Buy below the 8EMA when both lines are trending up. Now you are stealing my idea.
How can you tell when the trend is over?
- Thread starter mute9003
- Start date
No, I was discussing the 8 and 20 Pullback idea with Mon privately over the weekend. You're just trolling, without knowing the full story.
Man, if you have to resort to using indicators or oscillators, you will never be able to pinpoint exact reversals. It will be an approximation at best, and late-to-the-party at worst. By the time you jump in, provided it's a strong trend, you would be missing a big portion of it. If it were a weak trend, you would be getting in at the tail end of it.
Sure, nobody can identify the area of reversal every time, but you'd be much better off using pure PA than TA.
Sure, nobody can identify the area of reversal every time, but you'd be much better off using pure PA than TA.
TA is drawing silly lines across highs and lows which we can easily learn but how can we learn PA which I understand stands for “Price Action”?
Price action is like TA, which is like pretty young women jumping and spreading their legs over bars in a gymnastics competition.
It all looks well and good but as a spectator won't get you anywhere.
Refer to this link. It covers the basic essentials: https://www.elitetrader.com/et/posts/5745479/
But the most important PA any noob should become acquainted with is the concept of HIGHER HIGH and LOWER LOW. For a refresher, take a look at this video:
You are basically asking for the AI or ML rules to build the model "when does movement on time frame x for instrument y ends?"
A model is always an approximination of a realworld phenomenom. We try to find the best explanatory variables, only asuming causality.
It becomes more complex because definitions of "movement" and "ends" are variables by themself which need to be defined properly.
No model will give you an accuracy near 1.0, nor can you realistically expect it to do. In general, try to use the simplest and as less inputs as possible. Reduce and simplify. People sometimes say PA is better then indicators, but they are essentially the same. HH/LL, DB, DBP are in a sense also indicators.
Having said that and observing 'naked charts' for years, the indicators give by people above can be helpfull. Start with a simple TLB. Or plot a simple MA (like 20 or 30) and see what happens if it gets 'broken' or becomes flat.
"When does a girl in the bar will be an easy hit?" For starters, when she looks at you much in the eyes, playing with her hair and pointing her feet towards you it's probably a go. Next use the indicator "tell a bad joke", when she laughs, probability rises more.
A model is always an approximination of a realworld phenomenom. We try to find the best explanatory variables, only asuming causality.
It becomes more complex because definitions of "movement" and "ends" are variables by themself which need to be defined properly.
No model will give you an accuracy near 1.0, nor can you realistically expect it to do. In general, try to use the simplest and as less inputs as possible. Reduce and simplify. People sometimes say PA is better then indicators, but they are essentially the same. HH/LL, DB, DBP are in a sense also indicators.
Having said that and observing 'naked charts' for years, the indicators give by people above can be helpfull. Start with a simple TLB. Or plot a simple MA (like 20 or 30) and see what happens if it gets 'broken' or becomes flat.
"When does a girl in the bar will be an easy hit?" For starters, when she looks at you much in the eyes, playing with her hair and pointing her feet towards you it's probably a go. Next use the indicator "tell a bad joke", when she laughs, probability rises more.
Last edited:
I would. I'd be interested in anything you can show with even a 55% chance of winning.
There's a lot of advanced math that goes into some indicators. It's more than just drawing lines.