Ok, these are legitimate questions which I will try to help with, although I'm just an old retired rocket scientist. We'll probably have to get one of the devs in here to cover the fine points when one of them comes up for air.
But here's a good recent post that can provide a starting point:
http://bytemaster.bitshares.org/bitshares/2015/01/04/Delegated-Proof-of-Stake-vs-Proof-of-Work/The essence behind our thinking is that all systems tend toward centralization, and it is best to recognize this fact and specifically engineer controls into the system to manage it. Without such controls, Bitcoin has centralized into about six or so entities who decide what software is "official". People who join pools don't really provide any decentralization. They just lend (delegate) their processing power to somebody running the actual signing software. Pretty centralized and nothing really anyone can do about it.
Bytemaster recognized that providing a mechanism where stakeholders explicitly delegate their signing authority to authorized signers would allow token owners to reclaim that control - rather than leaving it to those who have unilaterally claimed the role by virtue of their ability to assemble a large centralized mining operation. If one of those Bitcoin central miners start's misbehaving, well there's not much the community can do about it. If the two biggest collude, well, game over.
Why 101 for DPOS? Well, its less centralized than six! But anyway, going from 1 to 2 delegates doubles your security. Going from 100 to 101 adds less than 1%. Yet each delegate added increases the cost linearly. So its simply a case of diminishing returns on how much you want to pay for security. Going beyond 101 simply doesn't buy you much. We chose it as the sweet spot, but 50 or 150 would still be reasonable.
Another factor he recognized is that there is really very little damage that a rogue delegate can do. Producing an invalid block simply gets her fired. Recognizing that everything a delegate does is inspectable and detectable, the shareholders can quickly detect any misbehaviors and summarily fire the offender. Again, stakeholders control their own destiny.
You do raise a valid temporary concern. What happens if the community of stakeholders decides to vest trust in fewer than 101 unique delegates? What breaks?
Well, 101 is just an arbitrary number that is much, much greater than six. It's entirely possible for the community of stakeholders to decide that, for a period of time, they want to hire the services of an individual or small business that brings so much value to the ecosystem that its worth assigning more than one delegate slot to that individual. Think of your favorite big name in the crypto industry. What if the block chain could hire her at the expense of multiple slots? Would the shareholders want to do that? Should they have the authority to do that?
Of course. BitShares is a company! If the owning stakeholders think that will make them more profitable and grow faster, why can't a company decide to do that?
In the short term, while shares are worth pennies, the hard limits on how much a single delegate can earn sometimes requires combining several delegate revenue streams to come up with a full time salary for a highly sought-after individual. By the time those shares reach 5 cents apiece, there should be no further need to merge streams and things will settle down to 101 independent delegates.
In fact, by the time BitShares reaches Bitcoin's market cap, each delegate will be one of 101 small businesses, selected by the stakeholders, each using a revenue stream of several million dollars
apiece to grow the ecosystem. Powerful stuff to look forward to!
So your concern is really a harmless startup transient. Security is already far better than it was for Bitcoin at this stage in its development.
As for Bytemaster having a lot of voting influence? Not really. He controls a small percent of the votes and usually waits for the community to show a strong liking for a candidate before deciding how he will vote. He can easily be overridden (in ten seconds) by any small coalition of voters, should they get the urge to do so. Most of the time they are content to rely on his judgement, but that is under constant vocal review every day at bitsharestalk.org.
Remember, BitShares is a company, not a currency. It is a unmanned, decentralized company that produces and trades interest-paying "smart currencies" as its product. So judge it by whether it is a good idea and implementation for a company, not a currency. Then you can get past all the accepted rules that (may or may not) apply to future currencies and see clearly what the investment opportunity truly is here.
If you want to get really excited about where this is going, read Bytemaster's latest article about BitShares as the future of exchanges. http://bytemaster.bitshares.org/article/2015/01/05/The-Future-of-Crypto-Currency-Exchanges/
And if you'd like to get your questions answered in person, Bytemaster hosts a public meetup using Mumble every Friday at 10 or 11 Eastern Standard Time. It's not unusual to have 50 people attending from most time zones around the world. Join us and ask your questions there. Then evaluate whether his live answers to a world-wide audience measure up to your expectations of a serious industry leader.