- This topic has 257 replies, 18 voices, and was last updated 8 years, 8 months ago by
.
- Posts
You are correct by trying to figure this out on >= M30 TF. There is no point in developing such system on smaller TF’s. Trying to find turning point based on PRICE leads to std. Fractals, which are lagging by at least 2 bars. Furthermore, the entry will be on the close of the second bar. Can we do something better than that? or more reliable, eliminating false entries? G.
Some more questions G :
1) If i look at your picture posted on post #9197 , the TrueBar is NOT one candle (60 minutes bar in TF H1) where we can find the red/yellow dot of OBV. But it contains some H1 bars. In this case, shall we take only one bar where the red/yellow dot printed, or we take the last bar of that zone (pink/blue zone) , or …. ?
2) If we calculate the sonar tools on TF D1 (like we do in H1 candle), do we need 7 channels (6 of them are inside the D1 candle itself). Do we calculate it at 5 minutes before the D1 bar close ? And how to calculate the H4 to M1 channels ? How many candles of each TF shall we take, and when its begin/end ?
Edit :
3) About finding turning point, shall we use fractals ? It will good better signal, but we are late 2 candles ….4) About your Gadi_NormalizeVolume. The white “doji” is actually a good hint of reversal, if its volume is bigger than average ? Or we just look at red (bear) and blue (bull) volume candle ?
Thanks in advance
Hi smallcat, according to GG53’s explanation in the Killing Zone thread at FF (post 1086), the white bar means: Volume is not moving the price in any significant direction (“Doji”). Vlan
Hi Vlan,
What is your rule in getting data from GadiVolume ?
Edit: Sorry … i forget that you gave the source code. I will look at it. Thanks——–
Below is my modified code, i do not know if i am right in getting data from GadiVolume, and not sure about the rule for up/dn swing.
I add MT4 standard fractal, so the signal will late 2 candles, but to make sure that there is a fractal, i call it after 3 bars (it will lagging 3 bars).
This is a quick and dirty code, so it is still in development, not ready to use ….
On image we can see your G_TurningPoints result too.Hi Vlan, What is your rule in getting data from GadiVolume ? Below is my modified code, i do not know if i am right in getting data from GadiVolume, and not sure about the rule for up/dn swing. I add MT4 standard fractal, so the signal will late 2 candles, but to make sure that there is a fractal, i call it after 3 bars (it will lagging 3 bars). This is a quick and dirty code, so it is still in development, not ready to use …. On image we can see your G_TurningPoints result too.
Hi Vlan,
Please look at the image. May be combining your code with my code will give better signal ???
Hi Vlan, Please look at the image. May be combining your code with my code will give better signal ???
The images look good. But, if I’m not mistaken, it seems like you had the ‘showSignalAtFractal’ set to true so that means all your arrows should in fact be 2 bars to the right.
To add to your puzzle pieces list: Currency strength. GG said that if implemented right, it has predictive capability. In an uptrend it would indicate currency weakening before actual turning point. But I don’t have such yet.
V.
Hi G, in MT4 price id given per Tick. At the time the candle close, it can be No tick. And if we want to check fractal at 2 previous candle, it can be we missed it. Does it not better to check fractal at open/new candle? Yes this will be lag 3 candles instead of 2, but i think it is safer.
To catch peak/valley is not easy, and we can be cheated. I just think of using channel like TMA / Keltner as filter, but TMA is repaint. Still trying to get channel/bands that similar to TMA, but not repaint. Keltner is nrp, but i like an almost balance bands like TMA or COG.
The images look good. But, if I’m not mistaken, it seems like you had the ‘showSignalAtFractal’ set to true so that means all your arrows should in fact be 2 bars to the right. To add to your puzzle pieces list: Currency strength. GG said that if implemented right, it has predictive capability. In an uptrend it would indicate currency weakening before actual turning point. But I don’t have such yet. V.
Yes, you are right. It is lag, because of fractal. Thanks for Currency Strength information. I have Gadi Currency Strength, will look at it. Or may be simplex’s compact currency strength?
Thanks
Vlan,
I slightly modified your code (see attached) – this relates to gg53’s last point that current bar has to: “Closes below the high (in uptrend) or below the low (in downtrend)”.
To be honest this is the most unclear statement – my understanding is that current bar close has to be lower (higher) that previous bar high (low).
I believe that this gives more accurate turning points…
Changed line 115 and 120 in your code.
edit: revised code is attached two message below
Vlan, I slightly modified your code (see attached) – this relates to gg53’s last point that current bar has to: “Closes below the high (in uptrend) or below the low (in downtrend)”. To be honest this is the most unclear statement – my understanding is that current bar close has to be lower (higher) that previous bar high (low). I believe that this gives more accurate turning points… Changed line 115 and 120 in your code.
My statement doesn’t relate to previous bar.
The CLOSE must be lower than the HIGH, suggesting that there is not enough “power” to continue in that direction and price fell from its High.
The opposite is true for DownTrend.
G.
My statement doesn’t relate to previous bar. The CLOSE must be lower than the HIGH, suggesting that there is not enough “power” to continue in that direction and price fell from its High. The opposite is true for DownTrend. G.
so basically, it must have a wick. When checking my FractalArrow database it shows that more than 90% of bars have their Close lower/higher than their High/Low. This was for the M5 TF but I think other TFs will have similar ratio.
Should the size of the wick be also evaluated?
V.
My statement doesn’t relate to previous bar. The CLOSE must be lower than the HIGH, suggesting that there is not enough “power” to continue in that direction and price fell from its High. The opposite is true for DownTrend. G.
so basically, it must have a wick. When checking my FractalArrow database it shows that more than 90% of bars have their Close lower/higher than their High/Low. This was for the M5 TF but I think other TFs will have similar ratio. Should the size of the wick be also evaluated? V.
Basically no. It’s just depends when the retracement started before bar close.
G.
Just to clarify: Volume greater than previous bar volume Range less than previous bar range A new high is made (in an uptrend) or a new low (in a downtrend) Closes below the high (in uptrend) or below the low (in downtrend) G.
Is there a mistake? Shouldn’t it be above the low?
I got this function:
if(Volume > Volume && MathAbs(Close-Open) < MathAbs(Close-Open) && Low < Low && Low < Close)
{
upArrow = Low – Distance;
}
if(Volume > Volume && MathAbs(Close-Open) < MathAbs(Close-Open) && High > High && High > Close)
{
dnArrow = High + Distance;
}
When I looked to the charts, I saw that there are candles, which have both: up and down arrows, which shouldn’t be in my opinion.
If current candle range is less than previous candle, and if current candle has higher high and lower low, than my code draws both arrows. One of them could be filtered with a trend indicator.
Hi G,
Please take a look at the image below. I do not understand sonar tools +1 , the time range & begin time of M30, M15 and M5 rectangle are same.
The rectangle of M15 is a little bigger than M30, but M5 is much much bigger. This is just example only, not based on the OBV red/yellow dot yet.
But the result is different. I am afraid that this is a bug …
Edit: Other question: Why the high of M15 is bigger than the high of H1 (although its height has small difference) ?? This is also can be seen at your picture.
I think ALL of the lower TF should be INSIDE the H1 candle, right ?Thanks
Hi G, Please take a look at the image below. I do not understand sonar tools +1 , the time range & begin time of M30, M15 and M5 rectangle are same. The rectangle of M15 is a little bigger than M30, but M5 is much much bigger. This is just example only, not based on the OBV red/yellow dot yet. But the result is different. I am afraid that this is a bug …
Edit: Other question: Why the high of M15 is bigger than the high of H1 (although its height has small difference) ?? This is also can be seen at your picture. I think ALL of the lower TF should be INSIDE the H1 candle, right ? ThanksHi G, i think i found it. I use kprsa’s RS code on getting min/max. I forget that in RS it counts candle to the left (from i to i+x). It should be to the right (from i to i-x), or negatif shift. Will try to repair it soon ….
Thanks
Just to clarify: Volume greater than previous bar volume Range less than previous bar range A new high is made (in an uptrend) or a new low (in a downtrend) Closes below the high (in uptrend) or below the low (in downtrend) G.
Is there a mistake? Shouldn’t it be above the low?
Yes, Sorry about that mistake. It’s impossible to close BELOW the Low…
G.
Hi Simplex, Pls see attached Multi-Donchian channel indicator I modified ages ago, I tried putting all the donchian channels with different TF settings but it looks like a mess, maybe got to do with my modification too, but I did it to show when price starts moving one direction. Pic shows M15. M30, H1, H4, D1 http://penguintraders.com/forums/topic/true-barscandles-and-market-sentiment/#post-8914
My “problem” with Donchian channels:
Donchian channels are based of Highest/Lowest of xx bars ago. Let’s assume xx=10.
If we use it on M5: what is the significance of 50 minutes ago ?!?! why not 52? why not 45?
If we use it on M15: what is the significance of 150 minutes ago ?!?! why not 170? why not 149?
etc…
etc…
So basically it’s an artificial, man made, arbitrary channel, doesn’t really reflect TRUE market (or “river” contour) condition.
G.
Hi G, i think i found it. I use kprsa’s RS code on getting min/max. I forget that in RS it counts candle to the left (from i to i+x). It should be to the right (from i to i-x), or negatif shift. Will try to repair it soon …. Thanks
This is the picture of the modified one …
My “problem” with Donchian channels: Donchian channels are based of Highest/Lowest of xx bars ago. Let’s assume xx=10. If we use it on M5: what is the significance of 50 minutes ago ?!?! why not 52? why not 45? If we use it on M15: what is the significance of 150 minutes ago ?!?! why not 170? why not 149? etc… etc… So basically it’s an artificial, man made, arbitrary channel, doesn’t really reflect TRUE market (or “river” contour) condition. G.
Hi gg,
Are u trying to mean just display an extended High & Low horizontal lines like M5, M15, M30, H1, H4 ,D1, W1, MN on the chart. Without actually complicating matters, namely just draw every higher TF high low horizontal lines on the chart then we see how it goes from there. Like Auto Higher TF High & Low Horizontal Extended Lines? eg. So it draws like every M5, M15, M30, H1 high low lines on H4 charts. (we skip the M1 high low or not?)
Best Regards,
LearnAlways
I only know I know nothing
Skype: learnalways@outlook.comMy “problem” with Donchian channels: Donchian channels are based of Highest/Lowest of xx bars ago. Let’s assume xx=10. If we use it on M5: what is the significance of 50 minutes ago ?!?! why not 52? why not 45? If we use it on M15: what is the significance of 150 minutes ago ?!?! why not 170? why not 149? etc… etc… So basically it’s an artificial, man made, arbitrary channel, doesn’t really reflect TRUE market (or “river” contour) condition. G.
Hi gg, Are u trying to mean just display an extended High & Low horizontal lines like M5, M15, M30, H1, H4 ,D1, W1, MN on the chart. Without actually complicating matters, namely just draw every higher TF high low horizontal lines on the chart then we see how it goes from there. Like Auto Higher TF High & Low Horizontal Extended Lines? eg. So it draws like every M5, M15, M30, H1 high low lines on H4 charts. (we skip the M1 high low or not?) Best Regards, LearnAlways
No.
G.
Hi G.
I’d like to come back to your ‘riverbed’. Did you have something like this in mind when setting up this metaphor?
Let’s consider NOW (see red circle) it’s 5:08. So I marked the ranges of the latest M5, M15, M30, and M60 candles.
Latest riverbed contour in red.
s.
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)
Hi G. I’d like to come back to your ‘riverbed’. Did you have something like this in mind when setting up this metaphor? Let’s consider NOW (see red circle) it’s 5:08. So I marked the ranges of the latest M5, M15, M30, and M60 candles. Latest riverbed contour in red. s.
Close, but not quite… M30 should extend past the already closed H1, M15 should extend to the already closed M30, etc…
By your screenshot:
When M5 broke the M15 it was time to enter a quick counter-trend LONG trade.
The overall sentiment is DOWN.
With std. Volume, the market will return to the M30/M15 range.
G.
Close, but not quite… M30 should extend past the already closed H1, M15 should extend to the already closed M30, etc… By your screenshot: When M5 broke the M15 it was time to enter a quick counter-trend LONG trade. The overall sentiment is DOWN. With std. Volume, the market will return to the M30/M15 range. G.
Hi G, so my tools is wrong?
Hi G!
Ok, let’s give it a try:
Close, but not quite… M30 should extend past the already closed H1, M15 should extend to the already closed M30, etc…
I needed a break from my integrated CIX, so played a bit with the conceptual ideas you provided here. I tried to come close to the sketches in your previous posts (here and here).
It’s not yet Christmas, but outside my window the first snow of the season is falling. So I decided to provide a toy for everybody around and make the indicator dynamic: double click the vertical gray line and drag it around: the channels will be redrawn according to this offset time given by the line.
For the coding newbies: this is a bit advanced, but maybe you’re able to figure out how it works. If not: any questions welcome.
For your concept, G.:
In previous posts, you suggested channels referring to 5, 15, 30, and 60 minutes high / low.
We also had some talk about splitting time windows according to geometric series. So I combined your riverbed model with a geometric series of 2, 4, 8, 16, 32, and 64 bars (default, can be changed by parameters). Maybe 2 bars is a bit too short and could easily be dropped.
What I’m not sure about is how much to extend the channels into the future, as referred to the offset bar. As an easy start, I just made them all the maximum channel depth (64 bars by default). IMO, this is a bit too much, but before fiddling around I just present my 1st prototype (attached).
Any feedback much appreciated.
simplex
Ahh yes: those points on bars 0 and 1 are just because I like to have something to double click on instead of calling the indicator by Ctrl-I.
EDIT 2015-11-26: updated version here.
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)
Thank you, @simplex for your indicator.
I’m sure I haven’t fully understand the idea of the price river flow. As far as I understand the idea is to identify price regions at which price can act with low force. That’s comprehensible. Thus, we could trade this fact by waiting that high volume forces price to leave its river bed and predict the revisitation. But what I can’t understand is why the revisitation is highly probable. For instance, if we use the river analogy, we can see that after a flood a river uses a new river bed and never returns to its old bed. Why should it be different in the markets?
Edit: Other question: Why the high of M15 is bigger than the high of H1 (although its height has small difference) ?? This is also can be seen at your picture. I think ALL of the lower TF should be INSIDE the H1 candle, right ? Thanks
- You must be logged in to reply to this topic.