Think it depends a bit how many chunks of bonds he can buy (isn't there some minimum block?)
Yes, the minimum is 100, so I guess at least initially he'll be working on one single block.
Game-theory wise it's almost a classic Prisoner's Dilemma scenario - where optimum strategy for bond purchasers collectively is if EVERYONE bids very high but the optimum for any individual prospective investor is to bid the lowest rate they're willing to accept.
Upon explaining
that mpoe loss ... Maybe there's indeed some room for improvement, looks like nothing is going to stop bondholders to increase requested interest and thus they prevent any MPOE profit (to be paid to shareholders). The competition on interest rate is ineffective, as bondholders seems to calculate that they won't get whole interest anyway.
Maybe bondholder with lower asked interest should be rewarded by lower risk (allocate lesser part of loss to him)?
The bond auction actually works as a disaster of commons, because everyone receives the same interest as the highest rate accepted, but each bondholder's chances of making the cut are an inverse function of the interest asked. Thus the selfish play is to ask for 0% and make sure you're always going to be on the top of the list, receiving whatever % the tail manages to negotiate.
In fact the optimum strategy for bond purchasers collectively is for someone (not for everyone) to ask high, but for each individually to ask low. On analysis this situation is slightly different both from the prisoner dilemma (where indeed best strategy collectively is for everyone to x but best strategy individually is for each to y) and from the classical disaster of commons (where best strategy for each individually is to take most and give least). Maybe Bitcoin has just created its very own game theory situation.
One thing's for sure - as a good portion of the time (including last month) the raised capital didn't cover needed capital
Actually quite the opposite, if you look at the 1st graph above. Most of the time raised capital covered needs. The exceptions are last month, when needs suddenly spiked up 335%, and April, where needs spiked up 330% (merely this stability in spikes seems remarkable to my eye). The bondholders quickly covered the difference in May, and May-Oct there's no red (ie, creditor of last resort contribution).
What possibly throws you off is the period before Mar 2012, but quoting from
the article introducing the bonds:
Were I investing in this, and if I knew others who were, then I'd be looking to collectively fund a bid at 100%, 1000% or whatever the max allowed is
Bond interest asked was originally limited at 2%, then this limit was raised to 5% on macroeconomic data, and after August proved that presumptions of macroeconomic stability were muchly exaggerated the ceiling was removed completely. It is possible it may come back but from what I gather somewhat unlikely. At any rate for now there's no maximum.
That aside, simply having a 1000% interest as the 100 BTC capper of a 10, 50, 100k BTC train of bonds all charging < 1% is obviously not a winning strategy by itself, as it will make all the BTC if and when the needs exceed capital, but it won't make very much otherwise. The agreement in the Board is that the only optimal strategy and only stable strategy is for each individual player to bid exactly the interest rate that he judges adequate on the market as he knows it. They don't have formal proof for this yet (it has been sought on and off for a little over half a year so far).
Is this what you were looking for?
Yeah well I saw that, but was curious if anything more elaborate was contemplated.