this is some great work. +rep
I had done some of this not to long ago myself but was to lazy to do the spreadsheet as I was keeping track manually. And I have to agree with you Entopia. The odd thing is the variance that you show with the 50 clicks. I found that when doing 100 clicks that the % rate is much less "spiky", almost like it needed to make up for the first 50 when going into the last 50. It may indicate that the cycle is not a huge one as I had first thought but smaller.
I would also love to see what your results are for failure%. I believe that is where the higher skill/high luck period will show the most interesting data.