Back in November, @fulltimegeek, @stellabelle, and some others (@aggroed?) delegated moderate stakes to dozens of minnows, of no more than ~5k to each.

What I noted and @abh12345 provided data for, is that all of the delegates decreased selfvotes, spread their votes to more minnows, and engaged more in comments.

I consider this a huge success, although the @haejin flagwars ended the experiment when those delegations were withdrawn so they could flag away at him. This is the great thing about such delegations: the stake, and potential for capital gains when the retention rate improves, totally inures to the delegator.

Prior to that I recall @ned delegating ~500k to a dozen or so folks, and all of them increased selfvoting, sold votes, etc... Except @surpassinggoogle, who is a saint.

I note the difference is in the amount. It seems people can handle ~5k responsibly, but not ~500k. I suspect ~50k would also be too large to expect people to treat with honor.

~5k just doesn't offer enough vigorish to folks to rampantly selfvote. It's not enough to fund a votebot, or to sell votes with, and so that would be my suggestion.

The massive stakes funding the votebots is voting now, and taking rewards from the pool, so parceling it out in moderate delegations won't change that at all, but will totally change distribution, and I believe that is the fundamental factor affecting retention.

My last information on retention is that ~10% of user remain after a year. Which means 90% give up before a year is up. This promises a short future for Steemit, and the votebots are making the problem worse.

I am plum tickled to be actually in conversation with @haejin and @transisto now regarding this very matter. Can't wait to see how that conversation goes! =)

Thanks!