northam,
in the long run I think the superspike's main contribution or impact will be as a conversation starter.
It will be somewhat like the "Crash of '87". Which, by the way, was blamed at the time on automatic computer trading.
I was away from the computer when the superspike happened -- I tend to not be a day-trader. I probably would have been paralyzed, and not have known exactly what to do while trying to figure out when to do it and by how much, if I had been watching it live. I would have obviously missed the sweet spot in terms of buying something. And as luck had it's way, none of my stocks were significantly influenced by it anyway, at least in my present time perspective and point of view. I was surprised to hear on the TV that a few very well known big stocks went practically to zero. What I find interesting is that when I later brought up a screen full of sector charts, many of them hardly budged and some of them moved a lot more. One healthcare sector ETF I track actually dropped 27% that day, and yet even after the drop on Friday (yesterday) it is only down 4% from where it's high was on superspike day, and 0.7% of that 4% drop was on Friday and it actually traded up about 0.7% part of the time on Friday.
I'm a fisherman. Most of the time I'm a caster, in which case it's all about feeling the line. Once in a while I want to take things a little slower and easier so I use a bobber, in which case it's all about distinguishing between tugs on the line from the deep and the ocean waves or ripples in the pond. Either way, Superspike Day doesn't tug my line.
------
As I recall, you might be using Highs and Lows to calculate some of your averages for the purpose of making channels. If that is true then that day would have really nudged a line which is calculated from Low data, and it would take some time to filter that out -- lingering memories of a distortion. I'm still encouraged to filter out outlying data. The extreme outliers are sufficiently rare that it makes it virtually impossible to find enough of them to do a statistically significant backtest.
