Now, when you say you have a system, you mean your own program? or a feature of IQFEED?
My own program built on top of IQFeed.
For the sake of clarity, here is a typical historical tick (for WFMI):
2006-07-31 16:00:00,57.54,400,2701862,57.5300,57.5400
In other words, at 16:00:00, a trade for 400 shares went off at 57.54. At the time, the bid was 57.53, the ask was 57.54. As you've read, DTN doesn't currently supply second resolutions on these ticks. But, the data comes as a correlated set.
In realtime, a packet arrives as a single unit which contains the bid, ask, trade, etc. I've found this data to be quite accurate, particularly for options. Again, it comes to you as a single unit, so you don't need to do any correlation.
Beyond that, you'd have to quiz DTN as to how they maintain "millisecond accuracy".
My own program built on top of IQFeed.
For the sake of clarity, here is a typical historical tick (for WFMI):
2006-07-31 16:00:00,57.54,400,2701862,57.5300,57.5400
In other words, at 16:00:00, a trade for 400 shares went off at 57.54. At the time, the bid was 57.53, the ask was 57.54. As you've read, DTN doesn't currently supply second resolutions on these ticks. But, the data comes as a correlated set.
In realtime, a packet arrives as a single unit which contains the bid, ask, trade, etc. I've found this data to be quite accurate, particularly for options. Again, it comes to you as a single unit, so you don't need to do any correlation.
Beyond that, you'd have to quiz DTN as to how they maintain "millisecond accuracy".
