Can a technical strategy be applied on any time frame? Eg 1min

It seems like when trading the hourly it works but on lower time frames like 5min etc it falls apart. Why? FX btw

 

The lower the timeframe, more noise is seen, more false signals in lower timeframes, more fake outs by most. In my back testing and trading, stops have to get larger for signals not to be clipped by noise, One has to study good deal to know probabilities of patterns well to understand what is not noise and what is firm signal. Larger timeframes usually means larger returns.

 

Of course, 5mn is 1/12th of 60mn so you have to scale expectations appropriately. It means your 60mn is not perfect, any discrepancy is magnified 12x, then you have slippage and commissions having a more pronounced effect, so that will magnify the imperfection 15x to 20x. It's just that simple, the higher the timeframe the more you can get away with.

At lower timeframes it's not noise, they are just submoves within the push and pull of higher timeframes, it's no different than a weekly influencing a 60mn, only you have very little time to react on minute charts and below. When you use HFT perfection (ms & sc) with long timeframes such as 240mn you get some funny results, +20% monthly return via one trade over a few weeks monitoring a few minutes per day.

 

It's a pretty common experience that the longer one trades, time frames lengthen, not shorten.

Think about that and you'll save yourself a lot of time and grief.

 

That was true from my own experience! Where were you when I needed this advice?

Regards,

 

Can I ask why more false signals & noise? My backtesting results from 2010-2017 shows ~ 60% win rate on hourly timeframe & multiple markets but completely breaks down when i trade it on 5min. Algos? Businesses exchanging money? I am very objective.

 

I wonder if HFT algos use their systems on higher time frames what the result will be? Not counting spreads.

 

Actually it's the opposite. I benchmark my intraday systems against longer term systems I designed. Both outperform the indices, but the intraday models grossly outperform (due to leverage) the long term models. I wish they didn't, as I'm getting tired of slaving away in front of my machines. People who fail intraday generally move on to higher time frames. No guarantee of progress there either.

 

Likely I was getting wacked for the first few years, like everyone else.

 

Eventually you'll realize that you're the weak link in your system, and you haven't included your burn-out in the model.

As Yoda would say, "Wisdom at that point comes."