Selecting a Timeframe for Your Charts

[blueberrycake 10]

Having recently transitioned from EOD trading where I used daily charts to intraday trading, I've been giving a lot of thought to the right timeframe (1m, 5m, 15m, CV, etc.) for my charts. After much experimentation, I'm coming to the conclusion that it's best to use a chart resolution where your expected trade duration for the type of move that you're trying to capture will typically be in the range of 2-20 bars. If you go below that (e.g. 1m chart while trading multi-hour swings), you'll find legitimate reversal patterns forming on the chart, which would rightfully put you out of a trade much sooner than with higher timeframe charts. That's not to say that you shouldn't fine-tune your entries/exits with a lower timeframe chart or confirm key support/resistance and trend with an even higher timeframe chart, but the key entry/exit decisions seem easiest on a chart where your trades have an average duration of 5-15 bars.

 

Is this observation consistent with the experience of other intraday traders?

 

-bbc

[HowardCohodas 10]

I'm a fan of the triple view. I prefer the approach where whatever time frame you choose to trade, you should use your same analytics by scaling the time frame and scaling down the time frame by the same factor. If all three agree, enter the trade.

[Kiwi 16]

It depends what you're trying to achieve bbc.

 

You don't want to go down so far that you get shaken out. But it might be that you get a superior pattern entry on a lower timeframe. On a longer timeframe you might treat a break of bars as a reason to exit. On the lower you might require a break of a more complex structure.

 

So it does depend what you design as your entry and management methods. It also depends whether you can maintain discipline at any particular level.

[SIUYA 11]

Pretty much agree with everything said above.

Possibly only adding that given your transition from EOD trading often its important to form the actual decision making process for direction on the higher time frame and just look for the razor sharp/exact entries on the lower time frames. ...... AND DONT MIX them..... this is where you will start to second guess yourself. Once you go down this path you will then find other errors and issues creeping in....all stemming from this one simple thing.

[wrbtrader 10]

Keep it simple via backtesting your method on different time frames and then choose the best interval based upon the results you're comfortable it.

 

Also, your results to determine a suitable timeframe will differ in comparison to someone else if you're trading something different, trading a different time of day, if you're not using an auto-trading system, your ability to manage a particular timeframe on your monitor et cetera.

 

I myself use multiple time frames which is why I use more than one monitor to manage multiple timeframe. Simply, if I get a trade signal on the 1min chart...I take it. Just the same, 15mins later and I get another trade signal on the 5min chart...I take it. I have on my monitors the 1min, 2min, 3min, 5min, 15min and 60min charts...trying to find trade signals on any of those chart intervals. Thus, I'm not married to one particular timeframe.

 

My point is if you want "more" trade opportunities...you're going to need to follow more than just one timeframe and learn to manage them on your monitor(s).

 

Mark

 

Having recently transitioned from EOD trading where I used daily charts to intraday trading, I've been giving a lot of thought to the right timeframe (1m, 5m, 15m, CV, etc.) for my charts. After much experimentation, I'm coming to the conclusion that it's best to use a chart resolution where your expected trade duration for the type of move that you're trying to capture will typically be in the range of 2-20 bars. If you go below that (e.g. 1m chart while trading multi-hour swings), you'll find legitimate reversal patterns forming on the chart, which would rightfully put you out of a trade much sooner than with higher timeframe charts. That's not to say that you shouldn't fine-tune your entries/exits with a lower timeframe chart or confirm key support/resistance and trend with an even higher timeframe chart, but the key entry/exit decisions seem easiest on a chart where your trades have an average duration of 5-15 bars.

 

Is this observation consistent with the experience of other intraday traders?

 

-bbc

[blueberrycake 10]

I myself use multiple time frames which is why I use more than one monitor to manage multiple timeframe. Simply, if I get a trade signal on the 1min chart...I take it. Just the same, 15mins later and I get another trade signal on the 5min chart...I take it. I have on my monitors the 1min, 2min, 3min, 5min, 15min and 60min charts...trying to find trade signals on any of those chart intervals. Thus, I'm not married to one particular timeframe.

 

I've been noticing something very similar in my trading of the currency futures (6E, 6J, 6B and 6A). Basically you often have very clear market structure in terms of S/R and trend on one timeframe one day, and then on a different timeframe another day. Initially I've been passing on any signals that weren't in my primary timeframe that I had done my testing with (15m), but I'm now thinking that I should retest my method (retracement to minor S/R w/ trend) on multiple timeframes, and potentially take signals in whichever timeframe has the clearest signal, then simply adjust the number of contracts traded to keep risk constant. Sounds like you have already taken this approach as well.

 

-bbc

[cuttshot 11]

I'm going to take a very different approach than the answers listed above. Try looking at range or tick charts. I have traded time based charts as well as range and tick and find that range and tick can really increase performance. The main problem I have with time based charts is that when markets start moving you can have a lot of action within that set amount of time. For example, if you are trading a 5 minute chart and you are waiting for the close of a candle to get in you might miss a good majority of the move. On the other hand, if you were looking at a 233 tick chart for example you might have 10 candles during that same 5 minutes (because candles are formed based on the number of trades or a certain range of price action in the case of range bars as opposed to a set amount of time). So you are able to get in sooner and capture more of the move.

 

You will also find that tick and range charts are a lot cleaner looking. This goes back to what I talked about above. When the market starts to move you are going to see a lot of nice smooth moves. When I trade using tick or range I don't find it necessary to look at multiple time frames for confirmation either. Looking at multiple time frames for confirmation can be very difficult at times. Atleast for me, when I do that I can almost always find a reason for me not to take a trade. I don't want to over complicate things. If I can look at one chart and base my decision off that one chart then I find I'm more focused and accurate with my trading.

 

Just something else you might want to take a look at.

[MadMarketScientist 11]

I'm onboard what cuttshot states.

 

I have a rule. If I'm daytrading it has to be tick, range or renko bars.

 

If I'm swing trading, time interval bars are perfectly fine.

 

I've tried to vary that over the years and my results are always better if I don't.

 

MMS

[BlowFish 10]

Horses for courses. If for example you want to see 'price rejection' at a level I still like time base charts. Holding ticks volume or range constant tends to obscure this information. Actually that is not completely true if you can display a time histogram on your charts. e.g A constant volume chart with a time histogram at the bottom shows these events well.

[MadMarketScientist 11]

I still feel you can just as easily see points of resistance/support on tick charts as time charts - to me possible even better since you may see more attacks of those levels inside of a time interval bar that you'd normally be waiting on to close. However, it is definitely true when it comes to renko - you will not see the extremes of moves that occur inside the value you are using for the renko bar -- it does not plot all price point moves.

 

MMS

[cuttshot 11]

I agree MMS that support/resistance are just as easy to see on tick charts. On a daytrading basis I have yet to see a time based chart outperform a tick or range chart in the long run. I have better consistency with these charts than I ever did on time based.

[wwalker126 10]

Its all about consistencies. All time frames vary from one strategy to another. Meaning, something may work well on a 5 min chart but simply won't apply to a 30min chart. It all begins with basic fundamentals. Find a strategy apply it, then improve. Just as market evolve over time, you as a trade must evolve with it.

 

*Good luck on your next trade*

[BlowFish 10]

Sure S/R are (arguably more so). They are great to see short term 'walls' of resistance. Price rejection is not 'S/R' though.

 

It is a data sampling thing....depending on how you sample data certain things are obscured some emphasised. One of the characteristics of price rejection is that it often happens quickly and on high volume/lots of ticks (a classic Wyckoffian selling climax if you like). It's not a question of 'out performing' it's a question of what market phenomena you want to observe and what you want to de-emphasise through the sampling process. Incidentally if your performance is better with constant tick charts more power to your elbow but it is a non sequitor to suggest that it is the charts performance.

 

A more interesting comparison perhaps is constant volume charts vs tick charts, they are much more analogous. The constant tick chart tends to emphasise what the small lot traders are up to (if that is your interest) as a 1 lot tick has the same emphasis as a 500 lot tick. have you guys looked at constant volume charts? Why do you prefer constant tick over them?

[MadMarketScientist 11]

BlowFish,

 

Good question. What's interesting is a few years ago I was convinced that the volume/share bars were smoother -- just slightly but I felt it was giving me a small, measurable edge. As to why I gravitated back to tick - good question (!) -- I think some of it is because it's easier to move between markets and pick a good time interval for tick charts. For volume, it always seems a little trickier since there is major variance - for example a 377 tick on several markets is ideal for what I do -- but when I try to match with volumes the numbers vary greatly. So maybe laziness on my part to do more work on it.

 

Do think they can be a great way to add some smoothness to tick charts.

 

MMS

 

 

Sure S/R are (arguably more so). They are great to see short term 'walls' of resistance. Price rejection is not 'S/R' though.

 

It is a data sampling thing....depending on how you sample data certain things are obscured some emphasised. One of the characteristics of price rejection is that it often happens quickly and on high volume/lots of ticks (a classic Wyckoffian selling climax if you like). It's not a question of 'out performing' it's a question of what market phenomena you want to observe and what you want to de-emphasise through the sampling process. Incidentally if your performance is better with constant tick charts more power to your elbow but it is a non sequitor to suggest that it is the charts performance.

 

A more interesting comparison perhaps is constant volume charts vs tick charts, they are much more analogous. The constant tick chart tends to emphasise what the small lot traders are up to (if that is your interest) as a 1 lot tick has the same emphasis as a 500 lot tick. have you guys looked at constant volume charts? Why do you prefer constant tick over them?

[cuttshot 11]

BlowFish,

 

I have not done a whole lot with volume charts. Most of my active trading is done with range charts at the moment. I really like how many nice clean quick moves these charts have because they react so fast.

 

I have recently started to look at volume charts again. You're right it would be an interesting comparison to look at volume vs tick charts. I agree with MMS that it does take some work to get the right setting correct going from tick to volume charts. At the moment I'm confident in my trade plan using range charts so it's difficult to motivate myself to mess with something that"s not broke.

[robertm 10]

S/R is visible on any timeframe (and most chart types) you consider has the liquidty and brokerage to allow you to trade it. The lower the timeframe the more noise you have to deal with though. If your candles consist of 3-5 tics (H-L), you probably aren't getting a good view of what's going on.

 

Just remember to point yourself in the right direction.

[BlowFish 10]

I have always thought CV 'smoother'. Though one problem is large block trades can distort them. The DAX sometimes has these particularly around expiration. So lets say you like a 500 contract bar (quite plausible for DAX) then out of the blue you get a 10,000 block tick as an off exchange trade is reported. They scale up well too.

 

Tick charts quite often have sideways 'blocks' (or so it seems to me) if you want to easily identify these pauses they seem pretty good for that. I quite like them on the fastest time frames (or should I say at the highest rate of sampling).

 

I agree robertm S/R is pretty easy to see on just about any chart type or periodicity. The more obvious it is (the more people playing it) the better. Of course the big question is whether it s holding when price revisits it at the hard right edge of the chart.

 

Bar types I would like to do more work on involve market delta e.g. constant delta, delta' shift'.