I Want Volume (spot Vs Futures)

[Kiwi 16]

The issue with the volume correlation is this:

 

Lets just imagine the 95% figure was true.

 

 

But lets say that when you looked at the data you discovered the correlation was true when the price was running. When it retraced and did an abc or abcde then, at critical times there was a low correlation. So, if you were (say) using the movement noise measure (cause it aint a tick either) to assist your decision to enter ... it would be useless.

 

 

So a 95% correlation could be useless if the correlation was true at some times but not others - and the other time was the important time for you.

 

 

But I think that without evidence its just another example of bs made into a factoid by quoting untraceable "statistics." And I am not implying that the person quoting it here is the bs'er ... with such statistics the bs may lie a long way up the train of "knowledge" being passed from mind to mind.

[Flojomojo 10]

This is interesting because I have just been reading hundreds of threads saying the opposite. What studies have been done that empirically show positive correlation between tic volume and the spot or even the futures market?

 

I did quite some reading myself and you are right that threads are contradicting. But so far I have not found a single post that actually threw in a well worked out study about this topic...which essentially leaves it as just plain gossip!

Snackly, watch out that you don't only process information that your mind wants to see.

[Flojomojo 10]

Thirdly Tic volume. Knowing how it's calculated, I can't personally take it seriously. Even at times when there's correlation there most definitely isn't a sustainable cause / effect line (I bet non of these studies undertaken included granger causality modelling or anything like that).

 

I did a very quick and incredibly dirty granger causality test and the only hint I found was that tick volume is granger causal to the range of a bar. (am not persuing this further for now) Nevertheless I got a bit sceptical about the Granger method as it essentially only checks if lagged values of a variable help forecast another. Also the method gets a lot more complicated if regressions are not linear.

[GammaJammer 10]

Bit confused - what did you test it against? futures volume?

 

GJ

[Flojomojo 10]

GJ...it was really quick and dirty only deserving a place in the gossip corner! I used this guys toolbox. It gave me a relationship between the range of 15 minute EUR/USD bars and the corresponding tick volume for the bars. No futures, just plain data that I recorded.

 

Flojo

[SwingTradinJay 10]

stack your charts FX futures + FX spot.

 

Use the volume on futures as a rough estimate, it's fine.

[Kiwi 16]

Thats an interesting statement STJ.

 

Years back I used to trade Eurofx futures on a 5 minute and 1 minute chart. Knowing the potential that volume has in Index futures and particularly in stocks I spent a lot of time qualifying volume price relationships in eurusd futures.

 

In the end I came up with one good correlation but it was an indicator of manipulation. I'm pretty sure that what was happening was that the futures were being manipulated so that profits could be gained in the (much bigger) cash market as the two markets have difficulty at certain times of the day deciding which is the tail and which is the dog.

 

Based on that (isolated really) observation I'd ask how much light volume in the futures will shed on the stochastic darkness of the cash market. I must have another look at futures and see if the intervening 3-4 years has changed my perception of price volume relationships so that I can see something different.

 

Whatever, if you apply volume price relationships to anything qualify it very carefully looking for both type I and II errors.

[alleyb 10]

The currency futures have a layer of bids and offers stacked which are arbitraged back to cash via the Block Trade and the EFP (Exchange Futures for Physical) facilities. Many retail accounts don't seem to understand this factor and believe that the volume is the be all end all. By way of explanation I will give an example. the market is say 1.2888-1.2889 in a 16 x 4 market. someone sells 22 lots all filled (due to a submarine bid) at 1.2888. the next quote is 1.2887 - 1.2889 in a 4 x 9 market. 1 lot trades down into the bid at 1.2889 and seemingly the market then shoots up to 1.2893-1.2894. The volume watcher would have seen volume come into the left hand side selling but the price action is now higher. this reflects the cash market where say 1.2893 is given and 1.2899 is then paid. What I find useful is not the volume per see but the pattern from the tic volume histogram that in MP terms I can associate with certain types of behavior. I agree with previous tail and dog comments. The only time knowing volume in FX is going to give you an edge is when you are a spot jockey and see the customer flow particular the orders coming from SAFE the Chinese vehicle or the Red Man (who is in fact a lady) the chief FX trader for the Russian Central Bank that is probably the worlds largest customers right now.

[kazos 10]

you can use IB with Amibroker.

or e-signal + amibroker.

 

These are nice

[kazos 10]

stack your charts FX futures + FX spot.

 

Use the volume on futures as a rough estimate, it's fine.

 

I would say that spikes in volume in futures fx can be used to estimate probable turning points in fx spot.

[trader273 10]

Don't know if this has been mentioned before, but this is something that was in the latest SFO magazine.

 

http://www.sfomag.com/article.aspx?ID=1315&issueID=c

 

It basically says that a 1 period ATR is pretty close to a volume calculation.

[BlowFish 10]

Don't know if this has been mentioned before, but this is something that was in the latest SFO magazine.

 

http://www.sfomag.com/article.aspx?ID=1315&issueID=c

 

It basically says that a 1 period ATR is pretty close to a volume calculation.

 

So in essence range and volume are closely correlated (which most know empirically). When they are not, that is the time interesting things happen. When they are correlated and at statistical extremes interesting things happen too.

[jf2000 10]

On the CMEFX web site they claim that $100B of FX futures contracts are traded daily through the CME.