Has anyone found that IB's smart routing adds a significant amount of latency? Just doing simple testing today and I'm finding that I'm missing a higher number of trades using their smart router than just direct routing (I can't confidently say that this isn't attributed to randomness though). Though my sample size is way too small, so thought I'd ask the forum.
If I may ask, how exactly are you doing the comparison between order fill rates? Doing something like sending 1000 orders and comparing how many get filled for smart routing vs direct routing?
I went through ~100 trades today running the code in parallel in two separate accounts. Only difference was smart vs directed. It's also worth pointing out that the products I'm trading will always be routed to the same exchange by the smart router.
I ran some similar tests a while back (also on small sample size). On average, SMART routing added just over 200ms to my execution times when compared to going direct. That was with "Seek Price Improvement" checked...somewhere on the bottom of my to-do list is to re-run the test with that turned off. Would be nice if IB offered some statistics on how that option performs in terms of fills and/or latency.