Found the mistake - error in bands calc on my part. Will be more prudent going forward - disregard earlier. My apologies. Andrew Moe --- Andrew Moe <moebot@xxxxxxxxxxx> wrote: > Found something that looks far too good to be true. Must > have wired something wrong. Would one of the crunchers > please have a look. > > Calculate 50 period bollinger bands for the vix at 1 std > dev. > > If today's vix > high bollinger band buy S&P at close, > exit > next close. Done in S&P points, cash index since '94. > Vix data is NEW method from cboe.com. I refit the dates > to > match the s&p dates. > > Record: 38-3 (93%) > Avg Win: 11.77 > Avg Loss: -2.56 > Avg Move: 10.72 > Std Dev: 11.02 > T Stat: 6.23 > > The reason I checked this was from recent discussions on > the list mentioning that it is roughly a 30 day predictor > of volatility. I figured that a moving average at double > that length should give some information. > > If correct, this seems to fit the theory that a sharp > spike > up in vix is highly bullish for stocks. > > Best, > Andrew Moe > > Data sets available for both vix & sp. > > > > > > > > >