This question has gone round and round in my head a number of times. I think it will be easier to frame the problem if we word the question as "What is the optimal number of masternodes?" and derive the masternode capital requirement from that.
One way to look at this is in terms of democratic representation. If there are 4000 masternodes and 4000 Dash users, on average everyone gets a vote (every proposal is near enough a referendum). However if there are 4000 masternodes and 40,000,000 Dash users (let's say defined by making a Dash transaction once per month), every masternode is representing 10,000 users. 400,000,000 Dash users is one vote per 100,000 individuals.
The larger the value of the masternode collateral, the more incentive a masternode operator has to study and understand the issues required in designing a cryptocurrency. An interesting point comes when masternode income is enough to sustain its operator. Let's take 50k USD/y as an example, as it is possible to live in most of Europe (even London) with that amount of income. Then at 10% annual masternode return, a DASH price of 500 USD/DASH and a 1000 DASH collateral would give pretty much every masternode operator a very livable income.
Would it be a good balance to have 4000 full-time masternode operators holding 500k USD collateral voting on proposals in the name of 10,000 Dash users each? What if the Dash price reaches 5000 USD and there are still only 10,000 users per masternode, are the interests of these 4000 masternode operators (now holding 5M USD each) still sufficiently aligned with everyday users, or are they going to start to believe they are more important, and try to turn Dash into a "settlement layer", like Blockstream has done with Bitcoin?
But say masternode operators start acting like banking oligarchs (anyone who has used the UK banking system may be able to appreciate the it appears to be run
for the banks, not
by the banks) – if they can be convinced that a collateral reduction is in their interest, how low is low enough? If you could slash the annual income of a masternode operator per node from 500k USD to 5k USD, you may kick out the rich and self-entitled oligarchs (or you may not, they may just continue to run more nodes), but you also substantially reduce the return of a node and the value of holding one. Will the new owners after this round of selloffs and purchases be as incentivised to dedicate their time to making good Dash decisions as the ones who made a comfortable living off it before?
I don't have any answers to these questions, a least not yet. However, I am fairly sure they are all valid questions, ie all the cause and effect I am assuming by asking them really does play out in the real world. (If anything looks suspect, just say!)
There is existing research into modelling and simulating both democratic and economic systems, and I'm just starting to look at it now. My hope is that it will be possible to create a theory for the effect the number of masternodes has on the Dash network. Not an exact, predictive theory like relativity or electromagnetism, but one that indicates the potential effects of changing the masternode collateral from 1000 to 100 to 10 DASH, given estimates of other parameters. Sometimes surprisingly simple models can give insights into sociological situations, for example
Schelling's Segregation Model – sorry for the alliteration, I couldn't avoid it
.
A useful comparison is the blocksize debate in Bitcoin, because there isn't a theory for the blocksize limit, which indicates the effects of a given blocksize limit given a certain number of transactions, nodes, high and low speed network links (etc), Blockstream has been able to turn the debate into a holy war with little more philosophical merit than the war between the Bigenders and the Littlenders in Gulliver's Travels over which end of a hard boiled egg to crack open. Dash has a
much better incentive base than Bitcoin, but that won't stop someone malicious coming along and trying to stall development by dividing the masternode operators. For example, if Dash starts to overtake Bitcoin, Blockstream may go back to their investors and say that their most profitable strategy is to build the Lightning Network on top of Dash, and that they will spend the rest of their millions lobbying the masternode operators to do so. Now? Convincing 4000 masternode operators each with a financial stake in the future of Dash each to do something stupid for their investment will obviously be much harder than convincing 5 miners with no stake in the future to do the same, but it will be yet an order of magnitude harder if there is a theory that says "According to our assumptions about how masternode operators work, changing the collateral/blocksize/fees/rewards like
that, will have an effect something like
this", because then everyone will be able to debate the cause and effect of those assumptions, instead of spreading FUD.
This is what has been going round in my head anyway. I've been reading about agent based modelling of social systems for a few years now and I've just got my head back into it to see what I can apply to Dash.
To return to your original points, I'll pick out these two, a possible advantage and disadvantage:
- Media bonanza - Dash now has the most nodes of all the cryptos.
- The % return per node may decrease (however this is likely to be offset by the increase in price)
So what I would like to know is: is I possible to create a model than embeds the current best understanding of how Dash works into a model, so that yet is possible to simulate the effect of "media boost from having most nodes" to "% return from nodes "? (Or whatever drivers seem most important.) It may not be easy to create this, but it would be an attempt to formalise the various influencing factors, and it may be possible to attract the attention of students, researchers and professors in Sociology and Complexity departments among others to contribute to the effort.
In short: whenever there is a debate over "Value X is too high/low", I think it could be of net benefit in the long run to create a mode for how how that variable behaves which can be simulated, in order to challenge the results of a simulation, rather than risk a stalemate of opinions. It will probably be hard work, but it may well avoid an equivalent of the Blockstream Blocksize Bamboozle being played against Dash,
This started off as a quick reply and rapidly turned into an essay (I do that a lot, sorry). I'm interested to know what anyone thinks though. I have some free time to spend on Dash, so if this seems like a productive avenue to peruse I'll delve deeper, if not, I'm interested to know why not, and more specifically why it might not solve the problems I think it might.