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

[birzos]

Can you elaborate on this part? For example if I go down to 1minute there are usual gaps between bars etc. Why does these anomalies affect the setups since its based on supply demand as well? I'm not trying to argue here, really asking questions.

Realised a little while after posting this will go nowhere. You are trying to take a working algo at high timeframe and curve fit it to lower timeframes. It just doesn't work that way, algos require contextual analysis built in to them to function, this is easier at high timeframes, the context changes infrequently. At low timeframes the frequency is magnified, so you need non-stop development. Computers are only as good as you program them, a 5pip anomaly at 60mn can be absorbed, at 1mn it will break the strategy, tomorrow it might be 7pips and next week 3pips. The lower the timeframe the more anomalies, most see that as noise, no, it is inefficient algo developers. The strategy has to account for it, meaning a ground up redesign not a top down refit, something that cannot be done without serious effort, and contrary to retails understanding of markets.

 

[digitalnomad]

Birzos described it well in regards to frequently changing context and non-stop development. Anomalies exist until they don't.

 

[ironchef]

Since order flow and speed edge is non-existent for retail traders we need to design predictive models within a time series with measurable entries/exits. Otherwise you're just shooting in the dark. You can use retail platforms and brokers if you have a grasp on your stats and execution capability. Not easy by any stretch of the imagination. Most are better off pursuing other endeavors, and invest long-term in diversified assets.

Every pro here told me I really could not and should not predict, especially using charts but charts told stories and actions that showed the intend of the market and I could trade using that info.

After about six months staring at charts, I am beginning to see the logic. Do you have any comments or words of wisdom for me in this regard?

Thanks.

 

[ironchef]

Realised a little while after posting this will go nowhere. You are trying to take a working algo at high timeframe and curve fit it to lower timeframes. It just doesn't work that way, algos require contextual analysis built in to them to function, this is easier at high timeframes, the context changes infrequently. At low timeframes the frequency is magnified, so you need non-stop development. Computers are only as good as you program them, a 5pip anomaly at 60mn can be absorbed, at 1mn it will break the strategy, tomorrow it might be 7pips and next week 3pips. The lower the timeframe the more anomalies, most see that as noise, no, it is inefficient algo developers. The strategy has to account for it, meaning a ground up redesign not a top down refit, something that cannot be done without serious effort, and contrary to retails understanding of markets.

Very helpful comments. Thank you.

May I ask you a question: What do you mean by "It just doesn't work that way, algos require contextual analysis built in to them to function"? Some examples of contextual?

Regards,

 

[digitalnomad]

Every pro here told me I really could not and should not predict.

Predict not with certainty, but by attempting to establish bias. You can use price magnitude, and even catalysts. Regardless, you need a trading model to navigate. Quant and statistical analysis are mandatory, or else you can't validate anything, unless of course your gifted enough to grow your account by staring at charts and clicking your mouse.

 

[lawrence-lugar]

It seems like when trading the hourly

People, or day traders, shouldn't scalp in quantity -- But only aim to capture the day's obvious major move(s) in the chart,

Why on earth would you scalp for small fish....Capture only the prized shark,

Expand your micro horizon,
Don't be tempted to quickly grab what's in front of you; let that small fish lead you to something bigger...let that potential snowball blossom into a small avalanche,

Don't follow each zig-zag of the snake...follow its general direction instead,
Which can be observed if you step up higher from a different perspective position,

 

[ironchef]

Predict not with certainty, but by attempting to establish bias. You can use price magnitude, and even catalysts. Regardless, you need a trading model to navigate. Quant and statistical analysis are mandatory, or else you can't validate anything, unless of course your gifted enough to grow your account by staring at charts and clicking your mouse.

Thank you. Good point about "to establish bias".

As for quant and statistical analysis, I doubt mom and pop retail traders like me can do that with nothing but an Excel spreadsheet.

So, we hope we can get some trading signals from staring at charts. To be honest, I cannot do it and have to rely on fundamental analysis to trade.

Regards,

 

[birzos]

Very helpful comments. Thank you.

May I ask you a question: What do you mean by "It just doesn't work that way, algos require contextual analysis built in to them to function"? Some examples of contextual?

Regards,

An algo is a fixed set of rules where if you have a take profit at 50pips but only hits 45pips it will not exit the trade, an experienced trader may see another timeframe anomaly and exit early. When you adjust the take profit level in the algo to account for the anomaly you are implementing context, but may reduce overall profitability.

https://www.cnet.com/uk/news/fujitsu-supercomputer-simulates-1-second-of-brain-activity/

The supercomputer can process 1/240000th of the human brain contextually, so algos are coded to the normal distribution, and fail miserably at the extreme sigma. The lower the timeframe the more frequently it will hit the sigma events.

This is an example of the algo failing due to an uncoded anomaly.
https://www.newscientist.com/articl...es-in-first-fatal-autonomous-car-crash-in-us/

 

[ironchef]

An algo is a fixed set of rules where if you have a take profit at 50pips but only hits 45pips it will not exit the trade, an experienced trader may see another timeframe anomaly and exit early. When you adjust the take profit level in the algo to account for the anomaly you are implementing context, but may reduce overall profitability.

https://www.cnet.com/uk/news/fujitsu-supercomputer-simulates-1-second-of-brain-activity/

The supercomputer can process 1/240000th of the human brain contextually, so algos are coded to the normal distribution, and fail miserably at the extreme sigma. The lower the timeframe the more frequently it will hit the sigma events.

This is an example of the algo failing due to an uncoded anomaly.
https://www.newscientist.com/articl...es-in-first-fatal-autonomous-car-crash-in-us/

Thank you for the explanations.

Thank you also for taking the time to answer my question.

 

[murray t turtle]

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.

%%
Easy for you to say,+ i agree, +your name is not Tony Crabel.That maybe true 80-95 % of the time?? The only 2 profitable daytrading systems i ever saw, the one that traded 20 % less made >>20% more money; but that was only 10 years, so, not a daytrading or any prediction.
Having done daytrading+ kept records, slippage, noise, comissions,P& L; i chose neither daytrading system.LOL

I may not do profitable auto daytrading like Mr Crabel ; but i like his/my rural locate idea