This morning I was going over the DOW charts and notice something that I hadn't before. What makes it very interesting in that the pattern is 12 for 12. I did not have enough data to take it back futher so it is not statistically viable, but I would not ignore it, since it is 12 or 12.
Before I explain it, I also want to mention that even though it is 12 for 12 it does not mean 100% probability. Too often correlation and mathematical probability is used to mean the same thing, which is not correct. The reason I mention this is that too often I hear 1 for 1 or 2 for 2 or 3 for 3, etc correlation and many assume that is high probability, when words like "WILL", "MUST", "DEFINITE" and "ABSOLUTE" are used. And frankly, those who use low levels of correlation to conclude high probability, may not have a firm/good understanding of probability.
OK, since we got that out of the way, here goes the 12 for 12 correlation. It is a subsector of my short-term analysis:
1) 12 of 12 - produce a minimum down move of 150 points. In one case it was 600+ points(one day intraday)
2) 11 of 12 - closed near the intraday lows(within 25%)
3) 10 of 12 - the following day was either flat or up
4) 6 of 12 - were or within 100 points of a short-term bottom.
5) 6 of 12 - produced rallies of a minimum of 400 DOW points. 4 of these rallies started the next day, and the other 2 started within 3 days.
6) 5 of 12 - had further selling of 200-400 DOW points before a short-term bounce
7) It did identify the strong rallys that started in MAR and JULY.
The one problem is that I cannot be sure if FRIDAY was the THAT DAY or MONDAY will be THAT DAY.
Wont have the time to do that thorough of an analysis on the NAZ/NDX but with a quick glance, the results appear to be similar, but not as many. |
