2
Spread of 0.25 as a share	5.2% of the distance	3.4% of the distance
Target at RR 1:2	2,413.8	2,419.0

What the table shows:

• The dollar risk is the same, about 96 dollars either way, because the lot adjusts to the distance. A tight stop is not “less risk”, it just means a bigger lot.
• The M15 stop is very tight compared with normal movement. In this series the average M15 candle range (high − low) is 3.87 dollars. A 4.8-dollar stop is only about 1.2 times one candle; 7.4 is about 1.9 times.
• The target has to move too. Both targets sit above the H4 high (2,410.7), and the H1-based target is another 5.2 dollars further. Check whether a major level is in the way.

The lot calculation is covered in detail in Position size from your Stop loss. To use volatility to set the distance, read ATR and Stop loss placement.

What we found in our backtests

Direction filters from H4. In our backtests on gold (XAUUSD) from 2022 to September 2026 (2022–2025 on 1-minute price data, 2026 on real tick data), we entered one trade a day at 06:30 Thai time (UTC+7), with a 5-dollar Stop loss and a 30-dollar target, and chose the direction with an H4 rule. Each rule was compared against two controls: buy every day and sell every day.

• Colour of the last closed H4 candle: beat both controls in only 1 of 5 years
• Colour of the H4 candle still forming: 0 of 5 years
• H4 close above or below the H4 EMA 50: at best 2 of 5 years, and the 5-year total of every rule was still inside the range you can’t tell apart from luck

A big-trend filter made things worse. In another test, a pullback system on gold M1 from 2022 to 2026 (older years on simulated ticks, 2026 on real ticks), we only allowed trades on the same side as a slow EMA (200, 300 and 500 M1 candles, roughly 3 to 8 hours, acting much like a trend read from a bigger chart). Average profit per trade (at 0.01 lot) fell from 6.31 to 0.98–3.53 dollars. The trades the filter removed were better than the ones it kept, so it wasn’t just fewer trades.

What helped was the size of the swing. Counting every time price pulled back to a Fibonacci level on gold M1 over 5 years of data (2022–2026), with no filter on swing size the result was the same as random. Only swings of at least 1% of price (at a price of 2,400, that is 24 dollars) beat random, by 3–6 percentage points. Swings of that size are structure you can see on a bigger chart, not minute-by-minute noise.

The lesson from our data: a higher timeframe is useful for telling you whether the structure you are trading is big enough. As a simple on/off switch for direction, we haven’t seen it work. All of this is backtesting on one market. Past results do not predict future results, and they are not returns you should expect.

Watch out for timeframe shopping

Timeframe shopping is flipping through timeframes until you find one that supports the trade you already wanted. M15 doesn’t look good, so you check M5. H1 disagrees, so you check H4. In the end every trade has a reason, but the reason came after the decision. It is the same trap as hindsight bias on charts.

Choose your timeframe set before you open the chart, not after you see price. One method, one set.

A simple defence: write your timeframe set into your trading plan, and record in your journal for every trade which chart decided what. If you look back and find trades that used a different set, that’s your warning sign. See Trading journal template for how to log it.

Common mistakes

• Several charts doing the same job. Watching H1 and M30 together is like reading one chart twice. You gain confidence, not information.
• Letting the lower layer overrule the higher one. Seeing three falling M15 candles and deciding the H4 trend is over, when on H4 it is still just one wick.
• A Stop loss from M15 with a target from H4. The stop is tighter than normal movement and gets hit often. Or the reverse: a wide stop with the same lot as usual, so the risk per trade is bigger than planned.
• Waiting for every layer to line up perfectly. By the time every chart looks neat, price has often run far from the zone, so you enter late with a wide stop.
• Forgetting that H4 candles depend on server time. A rule based on the colour or close of H4 candles on one server may see differently shaped candles on another.
• Stacking more filters from the higher chart. Each filter removes trades. Unless you measure that the removed trades really are worse than the ones left, you may be throwing away good ones. See Combining indicators without redundancy.