Every chart starts from the same raw material: a stream of ticks, each one a change in the quoted bid or ask. A candle is nothing more than a bucket. Put five minutes of ticks in a bucket and record the first price, the highest, the lowest and the last, and you have an M5 bar. Put four hours in and you have H4. Nothing else is added, and a great deal is discarded.
That framing settles the argument about which timeframe is best. There is no best. There is a bucket size that matches how long you intend to hold, how wide your stop can be, and how often you are willing to pay the spread.
The standard set, and what each is for
| Timeframe | Typical hold | Common use |
|---|---|---|
| M1 | Seconds to minutes | Scalping, precise entry timing |
| M5 / M15 | Minutes to hours | Intraday entries and management |
| M30 / H1 | Hours | Levels and intraday structure |
| H4 | Days | Swing structure, session-to-session bias |
| D1 | Days to weeks | Trend context, major levels |
| W1 / MN | Weeks to months | Position context, long-term ranges |
The cost consequence is the part people underweight. Trading M1 means paying the spread and any commission many times a day for moves that may be twelve pips wide. The same spread against a two hundred pip daily swing is a rounding error. Shorter timeframes need a genuinely lower cost structure to survive, which is why the scalping approach is so sensitive to execution quality and to which broker you use.
Server time quietly changes your chart
H4 bars are not anchored to your local clock. They start from the broker's server midnight, so a server set to GMT+2 and one set to GMT+3 produce four hour candles that begin an hour apart. Same market, same ticks, different bar boundaries, different highs and lows on the candle, different close prices. A pattern that appears on one platform may simply not exist on another.
The daily candle carries the same issue in a larger form. A server offset that places daily close at 17:00 New York produces five weekday candles. A server on a different offset can produce a small sixth candle covering the Sunday session, which distorts anything counting daily bars. Anyone using candle patterns or daily-close rules should check the server offset before assuming the chart matches the analysis they read somewhere else. The mechanics are set out in candle close times.
If your method depends on the shape of a specific candle, write down which server time your charts use and never mix analysis from a platform with a different offset. This is the least glamorous and most common source of two traders looking at the same pair and disagreeing about what happened.
Combining two or three, not six
Top-down analysis works when the layers do different jobs. The higher timeframe sets the context: is this trending, ranging or in a correction. The middle timeframe carries the levels you will actually trade. The lower timeframe times the entry and lets you place a tighter stop than the level alone would allow.
A ratio of roughly four to six between adjacent layers keeps them independent. D1 for context, H1 for levels, M5 for entry is a workable combination. So is H4, M30 and M5. What does not work is D1, H4, H1, M30, M15 and M5 open at once, because at that point there is always a chart supporting whatever you already decided. Two hours later you will be able to explain any outcome, which means the framework predicted nothing.
The rule I apply is that the lower timeframe may refine an entry and may never overrule the context. If the daily says the trend is down and the M5 shows a lovely bullish pattern, that pattern is a countertrend scalp with a small target, not a reversal. Traders who automate the alignment check tend to catch the contradiction faster than those who flip charts by hand.
Matching timeframe to the life you have
The practical constraint is attention. M1 and M5 require you at the screen continuously, because a setup appears and resolves within the time it takes to make coffee. H4 and D1 require you for twenty minutes a day at a fixed hour, and reward patience over reaction. Choosing a timeframe your schedule cannot support is the most common reason a tested method stops being followed.
Volatility matters too. The same instrument behaves differently by session, and a fifteen minute chart during the London open is a different statistical object from the same chart during the Asian afternoon. Our note on trading sessions covers when each pair actually moves, and it is worth reading alongside any timeframe decision, because an M15 strategy tested only across active hours will disappoint if you run it around the clock.
Changing timeframes mid-trade
There is one habit worth naming as a rule breach rather than a technique. Entering on M5 and, when the trade goes against you, opening the H4 to argue the position is still fine is a widening of the stop by another name. The timeframe that defined the entry defines the invalidation. If the H4 was the reason, the stop should have been an H4 stop and the size should have been calculated for it from the beginning.
Trading leveraged products carries a high risk of loss. The chart interval does not change that, and neither does moving to a higher one after the fact. What it changes is how many decisions you make per week, and decisions are where both the edge and the damage come from. The trade-off between them is the same one discussed in day trading versus swing trading.
"Traders ask me which timeframe works. The honest answer is the one you can watch properly. A brilliant five minute method you check twice an hour is worse than an ordinary daily method you follow exactly."
— Alex Onta, Executive Director, SINGUARD
Key Takeaways
- A candle is a bucket of ticks, so the timeframe decides how much detail is discarded, nothing more.
- Server time offsets shift H4 and daily boundaries between brokers, changing candle shapes and closes.
- Use two or three layers with a four to six times ratio, and never let the lower one overrule the context.
- Shorter timeframes pay costs far more often, so they need a lower cost structure to be viable at all.
Frequently Asked Questions
Why does my H4 chart look different from another broker's?
Four hour bars are built from the broker's server time, so a server on GMT+2 and a server on GMT+3 start their bars an hour apart. The price data is the same, the boundaries are not, which moves highs, lows and closes for anyone trading candle patterns.
How many timeframes should I use?
Two or three. One for context, one for the level, optionally one for the entry, each roughly four to six times the one below it. Beyond three you can usually find a chart that supports whatever you already wanted to do.
Is a lower timeframe riskier than a higher one?
The risk per trade is set by your stop and your size, not by the chart. What changes is frequency and cost: shorter timeframes mean more trades, so spread and commission are paid more often and the noise proportion of each move is higher.