Exchange price vs. resource price


#1

This is reference to an earlier topic here (the link takes you to @dirvine’s answer because it’s one of most interesting, but the entire topic is interesting):

One of the questions seems to be what should “lead” - the exchange price or the resource price (or perhaps self-adjust the latter so that it effects a stable exchange rate of Safecoin on various exchanges).
I find that question extremely interesting.

I’ve just started thinking about this, but my initial thinking is that the resource price (supply/demand) should lead and self-adjust purely based on supply and demand (I wouldn’t even use averaging).

Reasons:

  1. Supply/demand data should be available raw (I’d be happy with something like system load average - current, X min, Y min)).
  2. Whenever deliberate data/price manipulation (price management, price control) is introduced, it invariably results in unforeseen problems and reduces the transparency.
    Safecoin should be traded like a commodity (which it is) where the current price is how much you’d actually have to pay for SAFEnet resources right now and exchange prices would self-adjust.
  3. It’s pointless to have a stable price (“tuned” Safecoin exchange rate) denominated in a bunch of unstable cryptocurrencies. I
  4. (gold bug alert) Think the gold standard from 100 years ago (where the exchange price “leads”) vs. free gold (natural price of gold as determined by supply and demand)

I could probably come up with 3-4 additional arguments, but it’s late…

One “downside” that I predict would be raised if I don’t don’t mention it now: some may claim my preferred approach would create ample space for “speculators”, “middlemen” and “profiters” and is therefore bad.
That would be a false argument (please visit www.mises.org for details).

I’m sure there are other downsides to my idea… I hope to hear different views and counterarguments from others.


#2

This is a very interesting topic and we’ve hashed it pretty thoroughly, so I’ll share some of the treaded path, at least the takeaway that I’ve accumulated.

Basically, the price of safecoin is determined in the market of personal choices out in the world at large. It is not tied to any factor on the SAFE network directly.

The price of resources to be purchased from the network is determined by a network algorithm designed to keep network resources balanced. If there’s lots of extra resources available to the network, the price will be quite cheap. If it starts to get scarcer, the price will move up to balance it.

As safecoin gets more valuable, the desire to get it through farming will increase, which should provide plenty of incentive to provide excess resources to the network, in exchange for safecoin.

We went around quite a lot on this topic in June in the topic “Price discovery for purchase of network resources”.

It was a long hashing out and might be worth reviewing, though it might also be painful to slog through.

I’ve quoted below a sort of summary-of-my-understanding type post which David Irvine acknowledged as being on the same page with his thinking. It’s post 97 in that string, if you want to explore the context.

(Sorry, lost formating, had to make due)
fergish
17 Jun 2014

(quote)
"dirvine said:

“If every user gets free storage up to a value, lets call it network average (which does increase). Then it seems fine. If you have 4 1Tb farmers you have plenty of safecoin to pay for your account being very large as you earned and did not need to purchase them”
(end quote)

(My summary)

My vault/farmer nodes and my User Account are two very different things.

So you’re saying that I get plenty of safecoin sent to my User Account by my nodes running, so I can use it to buy more storage cap “from the Network”. The Nework has some sort of algorithm that it uses to determine the price per unit storage, of which I can buy as much as I want or am able to. This would then be assigned to my account by some sort of Network validated tag that enables me to store that much. The collected vault network keeps track of this.

So the price per unit storage is determined by an algorithm that the network executes in an attempt to keep the network in harmony, not any bid/ask sort of market. It’s sort of an internal supply/demand calculation rather than an external one involving people. Safecoin is required to buy, at whatever rate the network algorithm sets, which safecoin is buried for refarming.

Are we tracking?
(end quoted post)
[he replied that we were tracking]

Hope this helps.


Economics of paying app devs for GETs
#3

Thank you. I’ll have to sit on this for a few days and probably review the entire topic which I hadn’t been able to find on my own.

This part sounds suspect to me. So what do people on exchanges do - mindlessly make up numbers and trade hoping it’ll go up or down?

But the algo itself is man-made and should be (in my view) largely impacted by the current state of the market.

(You don’t have to reply, I just rehashed my earlier questions and I’ll have to think about the info you gave me. )

Perhaps it’d be interesting to run some simulations of what happens with these prices in various scenarios (algo scenario, market-determined supply/demand scenario, etc.).

I know the network can react to changes quickly (matter of minutes), but if the price doesn’t follow, then tail-ing the logs could become a lucrative method to predict price movements. I just need to build a log collector that sends this info to my Big Data cluster in the cloud where I feed it to my high frequency trading platform that frontruns the algo… I could even pay for that info (in Safecoins, of course).


#4

Correct, as with every other crypto coin!

EDIT:

This is not the evil HFT (high frequency trading) we’ve been seeing talked about - that is something more akin to insider trading (basically they intercept an order, go buy the stuff before the order reaches the market, and sell it to the person placing the order at an inflated price). Its pure parasitic theft based on no useful function.

What this is, is a normal, useful market process, akin to arbitrage, where people are trading on small differences in price that corrects those differences, causing a matching of prices across different exchanges/markets. The benefit is that people coming to one market can trust the price, and end up trading at similar rates to those in other markets. Its actually what makes market manipulation hard if you think about it.


#5

There’s a plan to do this I believe prior to launch which hopefully will sort all this out. I too struggle with this aspect even though its been explained to me - there still seems a bit of an oranges/apples situation to me, but I’m also pretty sure if it makes sense to the devs and all the techies on here then it’s most likely a case of me just missing something. (No need to explain futther - I’ll just wait for tests).


#6

Okay, you added some more content to your reply so I had to undo my like, sorry about that :wink:

Hmmm… Does that mean that you are now supporting my view that the actual supply/demand for useful work on the network will in fact lead the price on exchanges?
What is then the intended purpose official black box “price”? To provide an AM / PM fix?


#7

@janitor

@fergish stated what I believe is planned and seems sensible. I was answering your concerns about this, and saying that what is proposed should work fine (i.e. doesn’t create HFT scenarios or anything else that concerns me).


#8

I read about 40% of messages from that monster topic. Wow… My head will burst…
I noticed some guys like @Al_Kafir, @Traktion and @stuartcharles (especially his point about sunk costs) had similar ideas.

I must say I’m still not convinced.

Here I’ll jest mention some scenarios/arguments that I haven’t seen yet:

  1. Capacity should be priced at the margin. Bulk prices changes for capacity will spread havoc among existing users. It’s impractical (see: sunk cost) for anyone to say “Okay, now the price is 17% less, I’ll power off that sucker”. Or, if it is, it’s going to be even worse, because he actually will power off that sucker (and power it on just before the chunks are declared lost, to “register” as alive again)
  2. It may be possible to manipulate the price by making sufficiently large disturbances in the market. For example I organize a “Shut your Safe farm” event where users pool to (a) load up on Safecoin contracts for difference, and (b) synchronously power off their storage. Depending on how the black box works, that may or may not work. They could even be able to sell & profit (maybe even with doing the same in the opposite direction and power their units back on without losing revenue their chunks. While such campaign are currently hard to organize, few months down the road that won’t necessarily be the case.
  3. Related to 1) above - some formulas that were suggested in that post took into consideration “network utilization” (in terms of required capacity). If a farmer accepts to take care of X chunks of data at the current price over 1 year, then 2 weeks later the price craps out and he’s told now he’s spinning his disks for 50% of the price, few people will want to farm (or the churn will become enormous and network charges much more significant). In some other post here I mentioned how cache nodes won’t get paid for network and storage charges (I don’t know if cache will be stored only in RAM or also on actual storage, in which case storage should be paid for as well) so while I can’t even imagine how this will play out, I’m now even more interested in my idea that the network shouldn’t do anything in terms of adjustments, just output raw data (SAFEstats) and let traders, farmers, exchanges and individuals figure it out for themselves.
  4. Because of 1) I think we’ll see people switch between different networks a lot (SAFEnet and others) so if we have formulas that create some illogical situations (example: modern central banking) the volatility will be even bigger than normal.

There are way too many variables to consider. My current thinking is: the less is done trying to figure all this out, the better.


#9

Yeah, my head DID explode several times in that thread.

There are a number of variable factors, and it can be really confusing to talk about or even consider till definitions and categories are sorted out to a bit. Here are a couple things that seemed to stabilize the confusion for me.

We all want safecoin to be more that just a network token for buying storage. If it is just a resourse-buying token, there wouldn’t be much incentive to farm for it. So pegging it directly to a certain amount of resource wouldn’t make sense.

The ideal is for safecoin value to be market driven, such as bitcoin. We can expect volatility for a long time. As an example, bitcoin is just now showing some signs of stabilizing, though I suspect it’s got a long, long way to go. Bitcoin was worth nothing till finally someone accepted 10,0000 btc for two pizzas. Safecoin will be similar, except that it will start out with a minimal value of some specific amount for some storage cap on an account. But it’ll be pretty cheap! And continue to get cheaper over time. This is because putting resources on the network is a no-brainer, especially as storage devices, etc., get cheaper and safecoin gets more valuable.

One key for a node to be able to generate the most safecoin is being consistently online and dependable. So switching off will never be an advantage to an individual node.

Yes, humans will have to figure out the algorithm that the network uses to determine the safecoin price for increased storage cap, but I doubt that will be as complicated as one might think. Think of the shutoff float in a toilet reserve tank. If a super lot of resource is available, the price is low and people will buy up more. If the percentage of available resource (i.e., storage) is getting scarce, the price goes up till it starts to balance. (Setting aside attack vector consideration here, which should be looked at separately.)

The thing that helped me get this from the node view is that when you put your resources on the network, they are the network’s resources, not yours anymore. Yeah, you can shut them off, but that’s all you can control about it. Basically you are exchanging some portion of your computer and hard drive for the opportunity to earn safecoin. If you wish to sacrifice your earning potential, you can mess with the equation by withdrawing what you put in. But the marginal cost for keeping it there should be worth the difference. Can you earn a living as a professional farmer? Maybe. In some ways, I hope not. But the barrier to entry is very low, regardless, so participation is designed to be very broad rather than highly centralized like Bitcoin.

This is a confusing topic because “value” is a bit phantom because it’s largely opinion. Then there are several interrelated values that play against each other. Then there’s the problem of where the rubber meets the road with a solid price, but that’s related to concepts of value as well.

Anyway, this was more rambling than I’d planned, but hope it is useful. :smile:


#10

Yes, it does all seem really complicated - where I was confused was, I thought both resource price and Safecoin price were trying to be melded together somehow. As these are both derived by different means using different sources/inputs, I noticed that this was not feasible- an apple and orange situation. Currently, I’m in agreement to an extent with Janitor, in that Safecoin price does not need any kind of algorithm - the market will decide. The resource price I think, should just be a % lower than the mean/average…whatever, of all other providers - just keep resource price competitive…somehow…lol
To my mind it doesn’t matter whether you are buying resources, sausages or table-lamps with Safecoin - we don’t need any special algorithm to relate Safecoin to sausages do we?
I’m a decorator, so don’t rip into me too much…lol


#11

Thank you and @Al_Kafir for chipping in.

But if this sunk-cost capacity is not priced at the margin but “retroactively” (on the entire “float” of the toilet reserve tank), people will have to constantly re-evaluate their decisions and what may be feasible one day, it may not be 10 days later. And if this repricing happens all the time to the entire user base, that could “cost” a lot in churn and “re-chunking”. Since the bandwidth seems to be free (on SAFEnet) although it’s not, the extra rechunking could put undue stress on to that resource.
For example someone who could “normally” (I know…) store 1TB a month and “serve” 2TB of data because of the extra network churn caused by others may be able to store 800GB and “serve” 1.3TB which means he’d effectively subsidize this imperfection in pricing mechanism.

I think this is a very complex issue that’s not going to be solved in a “good enough” way in 2014.
I see constant “fine tunings” happening as the Devs learn about the network and how people use it and I’m slightly concerned that revenues aren’t distributed to caching and network nodes (I don’t know whether they should be, but it bothers me when something is free - that’s never right and it’s guaranteed to get abused :smile:).


#12

Ok, too technical for me, over my head…I’ll just go and paint something I think…


#13

To me the simple way to think is, the price of safecoin in $ or fiat will be 100% market driven, the farming rate will be dictated by the network. When it needs more space it will offer more safecoin, when it needs less it will reduce capacity (likely never to happen). This is like mining gold or bitcoin to an extent, easy to begin with and getting harder. How hard is being set at rate of network utilisation and reward.

The network needs to reward farmers and this is a good way. Pegging process or similar will be beyond it, but rate of reward is certainly within its remit.

I think there is no solid mechanism for rate of printing money, so one that is linked to a valuable resource seems to work. It did for gold for a while anyway and this time we hope the farmers will be the guardians of the worlds data, until we can get farmers built into everything. So you get a fridge for free, because it has a farmer in it, who knows, we get satellites paid for because they have farmers in them. The could charge for access until a cheaper one exists etc. I wrote about this a while back http://metaquestions.me/2014/02/19/machines-that-own-themselves-in-bitcoin/ and here http://metaquestions.me/2014/03/02/machines-that-own-themselves-in-bitcoin-part-ii/ There is a long way to go, but the value of true digital wealth transfer as a means of reducing costs and increasing efficiency has not even been touched AFAIK.

Anyway, the market decides the fiat value and the network decides the farming rate. Unless we can come up with a better scheme then this one seems very solid. Unless … You know I am always listening :smiley:


#14

Thanks for the clarification.

I have a question, though, on the above quote. This does not include one more factor, which we did go over, and I believe you put forward on at least a theoretical level. That is the price setting mechanism for purchase of storage resources from the network when someone wants to raise their cap. Am I not correct that this is also determined by the network, probably as an algorithm paired with the one governing farming rate?


#15

Yes that is right. The farming rate will be determined by the network as a whole. So we should be able to see what it sets and maybe publish that over time. It will tell us if we ask it :slight_smile:

I see it something like (made up numbers for example)

Day safecoin per 500Gb per day
1 1000
100 800
1000 200

So as the network grows the space is cheaper, but if safecoin value goes crazy (it might) then it may look like this

Day safecoin per 500Gb per day
1 1000
100 20
1000 3

OF course it could go the other way if the network needs more space. In that case it will increase farming rate, and take no notice of outside price, it will notice the oversupply if its too generous and reduce it quickly or the other way around. In essence balance of rate will be the networks job, but in doing so it ensures space/cpu/bw etc. are available. The price just means more folk will farm if its valuable and less if not, if it is less attractive the supply of space drops, the network increases rewards etc.

Hope that is what you asked :slight_smile:


Safecoin issuance algorithm
#16

Ultimately, there is only ever a single spot rate, at a given time, even at a gold exchange. Mines don’t get to set the price of dug up gold, prior to it going to the market.

Ofc, this doesn’t stop futures markets selling contracts at a fixed rate over a period. Perhaps such a secondary market will spring up to reduce the risk of farmers? I suspect if there is money to be made doing this, someone will provide such services.


#17

Thanks, David, but no, that 's not what I asked.

I’m looking at the mechanism for the network setting the price (in safecoin) by which one can increase access to resources (storage space) by purchasing it from the network–i.e., raising your storage cap by paying safecoin to the network, which safecoin would be recycled to be farmed again.

We discussed it previously under “Price discovery for purchase for network resources”. You definitely laid out that the network can and will be able to determine the cost per unit cap increase, and use this as a tool (along with farm rate, I’m sure) to keep the network resources balanced.

Did I misunderstand? I hope not, because I just posted a thread “Vault vs user account – an over view of user experience” which relied in part on that understanding. If you read that, you’ll see clearly what I’m talking about as separate from farm rate.

(Man, I hate to distract you from coding, so if need be ignore us here as we knock this stuff around.)


#18

Yes this is what I mean when I say the network will increase and decrease the amount of safecoin for farming. I think this is where I keep misunderstanding I call it quantity of safecoin and you are calling it proce of service in safecoin.

So on the farmer side that is OK, on the client side I see it as the network will charge for access at a rate. This rate can be set by us, but my preference is we use the global farming rate instead. So if farmers get paid X safecoin per unit of storage (which is forever) then a user pays X for that unit (forever). This rate will be a network rate though and base don the farming rate around a user, it should be strikingly close to the actual network average if you stopped all vaults and counted them.

So yes you are right and I was explaining farming rates again :-0


#19

No worries on that front I tend to look here between builds :smiley:


#20

I think I get it!! And I see why you felt you were answering but I didn’t get it.That makes it rather elegantly simple, actually.

It will be interesting to see how these details present themselves on the user interface end.