- This topic has 227 replies, 14 voices, and was last updated 10 years, 6 months ago by
.
- Posts
To FULLY test your results – use it with MT4 brokers that provide assets like indexes – such as DAX where the Volume is “true” Volume. G.
Hello G, I do not understand why the DAX volume would be any different to say the EURUSD, is it because the DAX volume is not based on tick data ? Regards pip
DAX is traded on close market, i.e. in a StockExchange like environment where there is TRUE volume. Same for Commodities.
In FX there is no such thing.
G.
Well, but it was not what I expected. 1. I haven’t expected that there is such a relationship. 2. If I would have expected it, I’ve thought it would be a positive correlation: up volume entering the market -> market rises. down volume entering market -> market falls. So the explation seems to be that market makers spread their money while entering market but went all out of the market at the same time!
Why don’t you look at the ForexGT_VolumeOsc posts?
High Volume extending ABOVE market “normal” signals = Drop in price. The opposite is true for extension BELOW “normal”.
Here comes the Elephants to “Pee”… (vertical dotted lines).
G.
That’s why I’m analyzing EXTREMES and not “normal” market behaviour.
Do you really expect the Market_Movers to enter in the middle of slowly trending market?
Or to INFLUENCE the market direction change, at certain price levels area?
The second one is what I’m after, and for that the Volume (or Ticks or Activity) IS a leading indicator.
ForexGT_VolumeOsc is proving that, on both M1 & H4.
G.
Great job – thanks for that:
For each M1 candle of 100001 EURUSD bars I calculated then the volume difference = up volume – down volume for each candle.
I assume your analysis is based on classical OHLC of M1 and not on tick data, and the logic of up / down volume is the same as discussed here – correct?
In my opinion the results described in your last paragraph are the most interesting ones. In your definition of
range5 = close-opendo you mean the signed value or the absolute onerange5 = abs(close-open)?Any questions, suggestions, comments?
I wonder how results would be if we left the boundaries of a static sampling rate given by M1 and analyzed tick data instead. Assuming roughly 100 ticks per minute that would lead to 10,000,000 samples. Now how to consolidate these data?
I think it would make sense to spot ‘unusual conditions’: extreme levels of ticks per second TPS (indicating no price direction), or points per tick PPT (indicating a price direction and strength of movement). Your findings posted above suggest (imo) that an intrabar volume / price dynamic could possibly look like this:
- extreme move in a certain direction associated with extreme values of TPS and PPT (intrabar)
- price move reaches a possible point of reversal: TPS will probably stay at a high level, while PPT (short term average) approaches zero
- price reverses towards its former level: TPS will probably stay at a high level, while PPT (short term average) crosses zero and turns in the opposite direction
Safely spotting the transition from phase 2 to 3 should give a nice entry. If we filter entry direction to be following only general trend direction (by currency strength or whatever) we should be able to find good trend / swing entries. Low delay in order execution and low slippage would be essential here.
Also essential would be to define what an ‘extreme move’ (see point 1 above) is. Take ‘volcanic eruptions’ after ECB announcements or similar events into account? Possibly better focus on ‘smaller’ extreme conditions and exploit the waves in calm seas.
Now the above mentioned N = 10,000,000 samples only make sense offline to verify or falsify my assumptions based on a significant amount of data (in R). If it’s possible to verify rules similar to my above model these rules should be working on significantly smaller samples like N << 1,000, or better N < 100. These rules would have to be coded in MQL.
Maybe the idea could be tested on larger timeframes first, because this data is at hand: replace above M1 or M5 timeframe by H1 or H4 and replace above ticks by M1 OHLC data. It’s certainly questionable to draw conclusions from the analysis of H1/H4 down to M1/M5 and ticks, but possibly the basic dynamics above is similar in both cases.
This is still coarse. Any thoughts?
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)
Why don’t you look at the ForexGT_VolumeOsc posts? High Volume extending ABOVE market “normal” signals = Drop in price. The opposite is true for extension BELOW “normal”. Here comes the Elephants to “Pee”… (vertical dotted lines). G.
1. I don’t trust voodoo charms (you comment on my ex4 file).
2. I don’t know what you exactly did to obtain values in your ForexGT_VolumeOsc. It could be any calculation, averaging, etc. Thus, just from an image without further info I can’t conclude anything.
That’s why I’m analyzing EXTREMES and not “normal” market behaviour.
Do you really expect the Market_Movers to enter in the middle of slowly trending market?
Or to INFLUENCE the market direction change, at certain price levels area?
I really have some problems with this EXTREMES as you have to define this conditions before you perform your analysis. Thus, the results may be biased. On the other hand I want to trade not only during rare market extremes.
Further investigations proved that the negative correlation is due to high volumes bars. Thus, I would suggest that those bars itself indicate your extreme markets. But nevertheless, the negative correlation persists.
Thank you for the real volume info G, I will have a look a bit further down the rabbit hole
Pip
skype : pipatronic
For each M1 candle of 100001 EURUSD bars I calculated then the volume difference = up volume – down volume for each candle.
I assume your analysis is based on classical OHLC of M1 and not on tick data, and the logic of up / down volume is the same as discussed here – correct? In my opinion the results described in your last paragraph are the most interesting ones. In your definition of
range5 = close-opendo you mean the signed value or the absolute onerange5 = abs(close-open)?It was not based on tick data and I derived the estimates of up/down volume from post you cited. I meant the pure difference (
open - close) and not the absolute value of this difference (abs(close - open)=|close - open|) (cf. second image of post where you can see that prize change = open – close) can obtain positive and negative values).Any questions, suggestions, comments?
I wonder how results would be if we left the boundaries of a static sampling rate given by M1 and analyzed tick data instead. Assuming roughly 100 ticks per minute that would lead to 10,000,000 samples. Now how to consolidate these data? I think it would make sense to spot ‘unusual conditions’: extreme levels of ticks per second TPS (indicating no price direction), or points per tick PPT (indicating a price direction and strength of movement). Your findings posted above suggest (imo) that an intrabar volume / price dynamic could possibly look like this:
- extreme move in a certain direction associated with extreme values of TPS and PPT (intrabar)
- price move reaches a possible point of reversal: TPS will probably stay at a high level, while PPT (short term average) approaches zero
- price reverses towards its former level: TPS will probably stay at a high level, while PPT (short term average) crosses zero and turns in the opposite direction
Safely spotting the transition from phase 2 to 3 should give a nice entry. If we filter entry direction to be following only general trend direction (by currency strength or whatever) we should be able to find good trend / swing entries. Low delay in order execution and low slippage would be essential here. Also essential would be to define what an ‘extreme move’ (see point 1 above) is. Take ‘volcanic eruptions’ after ECB announcements or similar events into account? Possibly better focus on ‘smaller’ extreme conditions and exploit the waves in calm seas. Now the above mentioned N = 10,000,000 samples only make sense offline to verify or falsify my assumptions based on a significant amount of data (in R). If it’s possible to verify rules similar to my above model these rules should be working on significantly smaller samples like N << 1,000, or better N < 100. These rules would have to be coded in MQL. Maybe the idea could be tested on larger timeframes first, because this data is at hand: replace above M1 or M5 timeframe by H1 or H4 and replace above ticks by M1 OHLC data. It’s certainly questionable to draw conclusions from the analysis of H1/H4 down to M1/M5 and ticks, but possibly the basic dynamics above is similar in both cases. This is still coarse. Any thoughts? s.
Great comment. I have to think about it. Regarding your three points I absolutely agree. What you’ve mentioned in point 2 describes exactly my idea of equilibrium. Here I also see a parallel to stopping volume and climax bars of the classical VSA approach.
Why don’t you look at the ForexGT_VolumeOsc posts? High Volume extending ABOVE market “normal” signals = Drop in price. The opposite is true for extension BELOW “normal”. Here comes the Elephants to “Pee”… (vertical dotted lines). G.
1. I don’t trust voodoo charms (you comment on my ex4 file). 2. I don’t know what you exactly did to obtain values in your ForexGT_VolumeOsc. It could be any calculation, averaging, etc. Thus, just from an image without further info I can’t conclude anything.
That’s why I’m analyzing EXTREMES and not “normal” market behaviour. Do you really expect the Market_Movers to enter in the middle of slowly trending market? Or to INFLUENCE the market direction change, at certain price levels area?
I really have some problems with this EXTREMES as you have to define this conditions before you perform your analysis. Thus, the results may be biased. On the other hand I want to trade not only during rare market extremes. Further investigations proved that the negative correlation is due to high volumes bars. Thus, I would suggest that those bars itself indicate your extreme markets. But nevertheless, the negative correlation persists.
I fail to see any big bars, except one, in my ForexGT_VolumeOsc.
I fail to understand your “rare” market opportunities counting.
You are right about Voodoo charms analysis. We need to wait for the source code to be released.
G.
I fail to understand your “rare” market opportunities counting.
Rare market opportunities: How have you defined your extreme market conditions you’ve analyzed? Only news? If so – news only happened relatively rare. And my aim is to trade not only during those times.
You are right about Voodoo charms analysis. We need to wait for the source code to be released. G.
Yes. Hope someone will have a good innovative idea.
I fail to understand your “rare” market opportunities counting.
Rare market opportunities: How have you defined your extreme market conditions you’ve analyzed? Only news? If so – news only happened relatively rare. And my aim is to trade not only during those times.
You are right about Voodoo charms analysis. We need to wait for the source code to be released. G.
Yes. Hope someone will have a good innovative idea.
For analysis I gathered 12 yrs of historical data, analyzing 56 major events per month (8 Currencies, 7 major pairs, one major event per month – at least) – and got some conclusions. That was the EXTREME analysis.
Once conclusion was solid – I went to code. From that point I can identify and trade not only news/announcement but almost any major move of the Elephants. You can see that from the posted charts with various of my Volume indicators.
As to ForexGT_VolumeOsc: You got the numbers almost right on your analysis post, now you have to define a “range”, limit that range to 80% (+40% & -40%) of the market.
Any above/below that is an Extreme volume movement. The task of making it an oscillator is trivial, as simplex said.
G.
“Events”, in the above post, are events that are known in advance, such as Unemployment, Interest rate, GDP, etc. – per country/currency. Those are scheduled and known (incl. Date/Time) on monthly basis.
G.
As to ForexGT_VolumeOsc: You got the numbers almost right on your analysis post, now you have to define a “range”, limit that range to 80% (+40% & -40%) of the market. Any above/below that is an Extreme volume movement.
Most probably a very useful statement. “You got the numbers almost right” may mean that I have to extend sample size?!
Thanks, G.
As to ForexGT_VolumeOsc: You got the numbers almost right on your analysis post, now you have to define a “range”, limit that range to 80% (+40% & -40%) of the market. Any above/below that is an Extreme volume movement.
Most probably a very useful statement. “You got the numbers almost right” may mean that I have to extend sample size?! Thanks, G.
5-7 bar range is within my calculations. But, as previously explained, I dont use “bars” as a measure, but Time or Pips.
G.
Oh, another question, G. I assume you’re averaging/summing up the up/down ticks. Yes? Would you reveal over which time/how many ticks?
Hold your horses !!!
tankhang just told me (via Skype) that he found a treasure, here:
and sent me the MQ4 source (attached).
This is very, very similiar to mine (difference in bar/timing calculation) .
Enjoy, and thanks tanghang !!
G.
That’s “all” Miss a headbanging emoticon … Thought your’s is based on ticks, not on OHLC bars.
Mine is based on ticks, but he achieved same results – so who cares….
I’m going to stick to my version – but now you can elaborate and enhance this one.
G.
The question is how to filter some bad signals …
Mine is based on ticks, but he achieved same results – so who cares….
Think this similarity proves that my former conclusions regarding info in ticks/bars were correct.
Hold your horses !!! tankhang just told me (via Skype) that he found a treasure
Less than 2k of code … best things in life seem to follow the KISS rule!
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)
The question is how to filter some bad signals …
Mine is based on ticks, but he achieved same results – so who cares….
Think this similarity proves that my former conclusions regarding info in ticks/bars were correct.
I think Ticks is more sensitive.
Differences (breakouts) are marked with vertical lines. I adjusted my params to roughly equal the default “14” in VZO.
In any case, you can see that EXTREME Volume (or “Elephants Activity”) indicates direction change.
G.
And one more thing to mention: in the source code posted by G. a link to a conceptual article by Waleed Aly Khalil was included:
http://www.ifta.org/public/files/journal/d_ifta_journal_09.pdf
I already downloaded that IFTA Journal more than a year ago, but only regarded a John Ehlers article worth reading in detail. Not a good choice, right?
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)
And one more thing to mention: in the source code posted by G. a link to a conceptual article by Waleed Aly Khalil was included: http://www.ifta.org/public/files/journal/d_ifta_journal_09.pdf I already downloaded that IFTA Journal more than a year ago, but only regarded a John Ehlers article worth reading in detail. Not a good choice, right? s.
I “drink” each morning any article about Volume in FX, Stocks, Commodities, Futures, etc. – instead of coffee…
G.
I “drink” each morning any article about Volume in FX, Stocks, Commodities, Futures, etc. – instead of coffee… G.
Java Coffee…….!!!
in big volume everyday… Yay… !! lol.. 
Have a nice week end Brothers…
MTH
Intuition, Experiences and Common sense..
http://www.binaryoptionsedge.com/The question is how to filter some bad signals …
Trade signals that are with the trend, i.e. in Down trend you trade only UP breakouts, and opposite in Down Trend.
That will reduce almost 90% of bad trades.
G.
in big volume everyday… Yay… !! lol.. 
- You must be logged in to reply to this topic.