Author Topic: Delegates responding to market rates, weird game theory with current set-up  (Read 1343 times)

0 Members and 1 Guest are viewing this topic.

Offline maqifrnswa

  • Hero Member
  • *****
  • Posts: 661
    • View Profile

I think what you guys forget is that BTSX is deflationary and it will always increase in value.

no, i'm quite aware of that  ;)
serving as a delegate should be the equivalent of buying BTSX at a discount. However, if the cost of running a node is more than the earnings, I'm running a node for the privilege of being forced to buy BTSX at a premium. If I think BTSX is going up (that is, always increasing in value as you say), why would I pay to run a delegate node when I can buy on the open market for less? That is the vector we're discussing. The problem is that delegates in that  situation have no recourse besides damaging the network by turning off their highly-voted node. They can't increase pay rate, so they can only discontinue service in order to stop losing money.


Here's another perspective.  The number of delegates are currently fixed at 101.  So as the transaction volume ramps up the delegates are only going to make more and more money.  It's not like the system requirements are going to scale at the same rate as transaction volumes.  If the transaction fee is controlled by the delegates then it's just up to the delegates to come up with some balance.  And really once BTSX are worth $1/ea or more than we'll be fine even at 95% burn rate. If BTSX doesn't get higher than that then we're all sort of out of luck anyway :) .

The real control we have to worry about is transaction fee.  Naturally if BTSX goes to $100 no one will want to pay 0.1BTSX so a balance will have to be struck.  Dynamic pay rates is just one lever, the delegates have others.

This is very true, transaction fee rates may too need a market. If that happens, then pay rates definitely need a market since delegates would have no control over their income.

EDIT: there/their oops
« Last Edit: July 31, 2014, 01:48:56 am by maqifrnswa »
maintains an Ubuntu PPA: https://launchpad.net/~showard314/+archive/ubuntu/bitshares [15% delegate] wallet_account_set_approval maqifrnswa true [50% delegate] wallet_account_set_approval delegate1.maqifrnswa true

Offline maqifrnswa

  • Hero Member
  • *****
  • Posts: 661
    • View Profile
I hate replying right after myself, but I think this may be more serious than I originally thought:

Assuming constant fees, and fees are in BTSX (not bitUSD). At some point the delegate payrate market will become efficient, and all nodes will have approximately the same pay rate a little above the cost of the node (for example, 5%). Imagine a shock to price (e.g., a one day drop of 10% in BTC, which has happened in the past). ALL nodes now are operating at a loss, and the correct thing to do economically is to turn off enough so the remaining nodes become profitable again. That will damage the network, which may cause BTSX to drop in value, which then would cause more nodes to turn off, until a new equilibrium is reached with only a portion of the network being active.

Of course, this outcome should be avoidable by selecting delegates that promise not to do that. But how long would you operate in the red before turning it off? $10? $100? $1000 lost? At some point, delegates not being able to raise their fees may lead to a breakdown in DPOS.

EDIT:
Not allowing delegates to raise their fees is the same policy as China not allowing short selling in the market. It sounds like a good idea, but leads to crazy economic outcomes. Even China had to give up on that idea:
http://www.bloomberg.com/news/2013-02-23/china-to-expand-short-selling-program-as-part-of-reform.html
« Last Edit: July 31, 2014, 02:09:14 am by maqifrnswa »
maintains an Ubuntu PPA: https://launchpad.net/~showard314/+archive/ubuntu/bitshares [15% delegate] wallet_account_set_approval maqifrnswa true [50% delegate] wallet_account_set_approval delegate1.maqifrnswa true

Offline Riverhead

I think that's one of the big advantages to a PoS system over PoW.  The cost to operate a delegate is substantially less than running a PoW GPU/ASIC farm.

For a real world example it costs me about $15/m to run a delegate. I was paying $1500/m in power costs (plus about $12k in hardware) to run a PoW farm.  Granted the ROI made sense at the time for PoW however there is no way I'm going to eat $1500/m "for the good of the network" but I'll hardly miss $15/m to ride out a low time.