Candlesticks candlesticks everywhere!

[djc0]

Something occurred to me today.

Candlesticks show, for the given time interval, the open and close (body) plus high and low (wick) points. The appearance of individual and collective candlesticks are commonly used to indicate trends and global buy/sell behaviour.

However what happens to the appearance (especially of key candlestick indicators) if you shift the time interval over which each candlestick is calculated. E.g. on 15M, shift the calculation by 5 minutes.

Won't that change the appearance of all candlesticks (especially the body, since the open and close will be different), perhaps even turning previously key trend indicators (e.g. doji) into much more benign markers (or vice versa)?

Does that make sense, and is there an answer?

 

[shareprofit]

yes your ? makes sense an the answer in a nut shell is YES.

If u change the time frame the result will / can be different. If you get the same result from multi time frames then it can add strength to the reading.

cheers.... have fun

 

[Kryzz]

candlestick patterns will generally carry more weight the longer/higher the timeframe imo

 

[djc0]

Perhaps i wasn't clear enough.

I'm not talking about comparing candlesticks between different timeframes (e.g. 15M vs. 3H).

What i'm talking about is, for a give timeframe (e.g. 15M), that there's nothing special about the starting point in time for calculating each candlestick, but we seem to assume there is. The starting point will determine (to some small or large degree) the appearance of the candlestick, and this is no unique.

Let me try and give an example.

1st candlestick is calculated across 0 to 15 minutes, 2nd candlestick 15 to 30 minutes, 3rd across 30 to 45 minutes, and so on.

Lets say instead i move the starting point for calculating each candlestick. Now it's: 1st across 5 to 20 minutes, 2nd across 20 to 35 minutes, 3rd across 35 to 50 minutes, and so on.

In this case the open, close, high and low points of each candlestick will be different (because you're now sampling different parts of the underlying continuous sequence). I.e. the candlesticks will appear different - that doji may no longer be a doji! So the appearance of individual candlesticks are not unique.

Have i explained that clearly enough (sorry if not!)? Has this been discussed before? Is this not a problem for using candlesticks to identify price movement?

 

[Mr J]

Won't that change the appearance of all candlesticks (especially the body, since the open and close will be different), perhaps even turning previously key trend indicators (e.g. doji) into much more benign markers (or vice versa)?

Of course, which is why a trader picks a chart that matches the timeframe that he or she will trade.

 

[djc0]

Of course, which is why a trader picks a chart that matches the timeframe that he or she will trade.

Sorry if i'm misunderstanding, but what do you mean here by "picks a chart"? If i'm watching and trading in the 15M chart (only), the candlesticks can be plotted many ways depending on where along the underlying continuous price line your software is calculating each.

 

[Trembling Hand]

djc you have a point. A chart will look differently if your 15min candles start at 10:08 instead of 10:00.

But you are assuming that candle pattern are a dead set certain indicator. They are not.

With that said it is important where a candle finishes. It tells a lot about what the smart money is trying to do. I have an example somewhere here I will try and fish out.

 

[Sean K]

I always pick a time period that will match what I want the chart to look like.

 

[alwaysLearning]

You want your candle stick shifts to be set such that the majority of other traders use that same shift. (which is probably the default setting in your brokerage software)

If you go messing around with shifting when a candle opens and closes then IMO that is defeating the point of the well defined candle patterns. You want your charts to be the same as institutional traders. (So a 1hr candles closes at 12pm for example and not 12:08pm.)

But since you would use the default shift, then if you change your time frame to anything (5min, 15min, 1hr, 4H, daily etc), at least you will consistency with what the majority of other traders see on their screen.

 

[djc0]

Thanks all.

With that said it is important where a candle finishes. It tells a lot about what the smart money is trying to do. I have an example somewhere here I will try and fish out.

Is that because of collective trader reaction to the candles themselves (in the same way that everyone assuming that past support/resistance will continue can create future support/resistance)? I'm guessing as well that different brokers/software packages plot candles differently (as well as using slightly different data)?

This is not a criticism of using candles to predict chart movement. Clearly for many this works exceptionally well. Just trying to understand the underlying "meaning" of what i'm looking at.

 

[Trembling Hand]

To my point previously. Here is a good example of candles being printed and why you want them in sync with everyone else.

http://tremblinghandtrader.typepad.com/trembling_hand_trader/2008/02/painting-a-pict.html

 

[djc0]

You want your candle stick shifts to be set such that the majority of other traders use that same shift. (which is probably the default setting in your brokerage software)

...

But since you would use the default shift, then if you change your time frame to anything (5min, 15min, 1hr, 4H, daily etc), at least you will consistency with what the majority of other traders see on their screen.

But only true for your particular broker ... ? The market is moving based on global buy/sell, and the majority of traders are probably not looking at the same candles as you and hence not reacting to the same patterns (i guess the broad patterns should be the same).

 

[Trembling Hand]

But only true for your broker ... ? The market is moving based on global buy/sell, and the majority of traders are probably not looking at the same candles as you and hence not reacting to the same patterns (i guess the broad patterns should be the same).

What

data comes from the exchange.

 

[alwaysLearning]

But only true for your broker ... ? The market is moving based on global buy/sell, and the majority of traders are probably not looking at the same candles as you and hence not reacting to the same patterns (i guess the broad patterns should be the same).

Break your day up into sessions.

Asian Session
Euro Session
US Session

Adjust your candles according to each session depending on when each market opens. (if you really want to)

Most brokerage software allows you to define your timzone. Adjust your timezone according to whichever session you are currently in.

 

[djc0]

What

data comes from the exchange.

The exchange provides raw or processed (e.g. into timeframes, candles, etc) data to the broker?

If not, then we'd need to assume that every broker is doing the processing in the same way (i.e. picking the same starting point for each candle). Otherwise different brokers will be presenting the candles differently.

 

[Trembling Hand]

The exchange provides raw or processed (e.g. into timeframes, candles, etc) data to the broker?

If not, then we'd need to assume that every broker is doing the processing in the same way (i.e. picking the same starting point for each candle). Otherwise different brokers will be presenting the candles differently.

Mate you are lost!

 

[alwaysLearning]

The exchange provides raw or processed (e.g. into timeframes, candles, etc) data to the broker?

If not, then we'd need to assume that every broker is doing the processing in the same way (i.e. picking the same starting point for each candle). Otherwise different brokers will be presenting the candles differently.

The candles should look pretty much the same. There may be some slight variations on when a candle on the daily chart opens/closes (depending on the timezone) but the end result is that the pattern should look the same.

You can test this.

1. Download MT4 from fxpro (the demo)

2. Set up an Oanda account. (demo)

Change the time zone in Oanda and see how it affects the candle patterns.

 

[djc0]

Mate you are lost!

Probably!! Working hard to find my way though!

 

[Mr J]

djco, the candles are calculated the same. Not all brokers offer the same quality data, and therefore candles may look slightly different, but this is due to data accuracy, not the calculation of the candles. All candles are calculated the same way. Time candles, for example, are all calculated by showing the high, low, open and close over a certain time period. This does not change.

I think you're trying to overthink this. Candles are very, very simple things, and it should be clear that the differently measured candles (ie. 5 min vs 60 min) should appear differently, and therefore give different signals. Think of different timeframes as stories within a story.

 

[Nathan G]

Your charting software is receiving price data from the server as a set of open, close, high & low values for a specific time period. These are just numbers.

The software will then draw a graphical representation of this data, in your case as candlesticks. How the data is drawn is controlled by your setup of the charting software (and whatever limitations/constraints it has).

If you don't wish to be constrained by "time" as being the x-axis on the chart, then look at chart types which don't factor time into thier representation, for example Point & Figure.