I have like 80-ish shift, still building up my bag. I voted twice, for 66 total. Basically the top people that profit-share.
I'm basically going to make 0 until I build my bag bigger, right?
And does it bother anyone else that with DPoS, most people don't know or care who they vote for, as long as they get paid? and the delegates themselves all just have tons of coins themselves, and keep making more, and then use them to vote for themselves perpetually? I love the speed and power of DPoS projects but the actual voting mechanism is terrible from a decentralization and governance standpoint.
Yea, basically I estimate you will need to hold for 6 months to get the two shift you spent on voting back. If you doubled you holdings to ~160 Shift it would take 3 months etc.. I would vote for at least the top 99 if you get some more for more payouts.
And yes, I agree DPos has a number of flaws, as once a delegate get's in the top 101 it is hard to get them removed. There are some interesting ideas to fix this posted on the Lisk (another DPos coin) reddit thread the other day:
https://www.reddit.com/r/Lisk/comments/80iw0y/im_starting_to_think_ark_fanboysgirls_are_right/ Going to adapt and recycle some old material here:
Delegated Proof of Stake (DPoS), like every other consensus system, has pros and cons. One of the larger cons, which is what Lisk is currently dealing with, isn't really innate to DPoS so much as to Lisk's particular implementation of DPoS.
The basic principle of DPoS is simple enough: Instead of everyone staking their own coins, they instead add the 'weight' of their coins together into pools that are run by specific 'delegates'. Only delegates can create new blocks on the network, and they are rewarded for this service. As a holder, you use your coins to vote for (a) delegate(s) of your choice. Anyone can attempt to become a delegate, but ideally you, as a holder, should vote for a delegate that offers more benefits than the others (see caveat below). These benefits usually take the form of redistributed rewards; i.e., the delegate promises to return X% of their rewards back to voters. But there could be other considerations: maybe a delegate also proposes to use some % of their rewards for advertising that will benefit the project, or maybe they propose to use some % to build and maintain a super-awesome server that will improve the network. So, in an idealized DPoS system, rewards are the main reason people vote for one delegate over another, but they are not the only reason.
In Shift's case, only the 101 delegates with the most votes at any given time receive rewards. As soon as they fall out of the top 101, they receive nothing--and the people voting for them receive nothing also. So
if DPoS is implemented carefully there is an incentive for delegates to always improve their service to the network in order to stay in favour with voters (again, see caveat below).
Caveat:
Of course, one problem here is that if someone (or an alliance) has enough Shift, they can theoretically vote for themselves with enough votes that they are basically locked into the top 101. Such delegates would be largely immune to the normal forces of the delegate market. More on that later.
But the real problem arises, as @McSunshine has noted, when one vote is actually more than one vote. I would argue that this is, essentially, a poor implementation of DPoS; it makes it much easier to abuse the system if desired. Perhaps as importantly, it also--to take @dhouse's point--devalues the significance of a token-holder's vote, since the incentive is to vote for as many delegates as possible to maximize return, rather than for the few that are actually doing the most for the network. This is the problem we see in Lisk, and it's a potential issue in Shift as well. Ark and NEO, on the other hand, solve this by only allowing one unit of the asset to equal a total of one vote, even if the vote is split among several delegates. I.e., you can vote 100% for one delegate, or 10% for each of 10 delegates, but the total can never exceed 100% and your reward is also non-summing (it either comes 100% from the single delegate you voted for, or 10% from each of the delegates you voted for, or whatever, but you don't make more for voting for more delegates). I'm not sure exactly what system Shift will implement with the new consensus mechanism, but I imagine it will address this very issue.
And furthermore:
Now, on the matter of removing delegates from the top 101, this '1 vote = 1 vote' business is actually pretty important. If 1 vote = 1 vote, then it is harder to accomplish self- or cartel-voting, AND voters are incentivized to vote only for the delegate(s) that provide true value--there is no deterministic financial benefit to voting for multiple delegates. Yes, someone with enough of the asset could still vote themselves in, but no one else has any incentive to vote for them and reward their misbehaviour. So the effects are minimized: it doesn't directly impact voters' returns or the overall health of the network (even 'bad' delegates still have to perform services for the network if they wish to earn rewards for themselves, after all). And the cost of doing this becomes increasingly prohibitive, so the odds of more than a couple of 'bad' delegates being able to hang around are reduced.
In any event, assuming Shift implements an improved voting system, the other issues we are seeing with delegates (low/dishonest payouts, etc.) will tend to sort themselves out over time as Shift grows, becomes better known, and its price increases. As that happens, the delegate 'marketplace' will become more competitive and delegates will need to work harder to remain in the top 101.
For example, look at Ark, which has a similar voting/delegate structure but is more established in the market. Most Ark delegates now share 90%+, and many perform additional services for the community as part of their proposal, as suggested above. As Ark has grown, the percentage being shared by delegates has steadily increased, and it will likely continue to increase. Needless to say, a delegate who doesn't provide value won't last very long (again, see caveat below).
Or, to use a less-closely-related example, NEO, which uses dBFT (Delegated Byzantine Fault Tolerance) and will ultimately have around 100 consensus nodes (delegates, more or less, except they set network fees rather than minting), officially anticipates that running a node will provide almost no profit, because voters won't vote for nodes that charge much/anything (voters also don't receive financial rewards for voting). Anyway, my point is that once a project is well-established, the benefit of being a delegate in a delegated consensus system is much more about network health and indirect benefits to your holdings than it is about direct profit. However, the evolution from selfishness (where many Shift delegates are currently) to altruism (where Ark is approaching) is slow, and depends on a well-conceived voting system.
So, overall DPoS is really quite robust,
if implemented with a carefully-considered voting system. Anyway, more info here if technical details are desired:
https://steemit.com/dpos/@dantheman/dpos-consensus-algorithm-this-missing-white-paper 
