Add: When looking at the top 50 in the rich list
https://chainz.cryptoid.info/nobl/#!rich , it looks like just under half are minting on a regular basis, and really wasting their coin age in the process since those are 7 million NOBL transaction sizes and up (the largest is minting a 143.5 million transaction size and is probably missing out on 90% of what could be minted if that were broken up into at least 70 smaller transactions).
I suspect the biggest single addresses are owned by exchanges. Possibly Cryptsy? Not sure if they want to spend too much time tinkering with all the wallets they have to achieve best minting results.
However you are right they lose most of possible rewards. So the incentive is there. PoS II works the way it was designed
Note to all people keeping their money at exchanges - it is fine if you are actively trading. However if you are not concerned with short term swings and want to keep your noblecoins long term it is better to setup your personal wallet and split your holdings into several transactions of sizes between 100 000 - 250 000 coins. I am happy exchanges stake coins since this helps securing the network, however it is unlikely they are going to do it in really efficient way. Common good is to achieve highest possible difficulty. And this can be achieved when people spend little bit of resources by staking when their computer is on and little bit of time adjusting their transaction sizes.
And if you're getting somewhere close to 8% (with optimized minting), that's another 800,000 NOBL a year for every 10 million you have, just for leaving your wallet open! (But only if that 10 million is broken into smaller amounts! That 10 million as one chunk might actually only get you 400,000 or even less! That's losing 50% or more. And imagine if that "chunk" is 20 or 30 million or more and you're only making 200,000 NOBL a year! That's when the complaints will start streaming in, and also when we start posting links to the original posts on the subject . . . that having been said, it might not be a bad idea to put together a definitive doc on this as soon as we can.)