2 statement that contradict each other that I've heard many times, what's your opinion?

New comment
 

Hi, I've met this 2 statement that's contradict with each other many many times when reading about trading strategy. Here's the statements:

   1. Not all strategy works on all timeframe. And what I take from this statement is that each timeframe act differently to each other.

   2. In multi-timeframes analysis there's a word 'confluence' with each other. So, there's a time when the price across multiple timeframes behave the same. 


I looked further about those statement and I think both is right. A question arises when someone talked about diversified their strategy using different timeframe instead of currency. How about your opinion?

 
I believe in first statement more. At least it applies to more circumstances.
 
akbar giro:

Hi, I've met this 2 statement that's contradict with each other many many times when reading about trading strategy. Here's the statements:

   1. Not all strategy works on all timeframe. And what I take from this statement is that each timeframe act differently to each other.

   2. In multi-timeframes analysis there's a word 'confluence' with each other. So, there's a time when the price across multiple timeframes behave the same. 


I looked further about those statement and I think both is right. A question arises when someone talked about diversified their strategy using different timeframe instead of currency. How about your opinion?

I suggest to do following test:

Take an EMA cross as a signal (or whatever you want as a signal) then take 100.000 candles from each timeframe.

Now run the signal on these 100.000 candles and see how many signals you will get.

I predict, it's quite similar.

Why I think so is because time-frames are fractal in their nature. It's like the Mandelbrot. No matter how deep you go, it will continue to "repeat" and show similar patterns.
 
Dominik Egert #:
I suggest to do following test:

Take an EMA cross as a signal (or whatever you want as a signal) then take 100.000 candles from each timeframe.

Now run the signal on these 100.000 candles and see how many signals you will get.

I predict, it's quite similar.

Why I think so is because time-frames are fractal in their nature. It's like the Mandelbrot. No matter how deep you go, it will continue to "repeat" and show similar patterns.

I agree with this.

 
Then why strategy works at some timeframe and completely failed (failed after optimization) at some timeframe?
 
akbar giro #:
failed

I think time frame is mostly imposed to chart. Who says there should be a time frame to a chart at all?

One time frame works and the other fails simply because (maybe) the market maker is taking action upon H1 and not M15.

 
Strategies that work on one time frame but not on the other are "time optimized"

What i mean is this: when you run an optimization, you optimize for price changes only. You usually have more or less no access to the time scale. What happens is, you actually optimize for both axis, time and price. But you do not actually take the time aspect into consideration. You just accept it.

Since the time axis is linear and has no actual "randomness" to it, like price axis has, you get as a result from an optimization run a fitting of price X time in the end result. But you don't consider the time axis in such way that you would close a position after a certain amount of time.

This also lays the foundation for renko charts, and why they are fundamentally different from price/time charts.

A renko chart is basically exchanging the units on the time axis from seconds or minutes to price chunks.

So a strategy that works on a certain timeframe is actually just curve fitted to the time axis. And therefore it cannot work on other time-frames.

I personally think a strategy that works only on certain time frames (in algo trading) is actually not a stable strategy.

It is something different when you execute this manually, as you have the aspect of human reactions time as well. We humans are slow, and not as deterministic as an algo. So you might fail to execute fast enough, or you miss it because you are busy with other things. This applies especially to low time-frames. But can also happen on higher time frames, depending on the time spent in front of the chart.

Here is an example: a swing trader usually works on higher time frames, because he will not be sitting in front of the charts all day, but if he were, he could also perform his swing trading on the M1 timeframe.

The smaller the timeframe, the higher your spread influence is on your trade. The higher your timeframe the less you need to care about overnight spreads. - This is well known, but it has actually no influence on the generation of signals based on candles. Because you are actually just zooming in or out.
 
Dominik Egert #:
But you don't consider the time axis in such way that you would close a position after a certain amount of time.

If I'm correct, to avoid 'time optimized' is to consider a time when exiting the trade by using time aspect into consideration (e.g exit at 5 bars later)?. I've rarely met a strategy that use exit at bar later except in daily timeframe.

Dominik Egert #:
I personally think a strategy that works only on certain time frames (in algo trading) is actually not a stable strategy.

 Could you elaborate more about being 'not stable strategy'? is it will become lose strategy in the future or is it unpredictable in making a profit?

Personally the strategy that I create is using indicator and not solely on price action. Is it because of that, that the strategy works only on some

 
The two don't even contradict each other as much as you might think.

Just because there are strategies that don't work in all TF doesn't mean ther can't be profitable confluences (in others).
 
akbar giro #:

If I'm correct, to avoid 'time optimized' is to consider a time when exiting the trade by using time aspect into consideration (e.g exit at 5 bars later)?. I've rarely met a strategy that use exit at bar later except in daily timeframe.

 Could you elaborate more about being 'not stable strategy'? is it will become lose strategy in the future or is it unpredictable in making a profit?

Personally the strategy that I create is using indicator and not solely on price action. Is it because of that, that the strategy works only on some

difficult to elaborate into one direction. As a matter of opinion, or fact, I see a strategy being able to make profits on one timeframe, but failing on a neighbor timeframe, is a strong indication for over fitting.

If you compare iE H1 to M1, IT might be due to spread, so it's not as strong an indication for over fitting, but maybe due to the way this strategy capitalizes on its profit taking.

It might be unpredictable, it might be a failure in the long run. Either way, it is in both cases an unstable approach.
 
akbar giro:

Hi, I've met this 2 statement that's contradict with each other many many times when reading about trading strategy. Here's the statements:

   1. Not all strategy works on all timeframe. And what I take from this statement is that each timeframe act differently to each other.

   2. In multi-timeframes analysis there's a word 'confluence' with each other. So, there's a time when the price across multiple timeframes behave the same. 


I looked further about those statement and I think both is right. A question arises when someone talked about diversified their strategy using different timeframe instead of currency. How about your opinion?

Trading multiple time frames is a way of entry and exit points diversification (staggered entries) to accomplish an heuristic curve forward fitting.

12
New comment