This topic contains 257 replies, has 18 voices, and was last updated by 
-
Posts
-
@pipatronic: Your output quantity is really amazing ;) Thanks!
-
This reply was modified 2 years, 6 months ago by
Anti.
Well I hope it is helpful to all, the biggest problem I have is lack of time, there is some really superb information on this forum if you go back and dig deep.
Regards
Pip
skype : pipatronic
Yes, maybe. However, regarding the true bar idea I had the same problems as you. One will just know the size of last true bar when the next has formed …
Awesome work Sir, thank you. As I know red/yellow dots lag one/two bar(s). For using this concept in our trading we should know on which dots we should pay attention and which dots we can ignore, because @gg53 told us that we never mind what happened between this dots. Any ideas penguins? I think we are in right path.
Also @gg53 told that :
In down-swing, some candles LOW is lower than the previous candles LOW in that swing, some are not.
But…
A candle LOW is always lower than all the previous candles HIGH, and this is the invariant for a down-swing.
If a candle breaks this rule, it starts an up-swing. It’s the basic nature of any bar-chart, so you don’t need parameters or indicators. Just “naked trading” (don’t do in the winter, you’ll catch cold)
Maybe this concept helps us to find exact dots?
Hmm … look at the examples in pipatronics video. The question is: How would you trade this. Before I did own experiments I’ve thought that we get trading signals whenever one true bar engulfs the preceding one. However, as you may see from the examples in the video, if we see the engulfing in most cases the trend is over. Ok, we could trade it this way: We wait until an unfinished true bar engulfs the previous bar and enter then. But this seems to be similar logic than trading band brakes using Donchian channels.
Maybe GGs concept of the river bed is more valuable. My idea for its us is that we compare the line for smalles lookback period to each higher lookback period lines. Whenever the median of the smaller channel is in the upper or lower third of the higher channel, we see a strengthen up/down trend regarding those TFs.
Hi @anti, can you show your thoughts in picture-examples?
Hi @Joakim,
I’ve tried to summarize it in the below image.
To show these yellow (smallest), red (longer) and blue (longest lookback period) channels, I’ve used Simplex’s RiverBed_v1_1_post indicator which has been posted at this forum. The yellow channel now should show the high/low range of last 5 candles, the red of the last 15 candles and the blue of the last 60 candles. Whenever the yellow channel is narrow and in the upper part of the red channel, we should se a short-term up trend, if it is at the lower third of the red channel, then we have a down trend. The same is true for the position of the red within the blue channel.
Jo do you know about clear method candles ?
skype : pipatronic
Anti, good post and excellant example
skype : pipatronic
Thank you @anti step by step, I hope I’ll understand how “to catch fish”
@pipatronic I have candles-indicator which was created by @takhang (sorry if I spell wrong). Do you mean this indicator?Jo, I am sure it will be the same but for anyone else here is a lnk
https://www.mql5.com/en/code/10623
skype : pipatronic
just did a comparison between red dot yellow dot (as in video ) vs clear method candles – more predictive vs lagging ?
skype : pipatronic
Hmmm…very interesting pip
Now put zig zag on (channels water ?), just for illustration as zig zag very laggy
skype : pipatronic
water volume over river bank
skype : pipatronic
@pipatronic: Is that the system you’re working on recently? If so, I’d either try to reconstruct and recode GGs OBV divergence indicator or I’d ask GG if you could get a mql4 copy of it. If you don’t have the mq4 file you could one day be confronted with a no longer working indicator …
@pipatronic: Is that the system you’re working on recently? If so, I’d either try to reconstruct and recode GGs OBV divergence indicator or I’d ask GG if you could get a mql4 copy of it. If you don’t have the mq4 file you could one day be confronted with a no longer working indicator …
I think GG53 doesn’t share his indicator’s open source code. Many times on FF and here he was asked, but he didn’t share, and I can understand him. But big big thanks to him for his work, and also he share with us without and tax
So we should work what we have 
for don’t lose importand indicators, I stopped auto-update my metatrader 
Anti, no I do not use OBV anymore but I do like to try and expand the brain (of what little I have) and play with old/new ideas
And yes always a worry with MT4 updates if you do not have the source (personally I like tomato)
Pip
skype : pipatronic
On the current time frame suppose we get a true DOWN candle close and we now have a true UP candle forming, if this candle comes with above average volume and the higher time frame trend is UP the results are reasonably predictable. I do not believe there is a strategy here but what is does give us is a better overall picture of the chart in front of us
Pip
skype : pipatronic
The function is quite simple (and logical):
Volume > Volume[1] and Range < Range[1] and (Low < Low[1] or High > High[1])
The analogy: more water (Volume) are coming into the “River” but encounter a narrow creek (smaller Range), so they must deflect and find new, different path. Roll up your sleeves. G.
After studying the volume concepts laid down by GG I found this one again. After adapting and filtering with higher TF signals I obtain sometimes quite accurate turning points. However, what me really wonders is this: We can often see turning points with a high volume and a high range. How does it fit in the water flow analogy? Why should the river change its direction after the flood leaves a big river bed?
Additionally I found a problematic behavior of tick volume in MT4 (and probably with other platforms, too) … a tick doesn’t neccessarily mean a chang in 0.1 pips. Sometimes it can mean a change by 0.2, 0.3, … pips. However, the volume only counts these changes as 1 tick. Thus, you can see very efficient candles where the prize path (abs(open-close)+2*wicks) is bigger than the volume itself. Is this a problem? Any recommendations?
After studying the volume concepts laid down by GG I found this one again. After adapting and filtering with higher TF signals I obtain sometimes quite accurate turning points. However, what me really wonders is this: We can often see turning points with a high volume and a high range. How does it fit in the water flow analogy? Why should the river change its direction after the flood leaves a big river bed?
Hi Anti, i have followed this thread for long time, but just as a good reader … :) I try to give my opinion on this : GG said about Volume(x) compare to Volume(x+1), Range(x) compare to Range(x+1), High(x) compare to High(x+1) / Low(x) compare to Low(x+1), Close(x) compare to previous candle (x+1) to get the Higher probability of turning point. But as you said, some times big volume and big/Small range does not change the direction. May be this idea can not be taken as main signal, but it is for filtering of other idea, for example in addition to Similarity method (just my Self opinion …) ?
Additionally I found a problematic behavior of tick volume in MT4 (and probably with other platforms, too) … a tick doesn’t neccessarily mean a chang in 0.1 pips. Sometimes it can mean a change by 0.2, 0.3, … pips. However, the volume only counts these changes as 1 tick. Thus, you can see very efficient candles where the prize path (abs(open-close)+2*wicks) is bigger than the volume itself. Is this a problem? Any recommendations?
This is interesting …. still trying to understand it … will post again later, if i find some thing …
Yes, surely. It may only give a hint. Thanks!
I’ve thought it would be a good idea to investigate the relationship between prize movements and volume. Thus, I’ve used my way of calculating the prize path, calculated it for 603,896 M1 candles and looked at the relationship between prize path and volume:
As you can see, there seems to be a linear relation between volume and the prize path. The red regression line explains more than 64 % of the total variation in volume.
However, the dots above the blue line indicate bars with more than average volume in comparison to their prize path, and thus more activity per pip. I think that these candles can be seen as exhaustion candles. It’s quite interesting to see how most of these exhaustion candles have relatively high volume and quite small prize paths (“ranges”). This is exactly what @gg53 mentioned in his formula (see 3 posts above this post).
Nevertheless, there are some low activity bars, too. They appear in the lower right part of the plot. I think they can be explained with an erratic prize movement in one single direction during the lifetime of the corresponding bar. With “erratic” I mean that prize developed with jumps (e.g. a jump from 0 ticks to 0.3 ticks, etc.).
If we now use the formula for the regression line and subtract the expected (average) volume for each bar from the real tick volume, we may get a measure for the excess activity … I’ll attach my indicator with the hope that we can continue our work on this topic. However, since we are most probably using different brokers, the formula I obtained for the relationship between prize path and volume may be different than the optimal solution for your broker. Additionally, the formula only applies for M1 charts.
Nice work!
I haven’t been around so frequently these days, so please forgive me if my thoughts on your latest post do not consider topics that have already been discussed.
When developing an algorithm for trading purpose I’m always heading for two major goals to simplify practical usage:
- If possible, the algo should be symbol independent.
- If possible, the algo should be timeframe independent.
You posted the following core algorithm in your source file:
buff=Volume-(12.12+148200*(2*(High-Low)-MathAbs(Open-Close)));
Generally, I think it would be interesting to learn about the constants (12.12; 148200) when looking at specific symbols to cover that part. There will be systematic variations.
The price part 2*(High-Low)-MathAbs(Open-Close) obviously is not timeframe independent, its value will grow by a factor of
sqrt(2) * (tf1 / tf2)sqrt(tf1 / tf2). Hence your constants will vary systematically over different timeframes.Considering the modification 2*(High-Low) / MathAbs(Open-Close) you might overcome this to simplify further usage. What should remain is non-systematic variations (noise) over different timeframes. Any thoughts?
s.
-
This reply was modified 1 year, 3 months ago by
simplex. Reason: Nonsense formula corrected!
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)
-
This reply was modified 2 years, 6 months ago by
You must be logged in to reply to this topic.
