HFT efficiencies

 Following the discussion at todays software I have put down some observations.
These raise more questions than the solve, but hopefully they can hep discussions.
see attached file