basic price action of S/R failures. but how often does it break S/R just enough to trigger retail entries only to go right back against you? Lots. How far should you allow it to move against you i.e. is it just a little noise or are you dead wrong on your entry? should you wait for the first pullback after BO to enter? are you sure you are in a trending market or just trapped in a channel where fading direction at S/R is better? should you just jump into every S/R failure and if it fails reverse your position? do you have that discipline? are you trading an instrument that allows that? lots of stuff to iron out IMO but that's trading IMO.
what are the signs of reversal vs pullback?
- Thread starter mute9003
- Start date
Here is a better example of following higher lows, or lower highs.
The green line assumes you've been long at where the line starts.
You could make a rule that sets a minimum percentage move you need to have a playable turn.
If you were to test this out, it would profit greatly, even when you are always in.
A problem arises when you start to include the high-low swings of each bar. So you're going to have more reversals that you would only be able to see if you also track bars on a lower time frame. You'll have to develop more rules to handle each contingency. That will lower performance. But if you try to follow the trend like this, good things can happen.
You can call the turns anything you want. You could call them a "breakout". Whatever it is called, you are looking for a break of the higher lows pattern, or a break of the lower highs pattern. Those are the only two patterns (which are really the same thing mirrored) that tell you there is a "trend". If that trend is no longer maintaining, then the break may be considered a "reversal".
Given this, what would we consider to be a "pullback"? We would have to come up with some other rule. But one thing for sure, if you can't/don't make money this way, you're not going to even know what is a pullback, or whether you could even make money on a pullback, let alone better money to get in "early". You have to make money this way before you can define "pullback" and/or "early". You can't counter-trend trade, or "fade" anything, until you have some idea of what the actual trend actually is.
Nor do you always have to be in the market. You could mine the statistics data to find filters, or to increase exposure at certain times.
Aloha Good1,
Thank you for the image. It sounds like you're referencing percent retrace. Is the previous with-trend segment what you're comparing the countertrend segment to?
In other words, if it's less than a xx% retrace it's essentially noise?
.
Hi mute,
My suggestion is analogous to what Expiated showed in his chart earlier, and I can add a few specifics.
Pullbacks to a well-tested medium-period MA in an uptrend or downtrend, followed by a strong bull or bear bar, increases the likelihood of a pullback happening. Medium-period EMAs and WMAs can work well for this kind of trading, but SMAs are not as efficient in my experience.
Yes, you could make a rule that says you'll ignore noise that is less than xx% retrace. This could be a rule for what you may end up calling a "pullback", but I am mainly focused on "turnarounds". Was the retrace significant enough to use it for a turnaround scenario.
Anyway you never really know what works, or what works best till it is put through a lot of data to generate statistics/probabilities/performance.
No nature has taken care of that for us with the Fibonacci sequence. It's nature's Quant.
What you're saying makes no sense.
Take a retracement of 60% for example.
Maybe 60% is somewhere on the Fibinnaci scale?
Doesn't matter. Whether it is or isn't, you still have to run a mountain of data through 60% specifically to have any idea of probabilities.
Don't like the probabilities? Ok test, 50%, 40%...any percent you want. You have to run each percent through a mountain of data.
If probabilities spike at certain Fibo points then maybe there is something to it. Otherwise, each percent will give you probabilities very near the percentage that is neighbor. Just pick the hottest probability out of 100 tests.
60% is not a Fibonacci level. It's .618... But it's close enough. You don't have to do anything....it's already done for you. Save yourself all that time and binge watch Cobra Kai instead.
Wait, so you're saying you're betting my big money on untested levels of Fibonacci? Are you also saying that .618 is the most profitable? On what basis?60% is not a Fibonacci level. It's .618... But it's close enough. You don't have to do anything....it's already done for you. Save yourself all that time and binge watch Cobra Kai instead.