Pictorial Analysis of Price Action

[Rapunzel]

Hi Folks, long time hope you're all great and not losing!!

I've been pondering the integration of an additional element into my trading decision-making process and am eager to hear your thoughts or experiences on this matter.

Currently, my trades are primarily informed by quantitative models that signal entry points. However, I've come to recognize the invaluable insights that can be gained from direct market observation — or price action (PA) — a technique some traders use exclusively with great success.

The challenge I'm facing is how to programmatically merge this 'screen time' experience with my quantitative approach, especially when timing is critical between receiving a signal and making an entry decision.

My idea revolves around developing an algorithm that can analyze real-time price movements (or possibly even visual representations of these movements) and match them with similar patterns observed during the model's historical training phase. This pattern recognition could then be used to predict the likelihood of a successful trade based on past outcomes.

In practical terms, when a trading signal is received, the system would identify the four most similar historical trades based on their PA characteristics. If any of these past trades were unsuccessful (i.e., they don't have a 100% success rate), I would consider bypassing the current trading opportunity.

Has anyone here experimented with integrating historical PA analysis into their trading strategy in such a manner? I'm curious about your insights, any potential pitfalls to watch out for, and suggestions for implementing this idea effectively.

Thank you.

 

[MACD]

Hi Folks, long time hope you're all great and not losing!!

I've been pondering the integration of an additional element into my trading decision-making process and am eager to hear your thoughts or experiences on this matter.

Currently, my trades are primarily informed by quantitative models that signal entry points. However, I've come to recognize the invaluable insights that can be gained from direct market observation — or price action (PA) — a technique some traders use exclusively with great success.

The challenge I'm facing is how to programmatically merge this 'screen time' experience with my quantitative approach, especially when timing is critical between receiving a signal and making an entry decision.

My idea revolves around developing an algorithm that can analyze real-time price movements (or possibly even visual representations of these movements) and match them with similar patterns observed during the model's historical training phase. This pattern recognition could then be used to predict the likelihood of a successful trade based on past outcomes.

In practical terms, when a trading signal is received, the system would identify the four most similar historical trades based on their PA characteristics. If any of these past trades were unsuccessful (i.e., they don't have a 100% success rate), I would consider bypassing the current trading opportunity.

Has anyone here experimented with integrating historical PA analysis into their trading strategy in such a manner? I'm curious about your insights, any potential pitfalls to watch out for, and suggestions for implementing this idea effectively.

Thank you.

That is a way that I trade and works well.

 

[Bad_Badness]

Essentially you are trying to integrate the concept of momentum. Hence basic physics applies, Movement along a direction, rate of movement, and rate of acceleration, all over time.

Work on developing those concepts, then add in additional analysis based on time of day, volume, and short term historical context (as in about 2-4x of the movement you are looking at)

From there you will find some fertile ground.

BTW: the papers on order engines on the NASDAQ web site provide insight on how they look at the issues.

 

[Rapunzel]

Thanks, volume is something that was not actually considered with my initil Idea. I wonder f the PA would already reflect this or how to extract that aspect and integrate it into part of the recognition process.

 

[MACD]

 

[Concinnity]

Thanks, volume is something that was not actually considered with my initil Idea. I wonder f the PA would already reflect this or how to extract that aspect and integrate it into part of the recognition process.

For me the PA seems to show this alone, particularly if you're talking about catching smaller moves in futures like 13-25 ish NQ points or 3-5 ES points. I've found no more success adding volume into it. For stocks I could see where volume may be more useful. Just didn't find the time invested in looking at volume worth vs tracking the other things I am doing. Now if I could have volume analysis fully automated and fed in an easy way to quickly identify it may be worth it.

What you're looking to do is very interesting. I can't fully quantify when I see it, but just the way price is moving in a pattern as well as in a micro sense you can see it setting up for a short squeeze or long squeeze sometimes.

Some things I look for are ranges, like if we make a range than drop below it hard it's often clearing out longs, getting in shorts and getting ready for the opposite move(which will often come back to at l east equilibrium of the breakdown point . Also if price action is moving down in a choppy way it's often baiting in longs and clearing out weak shorts, so it's temporarily bearish but this often ultimately ends in a strong move up. Of course macro context is critical as well and the amount of movement in the opposite direction can be greatly reduced or even nullified completely given context.

 

[murray t turtle]

Hi Folks, long time hope you're all great and not losing!!

I've been pondering the integration of an additional element into my trading decision-making process and am eager to hear your thoughts or experiences on this matter.

Currently, my trades are primarily informed by quantitative models that signal entry points. However, I've come to recognize the invaluable insights that can be gained from direct market observation — or price action (PA) — a technique some traders use exclusively with great success.

The challenge I'm facing is how to programmatically merge this 'screen time' experience with my quantitative approach, especially when timing is critical between receiving a signal and making an entry decision.

My idea revolves around developing an algorithm that can analyze real-time price movements (or possibly even visual representations of these movements) and match them with similar patterns observed during the model's historical training phase. This pattern recognition could then be used to predict the likelihood of a successful trade based on past outcomes.

In practical terms, when a trading signal is received, the system would identify the four most similar historical trades based on their PA characteristics. If any of these past trades were unsuccessful (i.e., they don't have a 100% success rate), I would consider bypassing the current trading opportunity.

Has anyone here experimented with integrating historical PA analysis into their trading strategy in such a manner? I'm curious about your insights, any potential pitfalls to watch out for, and suggestions for implementing this idea effectively.

Thank you.

%%
CHART traders do that all the time, many with discretion;
but i do better with business, charts-trades-investments.