E-Mini Futures Volume

[bassocat5]

I always saw differences between volume reported by the time and sales window and the volume study (I'm looking at ES #F and NQ #F). It is especially apparent on 1 minute charts. While trying to figure out why this occurs, I found that five 1 minute bars wouldn't add up to one 5 minute bar. I then found that if I reloaded the data from the server, the volume study numbers change and then match the T/S volumes (and the five 1 minutes now add up to the 5). Now I see that this condition only exists for bars that are produced "real time" and not for bars created using historical data. I noticed this months ago and reported it via email to eSignal and got a "looking into your problem" form letter in return. I have heard nothing since then and feel this is very important issue since it affects not only custom EFS studies but also basic studies like OBV and volume weighted MAs. I mentioned this in a Time and Sales forum but I think that my comments were overlooked since they were slightly off topic, so I created this new thread. Anyone have a comment or insight into this problem?

Rich

Rich,

Check out this link . Based on what you are describing, the discussions on this link may provide some help.

[bassocat5]

Thanks but that's not it

Thanks for the reply but that link isn't related to what I am talking about. It's not a timing issue, I can live with some volume slipping from one bar into another. In the case I am pointing out, large portions of volume just disappear entirely. In yesterday's NQ data, almost 10% of the total day's volume is missing. In one bad stretch, between 10:42 and 10:52am EST, almost 25% of total 1 minute bar volume was dropped. I have a fairly powerful setup, a 2.6 ghz CPU w/ 1/2 a gig of memory running win2k so I don't believe I am dropping this data. Besides, this type of volume dropout occurs even at 2am with only a few advance charts/efs studies running. Anyone else got a suggestion?

Rich,

I am assuming that the loss of volume numbers occurs in the efs. Based on this assumption, as part of your troubleshooting, what I would suggest is putting a debugPrintln statement in to print out the the current bars volume, the previous 5 bars volume and any volume summations you may want to make. Then run the code in Real time and what I believe you will see is that bars -1 through -5 will add up and be comparative to the 5 minute number. When in similar situations, this has been my experience. Something to try...

[bassocat5]

Debug Println

Steve,

I have a programming background and I have created an efs similar to what you describe. The volume does not end up adding up to the 5 minute. The volume DOES end up adding up to the the "normal" 1 minute volume study. Unfortunately, the 1 minute study numbers are missing trade volume (volume that magically shows up in the volume study when I do a reload of data from the server). There IS an EFS/ mini-futures problem that I know about where the EFS code is not running on every trade but on every change in transaction price(and/or every newbar). This condition might be related to my problem, but eSignal has not told me so. I was hoping that this thread would contribute to a temporary (if not permanent) solution for my problem. Once again, thanks for your input.

Rich