I always write my contracts so as to commit to less than I actually intend to deliver.  Hence with DMS I'm only commited to weekly reports - when in fact I'll be trying to provide them daily AND maintain near real-time access to asset lists.


It's better to under-promise and over-deliver than the other way around.

What makes you believe his comments had any validity?  If someone randomly assumes earnings are being retained that's no reason to believe it's actually the case.

Given that the contract explicitly states 100% of profit is paid as dividend it isn't actually possible for the COMPANY to retain earnings.  What IS possible is for evorhees to increase the working capital he provides (as part of his obligation in return for receiving all IPO proceeds).  That has the same impact (in terms of maximum bet sizes and risk of ruin) as if the company owned the float and retained earnings - but is not actually the same thing.

A strange world where providing something in return for keeping 100% of IPO proceeds and 87% of profit is "free".

The shares were sold as representing a share of profits.

If the pool was meant to be provided by shareholders then a portion of the funds raised from IPO should have gone for that purpose rather all of it going to you.

You received ALL of the capital raised from sale of shares, in part, because YOU were providing the working capital needed.  Accepting that risk - and providing that commitment - was all part and parcel of the original sale.

The key part of the contract of relevance is NOT the bit saying 100% of profits will be distributed, but rather the part defining profit.  Specifically:

"Net Profits here mean profits after explicit SatoshiDice development and Marketing costs (such costs, if any, will be described monthly in the Profit and Loss Statements). SatoshiDice does not pay salaries to any party, so no salary will be taken out of Net Profits. "

Only development and marketting fees can be treated as costs - NOT interest on loans or provision of capital.  And, for that matter, not any salaries or other costs related to running the business unless they're explictly either development and marketing.  In short there should be ZERO deductions in respect of anything which could be considered an operational cost.

How to do that? well of the 100% of IPO revenue you received (counting yyour partner as you for simplicity) and the 87% of profits you receive maybe you should have assigned some portion to cover YOUR end of the deal.  Which includes covering the costs of actual operation - one of which is the cost/rsik of providing sufficient capital.

Alternatively you could have done the contract differently - so that some portion of IPO funds was used as a float.  You could, in effect, change to that now - by GIVING to the company that portion of IPO funds that should have been retained if you'd IPOed with investors supposed to provide working capital.

What isn't acceptable is trying to charge investors twice - once so you can have it and then a second time so the funds go where they should have gone in the first place.

Another alternative would be to sell some of your own shares to raise cash to replace that currently tied up in the float.  That would be contractually allowed and is likely the best solution - even if it does produce a lot of whining about depressed share prices.

There's still provision for fees to be charged in some circumstances - from memory if burnside sells his controlling interest in LTC-GLOBAL.  So it makes sense for tha to be in there, even though it doesn't currently mean anything.

A good question - and illustrates the reason why, in nearly all cases, management fee should be based on profit rather than on sales.  I believe this set of securities to be an exception to the general rule and will try to explain why.  But first let me give an assurance - and explain why PART of your argument is wrong (the general question remain an entirely valid one though).

It is my intention that I will always hold a significant position in MINING and/or SELLING myself.  I'm not going to disclose which of those I currently hold most - and that may well change over time anyway.  My view is that NEITHER of MINING or SELLING are intrinsically better than the other - which one it makes most sense to hold depends entirely on the prices at which they trade (compared to the investor's own expectations of future difficulty).  So even if I hold a lot more of one than the other at present that ratio could reverse at any point in the future if prices change.

The danger (in terms of lack of incentive for me) is NOT if I go long on MINING rather than on SELLING - for 2 reasons :

1.  Even if were long on MINING right now, if prices move then my view could change causing me to want to move to SELLING.  Obviously that gives me an incentive to have sound investments - as when I change the income (and the reliability of the capital) becomes mine.
2.  If I'm long on MINING then good management of the capital increases the maximum payout I can receive.  Specifically, at any difficulty change where the increase in capital cover (expressed as days of dividends) is less than or equal to the number of dividends paid to MINING on the previous difficulty there would be NO dividend for SELLING and all capital growth would be retained as cover for MINING.

The danger in terms of commitment from me is if I sell ALL my MINING and SELLING and have no (or a negligible) stake at all left.  Now that will NOT be the case (and I'm fine for burnside to confirm at any time that I hold at least 3K combined MINING+SELLING - including PURCHASE as 1 of each.  Only time I'd fall below that would be temporarily if I were able to sell BOTH at a price exceeding my cost to replace them with a PURCHASE).  Obviously (I hope) I can't expose my personal portfolio - as that would reveal precisely what I held and what I had up for sale.

Even ignoring that, there IS still good reason to believe I'd be committed to doing a good job - even if I held no shares at all.

1.  My reputation.  If I plainly make bad investments, fail to maintain liquidity or don't try to invest at all then my reputation will suffer.  I like to think I have a good reputation so far for doing a good job - I don't want to damage that.
2.  My profit.  My current fee may be only on sales of PURCHASE, and not directly effected by how I manage the fund, but what is gong to happen to sales if I obviously do a bad job?  What will happen to sales if I obviously do a GOOD job and PURCHASE becomes tempting in its own right as medium-term investment?  The answers to those questions should clearly show a financial motivation for me to do a good job even if I held not direct stake myself.
3.  My initial intent.  It was never my expectation that the low management fee on sales would make me any significant profit.  My intention was always to be able to make a profit myself by trading/holding MINING and SELLING.  And that then leaves me getting direct benefit from sound management of funds - the same as everyone else.

I haven't rushed into investments yet - because investments have to be approved by a vote of SELLING (other than the one we already have).  And I didn't want to pass such votes immediately whilst I held an excessive chunk of voting power from my direct purchases and from my trade-in of LTC-ATF.B1.  I'll be posting later tonight with the first few proposals for discussion - but not initiating votes just yet.

Now let's look at the alternative ways I could have taken fees:

1.  As a percentage of all payments made out.  This is actually how I initially wrote the contracts.  I would then be effectively getting a percentage of all sales of PURCHASE AND a percentage of all capital growth.  I removed this for a bunch of reasons:
a)  It added a LOT of text and complexity to the contract - as there's a lot of places where payments get made out.
b)  As a result of a) it adds a lot of complexity to the calculation of management fee.
c)  a) and b) together mean less transparency - it's less OBVIOUS to investors what I'm taking and when.
d)  The net result would not be a lot different - because we'll only ever be touching low-risk investments (with commensurate low rewards) the bulk of fee would still be on initial sales.

2.  By taking management fee as a (much larger) percentage of (total returned capital - total received capital).  So I'd take my fee only when either the fund ended OR total payments to date exceeded total sales.  That means my reward isn't directly for running this - but instead a percentage of what profit I make using the capital.  

On the face of it that's ideal - as it obviously means I want to maximise profit from investment.  But unfortunately the truth isn't so simple.  If I get a significant percentage of profit but don't bear ANY losses then it actually gives a direct incentive to gamble or take high-risks.  I mean flipping the whole bank-roll on a 50/50 S.DICE spin would be MASSIVELY +EV for me.  If it loses I lose nothing - if it wins I get a good percentage of the profit.

Now I'd never do that sort of thing anyway - but my pay being based on profit is inevitably going to have some non-zero impact on how I assess investment opportunities.  Extreme examples like S.DICE are clearly not allowed by contract anyway - but it remains a bad idea to define in contract a financial incentive for a manager which encourages behaviour directly detrimental to one class of investors (MINING).  And that's a problem which exists far more here than in other securities : MINING and SELLING do NOT have the same interests when it comes to investments.

That is why MINING don't vote on investments - but SELLING do.  If MINING had to vote as well then in nearly all scenarios a rational MINING voter would vote no on ANY investment - as the majority of time they won't see the benefit of it.  But SELLING actually have an incentive to take a level of risk ABOVE that which it would be fair to impose on MINING.  So what keeps SELLING in check?  I do - as I control what votes occur and also whether or not I invest even when something is approved.  And if I'm the only thing ensuring MINING get treated fairly then I absolutely can NOT have a management fee structured so that I'm financially rewarded for risk-taking beyond even the level SELLING would want.  Do note that whilst SELLING should STILL have a very low risk threshhold it IS going to be significantly higher than MINING's - and there's nothing to stop a gambler or two getting a lot of SELLING and wanting to gamble.

In short, if you believe I'll act with integrity and commitment then the current system IS the best one.  If you believe I won't act with integrity or will be lazy then at least the current system gives me a financial incentive NOT to do really stupid things.  The current system means I have no conflict of interest between maximising my personal gain and ensuring fair treatment for both MINING and SELLING.

Hope that explains it a bit.

EDIT:  Just to add, at present I hold a 4-figure number of BOTH MINING and SELLING (I haven't been trying too hard to sell mine yet).

Dividend payment has processed.

The new wallet balance is : 359.54861969
If you add to that 100 BTC for our investments you'll see the result si very close to that in report:

NAV Post SELLING Div    459.54862112

The difference in the last decimal places is simply rounding errors - specifically the dividend actually paid is an exact amount of satoshis per share.  The spreadsheet calculates at higher precision.  I won't be posting verification of this comparison every day - but it IS one of the checks I'll be doing to ensure reported results are correct.  It is, of course, verifiable by anyone else who looks at the public wallet.

I'm doing the report BEFORE dividend is paid - hence PURCHASE ask (and bid) being taken down.  I'll still process exchanges near the end - those might make the share numbers different but changing nothing of importance (dividends are paid to PURCHASE as well as MINING so total dividends paid remains the same).  Posting it before dividends are paid allows anyone who is around to adjust their orders as they see fit.

Here's the calculation of management fee:

Sold   916
Price   0.066195
Total   60.63462
Less Fee   60.51335076
Man Fee   1.815400523

I placed 10K PURCHASE for sale yesterday and 9084 remained when I took the order down.  So I know there were 916 sales on which a management fee is due.  I can then calculate the total value of those sales, then take off the 0.2% site transaction fee and finally calculate management fee as being 3% of what was actually received.  Fee of 1.815 BTC has been transferred.

Here's the main data:

BTC Balance (BTC-TC)    360.72374078
10000 LTC-ATF.B1    100.00000000
TOTAL ASSETS    460.72374078
   
Outstanding MINING   7203
Outstanding SELLING   7203
Outstanding PURCHASE   90
Effective Units   7293
   
Block reward   25
Difficulty   15605633
Hashes per MINING   5000000
   
Daily Dividend    0.00016113
50 days (Min Liquid)    0.00805649
100 days (Forced Close)    0.01611298
365 days (Buyback)    0.05881238
405 days (IPO)    0.06525757
400 days (Post SELLING div)    0.06445192
410 days (Pre SELLING div)    0.06606322
   
NAV Post MINING Div    459.54862112
NAV/U Post MINING Div    0.06301229
Days Dividend Post Div   391.07
SELLING Dividend    -         
NAV Post SELLING Div    459.54862112
NAV/U Post Selling Div    0.06301229
PURCHASE selling price    0.06616290
PURCHASE buy-back price    0.06175204

Those paying close attention (or more likely just those reading this BEFORE looking in detail) will notice one surprising thing.

After yesterday's dividend there were 391.26 days of dividend left as capital.  After paying today's that has only dropped to 391.07.  Surely it should have dropped by 1 day?

The reason it hasn't is the sale of new units of PURCHASE.  Even after management fee and site fee are taken those still sell at a markup to current NAV/U.  That acts to slighjtly increase the number of days of dividend remaining.  The impact of that will fall as new sales of PURCHASE becomes a smaller percentage of effective outstanding units.  Revenue from investments will, however, have a similar effect - so days of dividend remaining will nearly always fall by less than a full day each day.

Because of the small drop in days dividend remaining (i.e. NAV/U) the selling price for PURCHASE also only drops very slightly.  This effect ensures that it will ALWAYS be better to buy PURCHASE before a day's dividend payments rather than after it (as the dividends you receive will always exceed the reduction in price - even if only very slightly).

If anyone has questions about any of these factors - or why I designed specific elements the way I did - then feel free to ask.  The only questions I won't answer are ones relating to how I think MINING and SELLING should be relatively priced - that's for everyone to decide themselves.

Looks like that's going to be a useful site - thanks for the link.

Very minor bug.  If you look at the first dividend for DMS.PURCHASE you'll see :

2013-06-11 17:01   ฿ 0.02078577   129   0.00016112   COMPLETE

However the dividend was actually 1 satoshi higher per share : 0 .00016113

As can be seen from my personal account's record of it :

Dividend History: DMS.PURCHASE Reinvest At Max ฿  Per Share
Date   Shares Held   Per Share   Total
2013-06-11 17:01:16   1   ฿ 0.00016113   ฿ 0.00016113

Seems to just be a diplay bug - presumably a rounding error when it recalculates div/share from the total divdend paid.

Also a feature request : If I schedule a divdend to be a fixed amount per share would it be possible to have the divdend list show it as that amoun per share rather than an estimated total to be paid based on a cached outstanding share count.  Would just be nice if, when I schedule a dividend, people could see what they'd get per share.  Obviously that only applies when the dividend is scheduled on per share basis - if the dividend is of a fixed amount then that fixed amount should be displayed as at present.

There's a bunch of reasons why TAT is higher.  The long contract for MINING is definitely one of them.  

One of the main reasons is just lack of supply.  If you look at what happened when the 21K were put up, the vast majority were bought in 2 big buy orders - so there's never actually been a chance for most people to buy them other than from the resellers.  If I'd been around when the batch went up I would have bought a bunch as well - not because I necessarily think they're worth .007 or more - but because there was always going to be people who'd buy them for 10% more.  I did that with them on Bitfunder (for LTC-ATF).

In comparison, MINING has near-unlimited supply (so long as there's people willing to buy PURCHASE and sell the MINING).  So there's no way for people to buy MINING with any great certainty of being able to sell at a profit.  That means they have to worry about what the shares are actually worth - rather than just following the "bigger fool" principle which is what flipping is all about.

But the extent of the price difference does surprise me a bit - given the enormous difference in buy-back terms.  I guess a lot of investors don't even read that far.  Or maybe they're ALL just hoping to sell for a bit more than they bought for - in which case some of them are obviously going to have an unhappy result (in that they'll have to keep shares they didn't intend to or take an immediate hit - there can only be so many cycles of reselling higher before the buyers dry up).  It's interesting to watch anyway - and it'll get even more interesting later in the week I think.

And do realise that the majority haven't actually bee successfully resold - total volume is 24642 meaning only 3.6k have resold so far.   Other than the two massive buys from resellers we've actually sold more hashes than TAT - the high prices are more because noone (other than the 1 or 2 resellers) has any to sell than because of overwhelming demand at a price over .007.

Yeah - like PMBs we don't pay fees.  Or maybe think of it as we mine 'virtual' blocks with no transactions in.  Of course we also don't get stales, orphans or power-cuts on our virtual mining - which probably more than compensates.

Seems the above is just a display error in historical dividends.  If I look at my personal account (where I held exactly 1 PURCHASE - I wanted to make sure someone got the dividend so there was a record of it) I see :

Dividend History: DMS.PURCHASE Reinvest At Max ฿  Per Share
Date   Shares Held   Per Share   Total
2013-06-11 17:01:16   1   ฿ 0.00016113   ฿ 0.00016113

So the correct dividend WAS paid - it's just being reported incorrectly in the security history.

I have noticed a tiny error myself.

MINING received a dividend of 0.00016113 per share
PURCHASE a dividend of 0.00016112 per share - 1 satoshi less.

Whilst I doubt anyone cares about 1 satoshi it shouldn't happen.  I know for certain that I entered the correct amount for the PURCHASE dividend per share - as I picked it from the list of previously used values.  And that list contains 0.00016113 but not 0.00016112.  I'll prod burnside about it - it looks like some very minor rounding error somewhere (am guessing a dividend per share is done by calculating the total dividend then dividing it by number of shares - rather than just by sending the specified amount).

Below is a copy/paste from my spreadsheet showing all calculations made in respect of dividends.  I'll explain all the lines today - on future days I'll just copy/paste the report here (with the calculation of management fee above it in same post).

BTC Balance (BTC-TC)    303.05291603
10000 LTC-ATF.B1    100.00000000
TOTAL ASSETS    403.05291603

The data above shows the value of the fund immediately prior to dividend payment.  This is AFTER the deduction of (any) management fees due.

Outstanding MINING   6248
Outstanding SELLING   6248
Outstanding PURCHASE   129
Effective Units   6377

The next block of data records outstanding shares.  Although the number of outstanding shares on BTC-TC is cached, the issuer can work out the exact number by looking at their own holdings.  Effective units is PURCHASE + (either of MINING or SELLING).  If there are different numbers of MINING and SELLING then this will say "ERROR" and it means I've just been very careless.  Effective units is the number by which NAV is divided to find NAV/U.

Block reward   25
Difficulty   15605633
Hashes per MINING   5000000

This data are the variables used to calculate dividends.
   
Daily Dividend    0.00016113
50 days (Min Liquid)    0.00805649
100 days (Forced Close)    0.01611298
365 days (Buyback)    0.05881238
405 days (IPO)    0.06525757
400 days (Post SELLING div)    0.06445192
410 days (Pre SELLING div)    0.06606322

This block of data show the calculated daily dividend and (for reference) various relevant values for "X days of dividends at current difficulty".)
   
NAV Post MINING Div    402.02539127 <-- This is the total assets calculated earlier less (daily dividend * effective units)
NAV/U Post MINING Div    0.06304303 <-- This is the NAV calculated in the previous row dividede by effective units and is NAV/U after payment to MINING
Days Dividend Post Div   391.26 <-- This is number of days of dividends left as capital.
SELLING Dividend    -         <-- If the previous row is greater than 410 then the dividend to pay SELLING is calculated here.
NAV Post SELLING Div    402.02539127 <-- NAV after paying any dividend due to SELLING
NAV/U Post Selling Div    0.06304303  <-- NAV/U after paying and dividend due to SELLING
PURCHASE selling price    0.06619518 <-- Price to sell PURCHASE at.  That's the NAV/U in previous row *1.05
PURCHASE buy-back price    0.06178217 <-- Price at which the fund will buy-back PURCHASE (or MINING + SELLING).  98% of NAV/U.

The actual wallet balance after this dividend (plus 100 BTC for our LTC-ATF.B1) matches the calculated post-div NAV to 5 decimal places.  The last few digits will rarely match due to rounding errors.

You will note that the days of dividend left after this report have dropped from 405 to 391.  That's not as alarming as it may seem - and today will be the only day it drops significantly (normally it will drop by 1 or less).  The drop is because 3% management fee was taken - when you buy a PURCHASE initially at 405 days' dividend, 12 of those days are going in management fee.  The remaining drop is due to transaction fee on sales and, of course, a 1 day reduction from actually paying a dividend.

If the management fee seems harsh, remember it's only taken once - there's no other management fees at any times.  A 3% fee is VERY cheap compared to just about every other security out there.  I just don't try to hide or the disguise the fee.

All questions, comments and suggestions for changes/additions are welcome.

Bid has been taken down.

As this is the 1st day of trading the management fee can be calculate from current assets.

Wallet (315.51791603) + Investments (100 BTC) = 415.51791603 BTC
A management fee of 3% is due on that - 12.465 BTC.

This will be transferred now and the balance after that used for calculating the report.

On future days the management fee will be calculated on new sales.

I've nearly made back the 15 BTC listing fees in one day - a pleasant surprise.

Please do not send any more PURCHASE to me until after the dividends have processed - I do not want PURCHASE sitting with me and not receiving dividends.  I'll process any exchanges made nearer dividend time.

Don't panic if you haven't sent in PURCHASE to exchange - all dividends paid to MINING also go to PURCHASE (as PURCHASE is a MINING + a SELLING).

I'll leave the ask-wall of PURCHASE up until a few minutes before dividend payment.

Details of the dividend payment and all other accounting info should be up within about 10 minutes of the dividend actually processing (which could be at any point in the 15 minutes after 16:00 GMT).  A fresh ask for PURCHASE will then be placed up at the new price (which will be very slightly above the current price - probably the only time its price will ever rise).

Per the contract I just exchanged in 10,000 LTC-ATF.B1 in return for units of MINING and SELLING.

LTC-ATF.B1 is a bond, issued by my trading fund LTC-ATF.  Although it's traded on LTC-GLobal (in LTC) it has a face value of 0.01 BTC on which 0.6% interest is paid each week.

No new LTC-ATF.B1 are being issued at present (it's actually unlikely any ever will be - as my belief is LTC-ATF can now raise capital cheaper than that if it needs it).  The bonds traded in are ones personally owned by me - either bought via the market or bought directly from other investors who wanted to sell back.  No bonds are availalable cheaper on the market (the cheapest Asks are at about 10% over face value).

There is no currency risk associated to these.  The face value is fixed in BTC - and backed by BTC-denominated assets exceeding the value of all debt.  The bonds are further backed by all LTC-denominated assets of LTC-ATF.  In total the bonds are backed by assets worth well over double the value of all debts/liabilties of LTC-ATF.  At this instant 91.28% of all of those assets are cash - either LTC or BTC on various exchanges (most of which is backing Bids).  The dividends are paid in LTC - this will be exchanged back to BTC (by LTC-ATF) at the same exchange-rate used when calculating the dividend payment for this fund immediately after dividend payment (because the debt is BTC-denominated LTC-ATF has to exchange some LTC back to BTC anyway to maintain its ratio of currencies, so this saves LTC-ATF some tiny fees whilst ensuring zero non-BTC exposure for this fund, even on dividends)

In short, they're about as low-risk as you'll get.  This is also likely to be about all of them that this fund gets (unless anyone wants to trade some in at face value).  The bonds are redeemable at 99% of face-value to LTC-ATF.  In practice I'll just buy them back myself at 100% of face value when the fund needs to get rid of them.  They will generate just over 0.6 BTC per week for the fund (LTC-ATF rounds dividends up to be round numbers, so the effective percentage paid per week is slightly over 0.6%).

As of this instant these now constitute around 25% of all assets of the fund.  That percentage will drop as more PURCHASE are sold.

10,000 LTC-ATF.B1 is 100 BTC face-value.
1532 each of MINING and SELLING have been transferred to my personal account along with change of 0.024744 BTC.

I will now update the fourth post of this thread so it can be used to calculate all of our assets with a reasonable degree of accuracy at all times.

A show of good faith would be paying out Nyan.A as promised.  Good faith is keeping your existing promises - not making new ones.

Just deleted a response from usagi - didn't bother reading it as about to head to bed.  Reason I delete his posts is simple - if he wanted to discuss things in one thread then his original one was the place.  You can't start a policy of deleting one half of a discussion just because you believe you're right then expect those on the other side to extend to you the courtesy you denied them.

Post in your own thread usagi.