I am currently evaluating the Interactive Brokers Accumulate Distribute algo. I am trying to determine how much of the algo executes on IBs servers, and how much in TWS on the client machine. I am finding this information hard to get from the documentation or customer services. Please could someone clarify which one of these scenarios is true? (Paging IBsoft! You seem like a very helpful fellow). a) all order management is handled from TWS on the client machine, each individual component order of an AD order is submitted and maintained by TWS b) once an AD algo order is submitted in TWS, the individual component orders are submitted and maintained purely on IB's servers, but additional condition filters such as moving averages/position are controlled by the TWS client c) like (b) except that all additional condition filters and data streams required by them (eg: moving averages of other stocks) are also executed on the IB servers without any round trips to the TWS client (c) is clearly the preferred option as the performance advantages with having all order management and conditional filters running with minimal latency on IBs servers would be huge. I would be very grateful if someone could tell me with authority what the current situation is?
Last I checked the answer was a), it's all done from the client. This is obviously sub-optimal, and, IIRC, IB was going move it server-side.
It's a bit moot now, because as it turns out Accumulate Distribute is not fully supported by the IB API. It's a real shame because it's a very powerful algorithm, but only the very basic functions are accessible through the API, none of the additional conditions which I need. I hope that IB can provide a full implementation soon, especially as there is no mention in the documentation that the API Accumulate Distribute is not complete.