Yeah, thanks, EB. Something that seemed odd struck my attention. Numbers are fascinating. I was trying to reconcile dwnlds by raw vote totals. The disparity is huge, even where an author breaks a series into 3 to N books--all having different raw vote totals for each succeeding book of the complete series. If you watch carefully, there are even some strange 'wobbles' of successive dwnlds and raw totals of votes. And, I would expect that plotting voting frequency or raw dnlds totals by increasing book count would give me 2 reverse J-curves, with maybe an anomalous point for the last book. So, getting a U-curve along with a reverse J, OR 2 U-curves is perplexing. Definitely, it's counterintuitive. But, thanks again, you gave me what I needed to have a shot at piecing my puzzle together.
Hey, for what it's worth, quick and dirty analysis shows you wrote 26 stories of file size 50 KB and up.[Anything smaller to me is a crap shoot, though I loved your shorty, 'Always a Marine' rated 9.23.] The range was 66 KB to 1437 KB, with mean 386.31 KB. Your average story got you 1315 votes, ratings range 7.13 to 8.55, mean rating 7.9435 (not too shabby, congrats). Going too far based upon linear regression/correlational data is foolhardy. But, I would bet with 98% statistical confidence that if you began writing stories of file size 1500 KB to 2000 KB, you would begin to rack up quite a few 9.00+ ratings and I'm 99.999% statistically confident that you would find more readers casting votes for your stories. Mirabile dictu.