I can't put my finger on how it's happening, but in one of my test algos, it seems like perhaps the universe is growing and growing until at some point after a year of simulation, I start getting stocks missing in my history() results, and things go downhill quickly from there (in my code). Anyone noticed anything like this?
I am trying to keep a historical log of all trades throughout history, in a DataFrame in my context, so that I can calculate un-leveraged theoretical performance in between rebalancings, for some equity-curve risk analysis/deleveraging, but having a column in a DataFrame shouldn't cause that sid to require a slot in the universe forever, should it?
Simon.
UPDATE: never mind, it was my bug :)