DislikedI tried reading the english translation for the russian books you cited, but I didn't get very far . I think my brain needs to be in a "special place" to understand Russian translated into english with google translate, hah hah. Maybe I need to drink vodka first. Anyway, I will try again sometime, but the gist of what I read reminded me a bit of Darkstar's book.Ignored
I understand that use of parametrics and transformation will help us have a better view of the data, but that's out indeed of my league (or perhaps I just require further practical exposure). Great to see that you weighing in this into the thread.
Disliked...Turns out that both leg length and percent retrace are log-normal (skewed to the high side)....
....Something interesting going on with leg length there, as its has a slight bimodal distribution going on...Ignored
IMO one would have to rerun the swing statistics again after optimizing the Swing Type criteria. To check if there is improvement or degraded odds, against the original baseline statistics.
Disliked...Here is the regression of retrace level against leg length. Note that there is a significant slope (p<0.0001). Generally, as leg length increases, percent retrace decreases...Ignored
I found the results (as attached) for all possible combination types; 18 types - URDR/URDJ/URDT, UJDR/UJDJ/UJDT, UTDR/UTDJ/UTDT, DRUR/DRUJ/DRUT, DJUR/DJUJ/DJUT, DTUR/DTUJ/DTUT, with 2 lengths being investigated- extreme and wick lengths, making a total of 36 sets of data. It turned out, at least in my (limited) judgement, to be relatively the same to the singular swing type results.
Attached File(s)
Results.csv
68 KB
|
266 downloads