Can a trading system be transplanted onto a different timeframe, and still remain profitable?

[Liberty Market Investment]

We've all heard that markets are "fractal", i.e. price patterns repeat regardless of the timeframe used. That's what makes it so easy for the guru to claim that the trading system they are offering are "universal" and can be traded on any time frame. The examples are numerous: from VSA to moving averages, from renko bars to heiken ashis.

But is this really true? Can you use what works say on H1 chart and apply it to M5 or M1 chart?

In my view, markets may be 'fractal' to some extent, but the price patterns on longer timeframes are qualitatively different to those on shorter ones. The underlying reasons include macroeconomics, cross section of participants, greater collective averaging, liquidity needed by heavyweight players, relative effect of news announcements, session considerations, etc. Longer time frames can offer more reliable analysis as each candle is representative of a greater cross section of participants.

This is easy to see on a typical hourly cluster chart of any liquid futures. Let's take Crude Oil as an example.

As you see from this example, volume analysis works great on longer time frames, with 2 ideal trade setups confirmed by using a simple range situation. The highlighted light blue clasters are those exceeding 8000 contracts traded.

What would we see on a 5 minute chart though?

Let's go down to a 5 min chart and try to apply the same approach, which is trading inside the range using clear volume signals. We will notice that excessive volume clusters (that is, more than 800 contracts traded) starts appearing right inside the range. The obvious reason for this is that the intraday traders have to scale in and out to keep up to their brokers' margin requirements, and therefore we would usually see big volume on the session open and close. If we put this information into our range analysis context, it doesn't add value to it. Rather, it gets in the way of clear market interpretations.

To sum up, longer time frames include a greater cross section of participants, their flow patterns are potentially more stable and consistent. Also, patterns like Support/Resistance on longer time frames have greater visibility as they also include S/R from all lower time frames. Therefore, they are therefore watched by a greater number of participants.

What about you, trader? Would you apply the same setup/system to all time frames? Please share your opinions.

 

[NotKnown]

The simple answer is yes and no. Volatility plays a large part. For what I do and my own tools, I can flick between time frames and more or less use the same rules to trade. Getting into the shorter time frames ie 1 min charts things tend to fall apart but mostly above that, things are ok.

 

[virtusa]

My opinion:

Good trading systems should be well balanced. It should not have a problem with a few unexpected, out of range, quotes and stay focused.

The system should work in any market condition, so the system needs some flexibility in interpretation and generating signals. It is rare that a system gets perfect conditions to give signals. Most of the time there is here or there something that is not optimal. This should not confuse your system.

As prices are essential in trading, and the only source of information, all the rest is deducted from the quotes. If you have no prices, you cannot trade.
Basically it should work in any timeframe as all timeframes use the same available information.

Because of the different lenght of the timeframes, reaction time can be faster or slower. The shorter the timeframe, the more fake signals you will get, and the smaller the ATR, so the smaller the profit potential.

 

[andre.salmeron]

Well they do work insofar as you adapt them to meet the requirements of a lower/higher timeframe. Ideally speaking, your system should indicate areas of high relevance in a given chart for you to use as a reference point, from which you'll set your stop loss and targets based on historical volatility (ATR, for instance). However, you can't expect prices to move as steadily as in a higher timeframe.

You mentioned volume on your example, so that too would be something you'd have to adapt into a lower timeframe. From an hourly to a 5-minute chart, you're breaking one candle into 15 ones, which can make it harder to evaluate where unusual volume is coming in (or not). So you'd need, for instance, a much longer moving average of volume to evaluate volume behavior over the same time period (say, the past 5 hours).

The issue with a lot of TA nowadays is that people (including those who teach or use it extensively) often lack a strong conceptual and theoretical framework for their own systems. IMHO.

 

[ironchef]

The issue with a lot of TA nowadays is that people (including those who teach or use it extensively) often lack a strong conceptual and theoretical framework for their own systems. IMHO.

What theoretical framework would you use for 50DMA/200DMA for example, or MACD, or Bollinger band...?

 

[vanzandt]

Not my thing.... but if this thread is meant to evoke a discussion regarding applying a strategy across one time-frame or the other... isn't the simple answer to use back-testing to figure that out?

Like I said however.... not my thing; however (comma)... in this particular case it seems like an obvious answer.

 

[andre.salmeron]

What theoretical framework would you use for 50DMA/200DMA for example, or MACD, or Bollinger band...?

Well for moving averages, they're simply averages of the closing price, not some magical support and resistance. I see a lot of people debating the "best" values for an MA and its calculation method (simple, exponential, Hull, etc.) when the fact is that it works if a lot of people act on it. Despite the simplicity, a lot of people don't realize that.

Same goes for Bollinger Bands: it is simply the "normal" deviation from an average (default 2 StdDevs from a MA). Well that might work if I'm looking for relatively short term stuff, but were I to go for longer-term investments I'd switch to a longer average or higher number of StdDeviations, simply because I can wait for more significant moves (20 SMA and 2 deviations gives a lot of useless signals). Yet, most people simply throw that into their charts and do whatever when price hits the upper/lower band.

I really don't like using MACD, particularly the idea that you buy on crosses or whatever. God, they are simply two moving averages and the moving average of their distance, but people act like it can tell you something other than that. When I say people need a conceptual framework in order to use these indicators, I am not talking about advanced stuff, but rather how the stuff they're using works and how to use them according to their interests, instead of canonical settings.

BTW, sorry if I can't quite express myself, like I said in another topic I've had a few beers.

Cheers, man!

 

[ironchef]

Well for moving averages, they're simply averages of the closing price, not some magical support and resistance. I see a lot of people debating the "best" values for an MA and its calculation method (simple, exponential, Hull, etc.) when the fact is that it works if a lot of people act on it. Despite the simplicity, a lot of people don't realize that.

Same goes for Bollinger Bands: it is simply the "normal" deviation from an average (default 2 StdDevs from a MA). Well that might work if I'm looking for relatively short term stuff, but were I to go for longer-term investments I'd switch to a longer average or higher number of StdDeviations, simply because I can wait for more significant moves (20 SMA and 2 deviations gives a lot of useless signals). Yet, most people simply throw that into their charts and do whatever when price hits the upper/lower band.

I really don't like using MACD, particularly the idea that you buy on crosses or whatever. God, they are simply two moving averages and the moving average of their distance, but people act like it can tell you something other than that. When I say people need a conceptual framework in order to use these indicators, I am not talking about advanced stuff, but rather how the stuff they're using works and how to use them according to their interests, instead of canonical settings.

BTW, sorry if I can't quite express myself, like I said in another topic I've had a few beers.

Cheers, man!

Not a problem. I was just curious since you mentioned theoretical framework for TA. I was hoping you had some insights on theoretical frameworks for TA.

Best to you.

 

[murray t turtle]

My opinion:

Good trading systems should be well balanced...…….
Basically it should work in any timeframe as all timeframes use the same available information.

Because of the different lenght of the timeframes, reaction time can be faster or slower. The shorter the timeframe, the more fake signals you will get, and the smaller the ATR, so the smaller the profit potential.

%%
Exactly virt-usa;
+ more data= more errors.
More errors/exspences on a daily chart than a monthly chart.
Another clue is some hedge funds do super well/most do not.But since some marketmakers/specialists do very well daytrading--exceptions to any rule...…………………………………………………………………………………………...

 

[Liberty Market Investment]

%%
But since some marketmakers/specialists do very well daytrading--exceptions to any rule...…………………………………………………………………………………………...

Really? Do they?