So, I had another thought. And call me crazy, but I'm going to throw it out there. What about a POS coin with demurrage.

The "problem" with POS coins: "inflation." Everyone wants high stake, but wants to control inflation. So we set a max coin cap, or we decrease the stake percentage over time to some eventual minimal inflation.

However, the attraction of the POS coin is that you can earn high %. But when that goes away the excitement is gone many seek greener pastures and newer coins with newer features.

That's where demurrage comes in, or negative stake (nstake). A coin could always offer high positive stake (pstake) because it will never suffer from inflation due to its nstake. The theory is that the nstake has a flag that is set when a number of coins is reached and increases to slow the pstake. Once a second coin point is reach nstake is increased above pstake to bring the coins down. It is a hysteresis effect to avoid market shock. There would be an ebb and flow around a soft coin cap chosen by the dev team.

Such a coin would have to have two stake ages computed: nstake age and pstake age. Nstake age would have to be faster than pstake age to hit every wallet - something like 6 hours. While pstake age would be longer: a day, 3 days, whatever.

In theory the coins would congregate to the clients that spend the most time staking. In other words the network would pay the nodes that do the most work for it by securing it.

Example coin:
POW coins: 0.5 mil
Minimum Coins: 1 mil
Maximum Coins: 2 mil
PStake: 100% @ 7 days
NStakeMax = 115%
NStake= ncoins/maxcoins * 100%
if(ncoins>0.7*max coins){
do{nstake= nstakemax}
while (ncoins>mincoins)
}

Assuming EVERYONE stakes - not likely, but a possibility
Year 0: 0.5 mil coins; nstake = 25% ; effective stake w/ compounding = 111%
Year 1: 1.05 mil coins; nstake = 53% ; effective stake w/ compounding = 60%
Year 2: 1.68 mil coins; nstake = 84% ; effective stake w/ compounding = 17%
Year 3: 1.44 mil coins; nstake = 115% ; effective stake w/ compounding = -15%
Year 4: 1.24 mil coins; nstake = 115% ; effective stake w/ compounding = -15%
Year 5: 1.06 mil coins; nstake = 115% ; effective stake w/ compounding = -15%
Year 6: 1.05 mil coins; nstake = 53% ; effective stake w/ compounding = 60%

That is only one example. You could change the nstake curve to increase the number of years a coin stakes with positive interest. And you could also increase the NStakeMax percentage to decrease the number of years of contraction - of course that may be more of a shock.  

Lots of variables to play with. If nothing else, it could be a fun experiment.