Seeking Discussion - GLBSE bonds for web hosting upgrades.

[burnside]

Hey, nice work, I didn't realize that you pioneered the perpetual mining bond.

Probably weren't expecting the crazy computing leaps so quickly.

I may be wrong, definitely wouldn't be the first time.  ;-)  And I see that difficulty hasn't changed much since Feb.

I still have a pretty strong feeling the bonds are going to take a beating in the next 6 months tho.

[1713308856]

1713308856

[burnside]

I may be wrong, definitely wouldn't be the first time.  ;-)  And I see that difficulty hasn't changed much since Feb.

Doh, then I stumbled on this later in the thread:

Re: GLBSE PureMining: Infinite-term, deterministic mining bond
March 03, 2012, 07:33:03 PM
...
ROI projections for the new price:

Issue price - 0.28 BTC
Difficulty - 1496979 (recently increased by 9%)
...

I didn't realize that you'd been updating the original post.  So what was difficulty at when you originally offered up the bond in Feb?

[Sukrim]

Just check it out on Blockexplorer/blockchain.info...

[burnside]

Just check it out on Blockexplorer/blockchain.info...

I don't see a historic difficulty chart on there.  (sometimes I'm half blind, heh.)

Found it here tho: http://bitcoin.sipa.be/

And here: https://bitcointalk.org/index.php?topic=2345.msg31405#msg31405

The difficulty has increased so fast that both graphs are logarithmic. 

[Bitcoin Oz]

I would suggest finding a way to get bitcoin income and then selling shares in it. Even if you use half the fund to purchase bitcoin income earning shares or bonds on glbse and the other half to invest in your USD company.

[Bitcoin Oz]

There is one mining bond with a difference on glbse and that is MOORE which offers growth in the value of the mhs/$ every week. Its remained fairly stable in price as a result whereas other bonds will continue to lose value.

[burnside]

I would suggest finding a way to get bitcoin income and then selling shares in it. Even if you use half the fund to purchase bitcoin income earning shares or bonds on glbse and the other half to invest in your USD company.

That's a good idea.  I could setup something like TEEK.  Problem is that whatever I do, I do not want to be exposed to the Pirate ecosystem.

[burnside]

There is one mining bond with a difference on glbse and that is MOORE which offers growth in the value of the mhs/$ every week. Its remained fairly stable in price as a result whereas other bonds will continue to lose value.

Yeah, I like that one.  Seemed to be well thought out.

[Meni Rosenfeld]

As such, selling a fixed asset, pegged to a fixed computing unit, is bound to follow the same valuation curve as servers do in my business.  That is, you buy it shiny and new at a huge premium, and you recycle it 3 years later, as by then it's no longer worth the cost of power to keep it operating.  The CPU miners have been through this.  The GPU miners are going through this right now.  Eventually the FPGA and ASIC miners will too.  It's inevitable.

Right. This is exactly the decision faced by someone wanting to purchase mining equipment - he wants his revenue from mining to cover the cost before his equipment depreciates. There's no magical way around it. If he believes it will cover it he invests, if not not. If everyone is afraid to buy equipment the difficulty will drop, increasing the profit for those who do invest. Mining bonds allow to make the same investment, but without the physical hardware purchase and operation. Meanwhile it allows the issuer to do the physical operation, without the risk of the investment. The raised funds go directly towards hardware purchase.

So as an investor, looking into one of these mining bonds, I have to make sure that the dividend payout covers the cost of the bond -before- it is essentially worthless.  If you look at the google doc I posted above, you'll see that none of the current dividend payouts cover the entire cost of the bond in under 6 months.

As said this is purely a function of the bond price. The price is determined by the market and it's ok to argue that they're overvalued. If you think they are, you shouldn't invest in them and should consider shorting them.

 The best right now is PUREMINING with a 220+ day payoff.

I think your numbers are a bit off, PureMining is currently the most expensive bond per MH/s.

 How much do you think a 1Mhash share is going to be worth when the difficulty is 10x what it is now, AND the payout has been cut in half?

Not much, and the same applies to anyone who invests in hardware. The one exception is BFL's trade-in program, which is an anomaly not accounted for in the design. It also should be a once in a lifetime event.

The same applies to mining companies too - they raise funds and use it to buy hardware. If the hardware devalues, so does the company.

 Special bonus to the mining operator, "The issuer can buy back the bond at a price equal to 1.2 times the highest price the asset was traded on GLBSE over the prior 744 hours."  So when people realize that the dividend on a 1MH bond is going to drop through the floor and the bond value plummets, the mining operator buys it back up and continues on his merry way.

There is no way for the issuer to profit from the buyback clause. If the price declines to the point where the buyback is cheaper than the original issue, it means the normal obligation of the bond has already decreased. Also, the money was already spent on purchasing hardware to back up the obligation, the issuer doesn't profit from the decrease in mining profitability, the whole point was that he's indifferent to it. The buyback exists only to deal with emergencies, and there must be such a clause - "forever" is a long time.

Thus, to me, making an investment in a fixed computing unit makes no sense whatsoever.  This is especially the case with the huge jumps ASICS are bringing.  Might as well have invested in 486's after the pentium had already been announced, benchmarked, and slated for delivery.

Then nobody should buy mining hardware. They can go ahead, I'll enjoy mining 7000 BTC per day on my CPU.

The way out for the mining operators and investors in their bonds would be to modify the contract to increase the MH output of a given bond 10x.

And they will use magic to increase the hashrate of their hardware?

However if I were an investor shorting the bond, this is the point at which I file a lawsuit against the mining operator for violating the contract.

The issuer's obligations are to bondholders, not shorters. He has the right to do better than stated in the contract.

There is one mining bond with a difference on glbse and that is MOORE which offers growth in the value of the mhs/$ every week. Its remained fairly stable in price as a result whereas other bonds will continue to lose value.

Yeah, I like that one.  Seemed to be well thought out.

Fixed hashrate bonds emulate hardware purchase. An increasing hashrate bond doesn't emulate anything and can't be backed by anything, it's a purely speculative investment not only for investors, but also for the issuer.

[burnside]

First off, thank you for your input.

 The best right now is PUREMINING with a 220+ day payoff.

I think your numbers are a bit off, PureMining is currently the most expensive bond per MH/s.

Not sure... that's not my spreadsheet, and it auto builds using real-time data from GLBSE.

Then nobody should buy mining hardware. They can go ahead, I'll enjoy mining 7000 BTC per day on my CPU.

I kind of feel that way right now... that nobody should buy.  Not that anyone would listen.  It's MAD... mutually assured destruction... all the miners are going to compete each other out of existence. 

The way out for the mining operators and investors in their bonds would be to modify the contract to increase the MH output of a given bond 10x.

And they will use magic to increase the hashrate of their hardware?

That is indeed the conundrum.  However we already know that some portion of the miners are getting "free" upgrades, so why shouldn't that be extended to the bond holders?  Yes, I realize that only half the cost of the upgrade is covered, however the resulting hash rate of the trade is 10x, which means the miner gets 8x for "free".  Normally I'd take the side of the miner and say a contract's a contract, but part of offering up bonds is PR.  You want people to be willing to purchase next time you offer up a bond.

However if I were an investor shorting the bond, this is the point at which I file a lawsuit against the mining operator for violating the contract.

The issuer's obligations are to bondholders, not shorters. He has the right to do better than stated in the contract.

Very good insight.  I hadn't thought about it like that.

[burnside]

However if I were an investor shorting the bond, this is the point at which I file a lawsuit against the mining operator for violating the contract.

The issuer's obligations are to bondholders, not shorters. He has the right to do better than stated in the contract.

Very good insight.  I hadn't thought about it like that.
[/quote]

It occurred to me, traditionally people trying to short the bond would be bondholders at various stages of the shorting process.  Not sure if that plays into it at all or not.

[Meni Rosenfeld]

Then nobody should buy mining hardware. They can go ahead, I'll enjoy mining 7000 BTC per day on my CPU.

I kind of feel that way right now... that nobody should buy.  Not that anyone would listen.  It's MAD... mutually assured destruction... all the miners are going to compete each other out of existence. 

By definition someone should buy. If nobody buys (or purchases an upgrade), the difficulty increase you speak of will not happen, meaning buying would be profitable after all. There's an equilibrium.

The way out for the mining operators and investors in their bonds would be to modify the contract to increase the MH output of a given bond 10x.

And they will use magic to increase the hashrate of their hardware?

That is indeed the conundrum.  However we already know that some portion of the miners are getting "free" upgrades, so why shouldn't that be extended to the bond holders?  Yes, I realize that only half the cost of the upgrade is covered, however the resulting hash rate of the trade is 10x, which means the miner gets 8x for "free".  Normally I'd take the side of the miner and say a contract's a contract, but part of offering up bonds is PR.  You want people to be willing to purchase next time you offer up a bond.

Because the determinism of the bond is important. When someone invest in an abstract commodity it shouldn't be affected by the specifics of the issuer's hardware choices or the quality of his mail service.

An upgrade plan would also be manipulable. I can say that I'll upgrade when the equipment arrives, but then keep it a secret for a few days collecting the profits. The first few days are critical.

I want people to buy bonds the next time on the grounds that I do exactly what I stated in the contract, not on the grounds that I offer bonuses on a whim.

[Sukrim]

 Special bonus to the mining operator, "The issuer can buy back the bond at a price equal to 1.2 times the highest price the asset was traded on GLBSE over the prior 744 hours."  So when people realize that the dividend on a 1MH bond is going to drop through the floor and the bond value plummets, the mining operator buys it back up and continues on his merry way.

There is no way for the issuer to profit from the buyback clause. If the price declines to the point where the buyback is cheaper than the original issue, it means the normal obligation of the bond has already decreased. Also, the money was already spent on purchasing hardware to back up the obligation, the issuer doesn't profit from the decrease in mining profitability, the whole point was that he's indifferent to it. The buyback exists only to deal with emergencies, and there must be such a clause - "forever" is a long time.

Actually there is, and I brought it up numerous times:
There is no way to verify 100% someone really has the hashing rate in hardware they sold. It is even argued, that this is not a necessary assumption in the first place and profits could come from anywhere, as long as they are paid.

With a crash in bond prices it is very well possible to profit from these clauses.
Initial price: 1 BTC
Dividends until a crash to 0.1 BTC: 0.1 BTC
Highest price taded after the crash: 0.2 BTC
Buy back for 0.24 BTC and be happy about it.

Also, if it wasn't traded for 744 hours, you buy back for 0.0*1.2?

This is why I'm now a strong supporter of "bonds have a certain face value and can only be bought back for that face value + slight premium". This is much fairer for both people holding bonds (they don't have to check every 744 hours if they can be screwed over) and bond issuers, who according to some contracts out there, would have a VERY hard time to reduce their debts.

[Meni Rosenfeld]

 Special bonus to the mining operator, "The issuer can buy back the bond at a price equal to 1.2 times the highest price the asset was traded on GLBSE over the prior 744 hours."  So when people realize that the dividend on a 1MH bond is going to drop through the floor and the bond value plummets, the mining operator buys it back up and continues on his merry way.

There is no way for the issuer to profit from the buyback clause. If the price declines to the point where the buyback is cheaper than the original issue, it means the normal obligation of the bond has already decreased. Also, the money was already spent on purchasing hardware to back up the obligation, the issuer doesn't profit from the decrease in mining profitability, the whole point was that he's indifferent to it. The buyback exists only to deal with emergencies, and there must be such a clause - "forever" is a long time.

Actually there is, and I brought it up numerous times:
There is no way to verify 100% someone really has the hashing rate in hardware they sold. It is even argued, that this is not a necessary assumption in the first place and profits could come from anywhere, as long as they are paid.

With a crash in bond prices it is very well possible to profit from these clauses.
Initial price: 1 BTC
Dividends until a crash to 0.1 BTC: 0.1 BTC
Highest price taded after the crash: 0.2 BTC
Buy back for 0.24 BTC and be happy about it.

He isn't profiting from the buyback clause. He's profiting from his successful short position on mining, buyback or no buyback.

Also, if it wasn't traded for 744 hours, you buy back for 0.0*1.2?

Despite being completely far-fetched, this is a good point, and I guess in this situation the relevant price would be the last traded price.

This is why I'm now a strong supporter of "bonds have a certain face value and can only be bought back for that face value + slight premium". This is much fairer for both people holding bonds (they don't have to check every 744 hours if they can be screwed over) and bond issuers, who according to some contracts out there, would have a VERY hard time to reduce their debts.

But the face value of a mining bond is 1 MH/s. Nobody can place a fixed BTC price on it, this is determined by the market and is constantly changing - if I buyback the bonds 10 years from now I don't want it to be based on the original issue price, but rather on their value at the time.

All that said, I agree it would be best to introduce a deterministic "extrapolated lifetime earnings" metric (which I introduced elsewhere but called it "projected") and assume that the bond value is roughly a constant multiple of it, and base the buyback price on it (with or without consideration to the market price). I'll probably do that if I create a new bond series.

[GeoRW]

Have you considered providing your services for bitcoins? Maybe you could find more "bitcoin" investors for your company that way.

[burnside]

Have you considered providing your services for bitcoins? Maybe you could find more "bitcoin" investors for your company that way.

Of course.  (see my sig)

[Sukrim]

This is why I'm now a strong supporter of "bonds have a certain face value and can only be bought back for that face value + slight premium". This is much fairer for both people holding bonds (they don't have to check every 744 hours if they can be screwed over) and bond issuers, who according to some contracts out there, would have a VERY hard time to reduce their debts.

But the face value of a mining bond is 1 MH/s. Nobody can place a fixed BTC price on it, this is determined by the market and is constantly changing - if I buyback the bonds 10 years from now I don't want it to be based on the original issue price, but rather on their value at the time.

All that said, I agree it would be best to introduce a deterministic "extrapolated lifetime earnings" metric (which I introduced elsewhere but called it "projected") and assume that the bond value is roughly a constant multiple of it, and base the buyback price on it (with or without consideration to the market price). I'll probably do that if I create a new bond series.

I would argue that *the dividends depend on* 1 MH/s but the face value is the price at IPO.

Even if the face value were 1 potato, you could at least hand me 1 potato to buy it back. Something that is a process and not a good can not be stored or transfered. I lack the correct financial terminology, but I guess you get the point. Something that might be close to what you imagine would be electricity markets, where you can buy/sell electricity in bulk. I don't know of any "1 kwh bonds" but on the other hand electricity doesn't generate money on it's own for dividends, mining does however.
Probably a better equivalent would be selling bonds with a face value of 1 hour of sunshine on 1 m². Depending on solar panels etc. you can "mine" a certain amount of electricity and sell it on the market for a certain amount of money. Still you cannot control some external factors like the weather and would have to factor these in. Honestly I still don't think that an immaterial face value is any good - and that it can NOT be measured in money by taking a trade price point of the last few days.

[Stephen Gornick]

I've researched the legality of offering up bonds in the US.  It seems legal, at least on the surface.  Small fry's are allowed to offer up bonds to private investors without having to register with the SEC.  So I just have to make sure that at some level my bond investors are "private".  I would thus define the potential pool of private investors in my bonds by saying our bonds are only available to those in the GLBSE Bitcoin Club, (aka, registered members of the GLBSE)  or something of the sort...  However... I'm not really sure if this would pass the test of an actual US courtroom or not.  The definition of a "private investor" seems somewhat vague.

I presume the "private investor" part simply means you didn't solicit the investor and that there was a prior substantive relationship.  As far as bonds not being equities and thus that affecting the requirement to register with the SEC, I believe that might be an inaccurate statement.

Section 2(1) of the Securities Act of 1933 (the “Securities Act”) sets forth a definition of the term 'security,' and the term includes, in addition to stock and bonds, any note, evidence of indebtedness, option or investment contract. This broad definition of security means that virtually any type of instrument in which the investor has a reasonable expectation of profit solely as a result of the investment of money will be subject to the rigor of the federal securities laws, regardless of the structure of the investment.

- http://bit.ly/voMtiX

Obviously for a business that operates online, locating it to a jurisdiction with lax securities regulations becomes a greater and greater opportunity.

[Meni Rosenfeld]

This is why I'm now a strong supporter of "bonds have a certain face value and can only be bought back for that face value + slight premium". This is much fairer for both people holding bonds (they don't have to check every 744 hours if they can be screwed over) and bond issuers, who according to some contracts out there, would have a VERY hard time to reduce their debts.

But the face value of a mining bond is 1 MH/s. Nobody can place a fixed BTC price on it, this is determined by the market and is constantly changing - if I buyback the bonds 10 years from now I don't want it to be based on the original issue price, but rather on their value at the time.

All that said, I agree it would be best to introduce a deterministic "extrapolated lifetime earnings" metric (which I introduced elsewhere but called it "projected") and assume that the bond value is roughly a constant multiple of it, and base the buyback price on it (with or without consideration to the market price). I'll probably do that if I create a new bond series.

I would argue that *the dividends depend on* 1 MH/s but the face value is the price at IPO.

Even if the face value were 1 potato, you could at least hand me 1 potato to buy it back. Something that is a process and not a good can not be stored or transfered. I lack the correct financial terminology, but I guess you get the point. Something that might be close to what you imagine would be electricity markets, where you can buy/sell electricity in bulk. I don't know of any "1 kwh bonds" but on the other hand electricity doesn't generate money on it's own for dividends, mining does however.
Probably a better equivalent would be selling bonds with a face value of 1 hour of sunshine on 1 m². Depending on solar panels etc. you can "mine" a certain amount of electricity and sell it on the market for a certain amount of money. Still you cannot control some external factors like the weather and would have to factor these in. Honestly I still don't think that an immaterial face value is any good - and that it can NOT be measured in money by taking a trade price point of the last few days.

With traditional bonds, what is called "the face value" isn't how much it was issued for, it's how much it is bought for at the end of the term. The face value doesn't change - if it's $1 then I'll get $1 for it at the end of the term, whether I buy it today or last year (that's why the face value is chosen as a round number). The traded price however constantly changes - the bond can be issued for $0.839, then traded at $0.927541 and then I can find someone on the street and sell it to him for $0.509172.

Same for mining bonds. Instead of "face value is 1 BTC, to be paid on January 12 2014", we have "face value is 1 MH/s to be paid to eternity". The issue price and traded price can vary (and there can be multiple issues at different prices), but the bond always has the same face value. As it happens, if there is a need for a buyback a Bitcoin price must be placed on it, and despite some shortcomings, relying on recently traded price fits the role.

Once again, if a mining bond is issued at 0.4 BTC, it does not still have a "face value" of 0.4 BTC 10 years afterwards.

[Sukrim]

And exactly that is a problem in my opinion.

If you were truly open to any price, you'd issue new bonds at 1 Satoshi each and pre-announce it, so the market can build it's own price in a bidding war. In reality you have some minimum price per MH/s when issuing a 1 MH/s bond (probably at least the price for the hardware + electricity for a year or so) that you can sell at. You don't set a lower limit to the price of buying back though, so the advantage is one-sided on your side.

1 MH/s in itself is not used to determine the value of a bond (the value is 0 according to your post I think) but only the dividends. People don't buy the bonds for their value but the possibility of having a ROI through dividends. Still you will not sell them at any price but can buy them back at any price, even if that would mean a net loss for people buying the bonds from you initially.

A more sophisticated  buyback clause could also be "1 bond can be bought back at (0.3 BTC - dividends per share since IPO) * 1.2 (if this gets negative, this is assumed to be 0) or 2 USD converted to BTC at the 24h average market value, whatever is higher" - if your minimum price is something a bit below 2 USD per MH/s.
This means you can still get very low buyback prices, but not arbitrarily low ones, will always have to pay up a proper amount for the debts you sold and it also tells that you initially priced your bonds in USD, as mining equipment is priced in USD as well (hence the not arbitrarily low IPO price).