Multi-Timeframe Analysis: Higher Timeframes for Direction, Lower Timeframes for Location

Different trends on different timeframes are normal

It is common for the same coin to be rising on the daily chart and falling on the 1-hour chart. Just as one 4-hour candle bundles four 1-hour candles, a higher-timeframe candle bundles several lower-timeframe candles and smooths away the ripples (see the timeframes article). A single pullback on the higher timeframe can look like a downtrend lasting days on the lower one. Multi-timeframe analysis does not try to remove this difference; it gives each timeframe a different question to answer.

Top down: direction, price zones, location

The usual order is to work down from higher timeframes to lower ones. First, on the highest timeframe, look at the direction in which highs and lows are stacking up and at the slope of the moving averages. Next, on the middle timeframe, mark meaningful price zones such as the latest swing highs and lows or support and resistance. Finally, on the lowest timeframe, see where price currently sits relative to those zones and where the structure is that a stop could be placed against. The higher timeframe answers "which way is the market flowing?" and the lower one answers "where in that flow are we now?"

• Higher (e.g., daily): direction of the flow, slope of the moving averages
• Middle (e.g., 4-hour): key price zones, the latest swings
• Lower (e.g., 1-hour): where price is now, structure for a stop

The ratio between timeframes

A common convention is to space the three timeframes roughly 4 to 6 times apart. Daily to 4-hour is 6x, and 4-hour to 1-hour is 4x. If the ratio is too small (1-hour and 2-hour), the two charts show almost the same thing; if it is too large (weekly and 5-minute), the moves in between drop out and the two no longer connect. Because crypto trades 24 hours a day, one weekly candle is seven daily candles and one daily candle is 24 hourly candles. Which combination fits depends on how long you intend to hold a position, and there is no set answer.

Higher timeframe: see the direction

The figure shows 60 hypothetical daily candles. The close is above the EMA20 (the 20-bar exponential moving average), the EMA20 is above the EMA50, and both lines are rising. Highs and lows have also stepped higher one after another. On the daily chart this can be summarized as being inside an uptrend. The last two candles, however, are bearish. These two days, which look like a small pullback within a long advance on the daily chart, look quite different on the 1-hour chart in the next figure.

Lower timeframe: see where price is

Spread the same two days over 48 one-hour candles and you get a move in which highs and lows step lower one after another, with price below the 1-hour EMA20. Looking at the 1-hour chart alone, it is a shape you could call a downtrend. If you then carry the daily EMA20 over as a horizontal line, you gain information about location: price is currently coming down toward the higher timeframe's reference line. A common way to use this method is to draw the price zones set on the higher timeframe onto the lower one and watch how the lower timeframe's structure changes near them.

Two readings when the timeframes disagree

When the higher timeframe is rising and the lower one is falling, there are usually two readings. One is a pullback (a dip) within an uptrend: the decline on the lower timeframe stops near a higher-timeframe price zone, and the lower timeframe's highs and lows start rising again. The other is the start of a trend change: the decline on the lower timeframe continues until the higher timeframe also closes below its most recent low. The problem is that in the moment you cannot tell the two apart. Even a large decline starts out looking like a pullback on the lower timeframe. That is why the method is sometimes used less to pick an interpretation than to decide in advance where you will admit you were wrong (for example, a daily close below the most recent swing low on the daily chart).

The scores in the Multi-Timeframe Trend Matrix

This site's Multi-Timeframe Trend Matrix checks three conditions on closed candles for each of the 15-minute, 1-hour, 4-hour, daily and weekly timeframes: whether the close is above the EMA20, whether the EMA20 is above the EMA50, and whether the EMA20 is higher than it was 5 bars ago. Each scores +1 or −1, so the total is one of +3, +1, −1 or −3. When the daily and weekly are both up but the 15-minute or 1-hour is down, it is flagged separately as "Pullback in uptrend". The score summarizes where the moving averages sit, so it only shows whether there is a trend right now; it does not mean the trend will continue.

A common trap: more timeframes make the answer easier

With five or so timeframes open, at any given moment some are rising and some are falling. That makes it easy to pick whichever timeframe agrees with the direction you had already settled on and use it as your evidence. This is confirmation bias at work, hopping from one timeframe to another (see the article on the limits of chart analysis). One way to reduce it is to decide in advance which timeframe plays which role and not change that midway. Also remember that the higher the timeframe, the later its signals. For a change in the weekly trend to be confirmed, several weekly candles have to close, and by then price has already moved a good deal.