Minimizing noise in Trend identification?

[Zwaen]

%%
I, also noticed that, on increased time frame;
but with a lot of trend study, lots of trend measure tends to increase tons + tons of profit.
Longer time frames can also increase pound + pounds of profit;
but may not be quite as fun or have as much bid\ ask+ slippage fun
Noise may not be a bad thing;
shooting sports, hunting seasons are full of noise, ear muff makers sell plenty of ear muffs LOL.[Edit, Iron chief/ MACD =too far from price [for me], its below\ volume on charts ;
a moving average is near price, or easily glanced @ even if 200dma is far from price ]

The day I went up in timeframe - "swingtrading" - was the day everything became a lot easier. A lot.

 

[ironchef]

And when you increase the timeframe, you will notice that you will be missing
tons and tons of trading opportunities.

Then traders will become very angry/frustrated/mad and suffer sleepless nights.
Then traders will do revenge trading and those things.

Words of wisdom from a wise man.

To each his own.

 

[ironchef]

A headache and prove nothing.

Thank you for your encouragement.

Here is why I want to do it:

I noticed, even though win rate is ~50/50, 75% of the days were green, why? I stopped whenever sum of trades turned green. Losing days usually happened when I over traded. To me, that could be some sort of martingale effect and not that I have an edge. If so I better find out for sure before I commit to full size.

 

[ironchef]

The day I went up in timeframe - "swingtrading" - was the day everything became a lot easier. A lot.

Yes and Buffett said his time frame is never, a lot easier than swing.

Kidding aside, for intraday I am not convinced longer time frame is better. It all depends on your method. It is always a trade off.

 

[Zwaen]

Yes and Buffett said his time frame is never, a lot easier than swing.

Kidding aside, for intraday I am not convinced longer time frame is better. It all depends on your method. It is always a trade off.

I would guess it would be less variables influencing price, significantly.

 

[themickey]

I am not convinced longer time frame is better. It all depends on your method. It is always a trade off.

No matter what time frame one chooses, there will be stress which accompanies it.
Short term trading has its stress; more noise to deal with, more trading action, increased trading tempo, more decision making, more mistakes....
Longer term has its stress; holding overnight/weekends, impatience, boredom distraction, larger amounts of money tied up maybe doing nothing or turning red....

We're all wired differently so just find your own niche which is most comfortable.

Intraday, I would imagine it is the same, a 1 minute or 5 minute bar trading will have more noise than 1 hourly or 4 hourly bars.
Obviously longer tf bars have more drawdown and thats what we live with.

My trading will experience several thousand dollars of drawdown each day, most guys can't hack that sort of pain, but the drawdown is not really important, it is whether the system is logical and runs intune with how the competition code their robots.

Yesterday was a red day on asx but I was up a small $3k in the morning which I thought was ok for a red day, but towards the end of day down a couple of hundred, it didn't fuss me, as today is a new day.

 

[ironchef]

I would guess it would be less variables influencing price, significantly.

But sometimes, in real time, in shorter time frame, the dynamics of price movement tell a story?

 

[ironchef]

Markets are (again IMO) much better approached or viewed using ensemble methods.

What is ensemble methods?

 

[Specterx]

What is ensemble methods?

Just using a diversity of models or inputs to come up with a prediction/forecast/hypothesis. You have a trend metric which you feel gauges price momentum reasonably well most of the time; from there, it's generally better to add factors to your model describing other states or features of the marketplace besides price momentum, versus torturing the metric further with advanced math. At least, that's my opinion and experience.

 

[ironchef]

Just using a diversity of models or inputs to come up with a prediction/forecast/hypothesis. You have a trend metric which you feel gauges price momentum reasonably well most of the time; from there, it's generally better to add factors to your model describing other states or features of the marketplace besides price momentum, versus torturing the metric further with advanced math. At least, that's my opinion and experience.

Thank you for the explanation.

In other words, you use multiple factors, e.g., momentum, volume, forecast, filters... Out of those, you use a system to rank/factor and come up with a verdict for trading? I guess the devil is in the details.

Regards,