Thought I'd start a thread discussing common problems that aren't so obvious in the beginning. It would be nice to hear from others, so people here will not have to spend a bunch of time on a problem only to realize that others have already solved it.
Here are two quick ones on using last trade
Issue #1: This one should be fairly obvious for anyone who has traded before and it's related to time. You run the risk of operating off of stale information. Take this extreme example, suppose the last trade happened an hour ago and it's quoting at values extremely far from the last traded value. Basically, it's not a realistic representation of what you can actually execute at.
Issue #2: Things here will get slightly quantitative and I would love to hear other people's feedback on this one. The issue I see here, is that if you examine trades at high granularity you'll end up with false signals. Here I've plotted the first 1000 trades executed of SPY on the morning of 10/12/2010 on BATS. The value looks like it follows some kind of autoregressive process, however that's just an artifact of trades going through between the bid and ask. It's quite misleading if you're not careful...
That's all I have time for, for now. Feel free to add to the discussion.
-A
