Final Comments.
- Even if I'm not using real money yet, I take the decisions to buy and sell in real time. (if I'm in front of the computer when the signal is given --otherwise I use a time and price biased to the "worst" case)
I've been running this program only for 6 days, so any conclusions are likely to be very premature, but I'll throw a couple anyway:
- I'm using 1% and 3% as triggers for going short, and long. The actual data values are usually far more extreme than that. I was tempted to change them to (let's say) -15% and +15% but it doesn't seem to affect the results much. (and the "one trade per day only" rule takes care of the whipsaws better, it seems)
- SPY seems to have less volatility than QQQ, but ADVN seems to be more sensitive than ADVNQ. ADVN has triggered changes earlier than ADVNQ in all cases. (so far, of course) It would be interesting to trade QQQ using ADVN as trigger.
I'll keep testing this system as it looks that there might be something to it.
If there's interest I'll keep on posting results as they become available.
News 
Market Data 
Discover 
