Taking a break from MPF work. I can't see how to get it working right. I do have a few other ideas of how to link 2 priceframes / timeframes together though.
Resuming classification of states for sequencing;
http://www.forexfactory.com/showthre...25#post7690925
http://www.forexfactory.com/showthre...47#post7691947
Got the corrected statistics up for both lower and higher priceframes;
Its interesting to note that the distribution stays pretty much the same with acceptable variance, even thou the priceframes are different. I do want to test if the distribution stays the same across 'seasons' e.g. 1999, 2000, 2001...etc, to individually test meaningful samples of sufficient size. I have my doubts actually.. e.g. trending season may have skewed T3R2T1.
Will revert with the results soon. If the distribution remains consistent, this means past odds are a reasonable representation of future odds. Yes, I know we are 'taught' that the past has no bearing for the future. However, it must be relative to context; some things do have weight, some things don't. I never like general sweeping statements without context.
http://i62.tinypic.com/2qk43gg.png
Building an engine to read the odds; doing to write a new one as the old one isn't so great, but the general idea is the same. What a coincidence that the lower and higher priceframes are in the current same states?
http://i57.tinypic.com/2nb9bfl.png
Resuming classification of states for sequencing;
http://www.forexfactory.com/showthre...25#post7690925
http://www.forexfactory.com/showthre...47#post7691947
Got the corrected statistics up for both lower and higher priceframes;
Its interesting to note that the distribution stays pretty much the same with acceptable variance, even thou the priceframes are different. I do want to test if the distribution stays the same across 'seasons' e.g. 1999, 2000, 2001...etc, to individually test meaningful samples of sufficient size. I have my doubts actually.. e.g. trending season may have skewed T3R2T1.
Will revert with the results soon. If the distribution remains consistent, this means past odds are a reasonable representation of future odds. Yes, I know we are 'taught' that the past has no bearing for the future. However, it must be relative to context; some things do have weight, some things don't. I never like general sweeping statements without context.
http://i62.tinypic.com/2qk43gg.png
Building an engine to read the odds; doing to write a new one as the old one isn't so great, but the general idea is the same. What a coincidence that the lower and higher priceframes are in the current same states?
http://i57.tinypic.com/2nb9bfl.png