What are the odds of correctly predicting if the S&P will be higher or lower for 10 days in a row? 20 days in a row? What's the longest anyone has done? The ESPN "Beat the Streak" contest has me thinking along this line.
thats easy considering an adge of .04p, each prediction will come at .54p of success. hence, 10 days is a row would have the awesome probability of .54^10 = .0021%, thats odds of <b>1:394</b>. isnt that nice? 20 days in a row? LMAO do the math
Or....... http://www.elitetrader.com/vb/showthread.php?s=&threadid=138462&perpage=6&pagenumber=22 I generated this forecast at data point number 30(roughly) on X-axis. I've been updating the underlying price every week/two weeks since then. It's not exactly right for 10 out of 10 days straight, but you get the idea. Just scroll down a bit, and click on the first attachment on the page. Pizza out. -C.
The actual odds are unknowable since there are too many variables to account for. One has to make assumptions, such as prices moving in a perfectly Gaussian manner. Those assumptions are almost always wrong. Remember: An assumption is the root of all screwups.
It also depends on WHICH 10 days. There is a huge diff between "I will predict starting on date X" vs. "During a 4 year period, I had a winning streak of 10 days."
I would think the odds would be better than flipping tails ten times. The odds of flipping tails ten times in a row would be 0.5^10. As a worst case scenario, you have a 50/50 chance of predicting market direction (up or down). If you happen to have some sort of edge, this would bring your odds to better than 50/50, thus better than 0.5^10. Edit: Although unlikely, I suppose the odds of predicting market direction are worse than 50/50 considering the possibility of an exactly flat day.
So- working backwards, *if* you predict the market direction 10 days in a row, does that mean you have an edge, and can you calculate what that edge is?