13 Steps to Enlightenment
falsifying stock-to-flow — the final part
my investigations are probably a bit autistic and “too deep” buhu, because I´m a poor student. I don´t hate YouTubers and Crypto-gurus (they are entertaining, don´t want to miss them) but those people don´t rely on their predictions. They earn fiat money. Being wrong is not an existential threat to them. So if you are an (probably young) interested person (older persons with good status don´t care for deeper research) and you find something wrong in this article, then burn it with fire. The following is not financial advice! but as long as you are no social -media star, your portfolio is depended on verifiable facts and not on nice narratives. So don´t hesitate to correct me.
Introduction to the last act
In Part two of my trilogy I wrote about the history of valuation-models for Bitcoin. Not all, but some models were based on scientific studies. One class of them are the network-effect models: For many communication networks there is a correlation between number of users (n) and the market-capitalization of the corresponding stock. The relation is described by network-effects.
>>Note that the following is dynamics and not statistics!<<
1) Intrinsic Value Mr. Buffet
The fair value of a communication Network — what science says:
The number of users (n) determines the value (V) of the network. The relationship follows either Metcalfes-”Law” (V=n²) and sometimes Reeds-”Law” or Odlyzkos-”Law” (V=n log(n)). *Those laws are hypotheses and not laws in the scientific sense.
V=n log (n) grows much slower than V=n² or n/2 * (n-1) which is exponential rise in value. It all depends on the initial size of the network.
Network-economics:
Star-Investor Warren Buffet said, there is no intrinsic value to Bitcoin [1]…
Metcalfes-Law is validated for social-media networks like Tencent and Facebook [2], for Trustnet and the US-Dollar and even for Crypto-Currencies like Ethereum, Bitcoin and Dash[3,4,5]. It is a simple relationship somewhere between Value= 1/2(n)*(n-1) and n². If you wonder, “then why Facebook´s value is not 2.3 billion squared?”…well, it is the number of daily or monthly active unique users and not accumulated accounts and no, Facebook is not one network but many small networks — effectively there are language-barriers (For details: see [6] Zhang et al. 2015; [7] Paterson 2019).
“There is compelling evidence that suggests that the growth and price of bitcoin and other cryptocurrencies are likely to proceed according to a relatively straightforward mathematical model similar to the growth curves of Facebook and other networks” (Paterson 2019)
2) Bubbles — the events on top
The “bubble” is any deviation of the null-hypothesis (base trend), described by the network-model (Alabi 2017). Bitcoin is at its very core → a communication network, made possible by a communication protocol. Not a damn stock! A financial transaction is a very primitive form of communication, where there are no strong language barriers.
You can call it: boom = over- and bust = under-valuation. This implies that there is a fair or intrinsic value.
3) Value — just what people think?
A) Yes, value is what people believe/expect…
B) but the believe-system of animals and humans adapts to hard properties of reality. Believe is not pure imagination. [Matching Law]
For example: People invent `dead fish´ and believe that it is a great store-of-value. Suddenly the fish starts to rod and immediately people adapt their believe to the new available information (it is called learning 🤫).
Philosophically you can debate whether the value is really “intrinsic”, since it is observer dependent (epistemic) and not observer independent (ontological)… well, it exists only in the domain of human knowledge but it corresponds to ontological real/physical/hard properties of the valuable asset:
- non-inflateability/ unforgeability
- durability/robustness against external stressors
- connectivity
- …
In short: some utility which helps you or peers or the community to maximize evolutionary fitness (Expected Utility Theory).
The Efficient Market Hypothesis
We all know, that humans can be wrong, very wrong about certain properties of an asset. Driven by strong emotions and underlying cognitive biases we tend to be naive about reality. FOMO and FUD dictate Boom and Bust Cycles and still, markets are really efficient.However efficient markets can still be wrong.
Bubbles are driven by network effects, but those effects are an additional layer on top of the technical use. Bubbles are speculation driven.
Bitcoins price explosions are not exponential, they are super!-exponential. Within finite time → the price approaches infinity. Of course it will never become infinite, but it explodes towards infinity. This is called a finite-time-singularity [8].
4) What ALL Crypto-YouTubers get wrong!
The standard model for population growth — Malthus’ model describes exponential growth (*population could be anything e.g. people, price, bacteria, speed, mass,…) as follows:
This 👆 is exponential growth. When r=2x and the time period is one month, than price is 2x after one month and 4x after two month and so on… BUT this is not what really happens!
Super-exponential Price Jumps
Exponential is an understatement! In reality is much more extreme:
Whenever a Youtuber says “Bitcoin has shown exponential price movement”…he/she has NO clue what exponential means. When the exponent (delta; δ) is slightly positive, which means bigger than p(t)¹ (which is the normal — “no dependency on size of p”-state of the Malthus model), then the function explodes in finite time. Of course a delta of 1.2, grows faster than a delta of 1.1, but even 1.000001 is sufficient that the function explodes in finite time. This is called a finite-time singularity.
Too fast to react!
Black holes are finite-time-singularities, financial bubbles can be finite-time-singularities [8] and of course nothing in this material world gets infinite. Before it gets there, it collapses. However , infinities are there in the dynamics of the system.
But forget the math, just look at the acceleration:
The time period for price- (population)-doubling gets smaller and smaller → approaches ZERO, so that the price in a few seconds would be @ infinity . So, when the halving-time goes to zero, the price approaches infinity. → It skyrockets. 🚀
Why it necessarily has to collapse — and why it does not matter…
Whenever a YouTuber comes with reasons for the crash, you should now understand, why this is absolute meaningless. Any growth-dynamic approaching infinity is unsustainable… so anything, ANYTHING would cause the pop. It does not matter if it is the China-FUD or the Korea FUD, there is no meaning in the external trigger, it would collapse anyways. The system internally matures towards instability [9]. Like a drop of water getting bigger, until it reaches a critical level an drops down. Like a ripe apple on a tree… does it make a difference, whether it is a happy bee sitting on the apple, or the rain, or the wind causing the apple to fall? Obviously not.
5) The bigger picture — why we will win
Bigger trends beneath the bubbly surface
Network-Value is the very basic (utility derived) value of a communication network like the internet or bitcoin or even more complex networks like social medias. No matter how crazy the bull-runs and crashes are, in the long run, a one-billion users network will be at least as valuable, as the utility the users derive from the network. A one-billion-users-network will never be worth nothing. Any useful tele-communication network has at least its network-value. Of course, in the short run the panic can lead to an undervaluation same as the hype can cause a massive over-valuation. In early 2018 Sornette et al. have ex-ante! predicted the level Bitcoin was collapsing towards at the end of 2018 [10].
And now people come with stock-to-flow 🤷♀️ 🤷♂️?
Stock to Flow (S2F) is simple the ratio of what is already there (stock) and what is left there to be mined, or more precise the flow. So, rareness and high demand should be a good thing
6) Rareness vs. Scarcity - Why Crypto and not Krypton?
Now, we all know that rareness and scarcity is not the same. Not everything which is rare, has a high value. Whale puke or shit or what ever it is, the black amber is super rare poop of whales 🐋 💩, chemically modified by sea-water over the years. It smells 👃 good, hence people want it for perfume 🤳🏽. But today we can synthesize and forge it. So it is not that “hard” anymore. Back than, it was traded against gold and eunuchs. 💁🏻 💁🏻♂️ 💁🏾 💁🏾♂️
As introduced in the first Part, tulips 🌷 back then were very rare, because the tulips of interest were modified by special viruses [tulip breaking virus], so it was not just about dumb flowers and onions…sure today a special tulip bulb is not that “hard” anymore but back than it was totally rational to use it as a store-of-value and speculative vehicle.
So, it needs to be rare + unforgeable. You cant economically efficient counterfeit bitcoin or gold. Many other rare elements like krypton or xenon, can be synthesized 👨🏾🔬 ⚗️.
Rare mutations 🧬 are also unforgeable and hence can be very scarce… do you want some? Well, of-course some mutations are traded (e.g. rare animal breeds 🐶 🐱) and in the 16th century for kings and dukes, it was desirable to have people with rare mutations around, those were highly rewarded positions but not every mutation is of interest and of course we are living beings and not goods.
7) The Alchemy behind Bitcoins Hardness
So seriously, why bitcoin, why gold? When we have clay…
Of course, plutonium and uranium have high value but are inconvenient as everyday money ☢️… it is about convenience. We all know that good “money” is not only hard, easy to mint and hard to forge but also portable, dividable and easy to store (durable).
After we had a silver-standard and a gold-standard, we went to materials with less material value than nominal value. In Japan during the last phase of the second world war, the government made the money less and less material and at the end they moved over to clay-coins (which were never used because the war ended)[11].
Humanity has learned to deal with such “money”, but still, gold and silver are very widely used for storing value. Russia once tried to establish a platinum-standard — platinum is not clay ☝ ️it is valuable… but they never succeeded, because the minting process is way more inconvenient than minting gold-coins. But the biggest reason why there is only one, sometimes two winners, has nothing to do with the material, it is because of network externalities.
Bitcoin a Gold-Mimetic — Metallism out of the lab
Bitcoin is a gold-mimetic, it behaves like gold, is some sort of synthesized artificial gold. But hey, clay or paper with flexible-supply can also be as valuable as a metallism-based system. Nominalistic clay-coins are just not store-of-value (so what?) but they are not fiat systems either. A decentralized inflating crypto-currency is not centralised-fiat and even if it was, fiat is not worthless. However, it is OK to be fixed supply and rare:
So “rare” is not sufficient, it needs to be scarce (rare + high demanded). However, it is in the human nature to demand rare items and trade it for high prices, hence rareness can provoke scarcity. When scarcity can not be addressed with forgery and synthesis, you have an unforgeable costly good.
Bitcoin is “grand complication” — the Rolex Problem
Watches like Rolex are limited edition watches but easy to forge, so watchmakers build in “proof-of-work” by introducing so called “complications”.
Of course the replicas get better and better, many Chinese manufacturers these days can produce Swiss-Made-grade tourbillions, which leads to price decrease of Swiss-Made-grade watches (Federation of the Swiss Watch Industry 2010). Using expensive materials is as secure as make a watch “limited edition”… the only secure way to safe the extraordinary status of high price watches is…
…introducing more complexity — “grand complication” 🙌
(moon-phases 🌕 🌖 🌗 🌘, 1000 year-calendar 🗓, complex tourbillions 🌪, build in melodies 🎼,…) as watch makers call it. So, to a certain degree the labor theory of value holds. Some of the most expensive watches have up to 57 complications.
(Un)fortunately, complications are of no economical value. Who uses a moon-phase calendar?… 🤨. Bitcoins proof-of-work is also not coupled to any useful work. I am not sure if Nakamoto deliberately chose useless work, so that Bitcoins value cant be associated to the economic value of the work, which could have been used as a source for pricing… However, in crypto there is also useful proof-of-work. It existed long before Bitcoin. Instead of brute-forcing against a Hash, you could use protein-folding [12]. Both brute-forcing and finding solutions to protein folding belongs to the class of embarrassingly-parallel problems. The only difference is that protein-folds or rendering is of social and economic use, while mining is just costly-signaling or conspicuous-consumption. Probably it makes sense to use useless-work.
8) Which Parameter to use for Prediction?
Network-Externalities
Lets view gold as a medium for communication (transactions are communication), so → gold is its own protocol. The more people are using gold, smartphones, Facebook, Bitcoin …the bigger the positive externalities of using it and the bigger the negative externalities of not using it. So again, the number of users is the most important parameter of valuing a network asset.
Scarcity and S2F are “Wired Proxies”
Nakamoto hoped for a “positive network effect”. If he was right, then stock-to-flow and number of users should go hand in hand. One could say, that stock-to-flow is a proxy and not really the predictive causal parameter. But this is only half the story, since scarcity gives birth to knew users.
9) Swallow this A…Austrian!
Whenever there are discussions about Bitcoins value, most people immediately come up with the fixed supply of Bitcoin, they think “Austrian”. The Dollar is the most valuable currency the world has ever seen and it has potentially “infinite” supply…
so it is valuable because of networkexternalities…scarcity is important for store-of-value properties but not for value!
Even Satoshi Nakamoto himself didn't wanted a fixed supply, but since there were no decentralized multi-party-oracles and no smart-contracts to match supply with real-world demand,… he made it fixed (see previous article). Satoshi was not able to build Ethereum, → so he built Bitcoin instead.
„because I don’t know a way for software to know the real world value of things. If there was some clever way, …the rules could have been programmed for that”
Nakamoto knew about the positive feed-back-loop BUT what the Austrian-narrow-minds want to overlook, is the fact that Nakamoto also knew about Metcalfes-Law. But today we know, that Nakamoto was naive about scaling-issues!
“As the number of users grows, the value per coin increases. It has the potential for a positive feedback loop; as users increase, the value goes up, which could attract more users to take advantage of the increasing value.” (Satoshi Nakamoto 2009)
10) Digital Gold narrative — dream or reality?
The good thing is, that when it comes to value, the network-effects are not dominated by the number of transactions, but by the number of users. So even if nobody pays his/her Starbucks coffee with bitcoin, it can still have really enormous value!
In his article, Woobull concluded, that a store-of-value needs to have the following properties:
“(1) Security
(2) Credibility of fixed monetary policy
(3) Liquidity / Lindy Effect / Infrastructure / Ecosystem
(4) Governance / Adaptability / Community / Game Theoretic RobustnessI can rewrite this as:
(1) don’t break on me
(2) don’t cheat me in the future
(3) bring more money in
(4) look after the above”
[Source: “Utility is a red herring, look for organic store of value” (Woobull 2019)]
Let´s rewrite this as
Store-of-Value IS Utility
(1) Attack resistance
(2) Szabos Law (“don´t change Bitcoins protocol”. Hint: supply can be changed)
(3) Metcalfes Law: high number of users (implies scarcity)
11) Szabos Law was already broken — TFW supposed Gold changes its policies
which Bitcoin do you mean?
Stock to Flow is definition agnostic! It says “Bitcoin” will be worth x-Dollar
Hey but which “Bitcoin” do you mean? No, I don´t speak about BTC vs. BCH, vs. BSV vs. my ass coin
We speak about THREE different Versions of BTC!
The original Token held by Satoshi Nakamoto are based on the P2PKH-Standard, the addresses start with “1” (so called legacy addresses)…suddenly Bitcoin-Cores “SegWit” — a softfork (Protocol update without chain fork) came around and changed a little detail… 🤭.
Have you ever seen your coins with your own eyes? 👀
Now there are also nested SegWit addresses “3” after P2SH and native SegWit “bc1”…guess what most people hold?…right SegWit´s “bc1”-token, which are not down-compatible to the original P2PKH-Standard.
Well, SegWit is more convenient, but guess what happens in the case of a hard fork… → Those represented on the legacy-chain (people like Nakamoto) get both versions and can dump one. We, the ones holding the new-bitcoins…can only be dumped. Sure it is “unlikely” but Bitcoin is not Bitcoin.
but there is another technical problem…
12) The “no-need-for-scaling fallacy”
Unfortunately Nakamoto consensus does not scale above a ceiling-constant K, this is why Ethereum abandons the old paradigm, you can tweak some parameters like Block-size but this does not solve the fundamental issue of the need for long block times of around ~10 minutes (the reason you find in Part 1). Shorter Block-times trade speed for security. NO THANKS!
Fees are as high as people are willing to pay!
So with this limiting constant K and an increase in demand → fees shoot up. In December 2017 fees were as high as $60 and on Ethereum when demand was really high, people paid as much as $6000 for a single transaction!
So tell me Austrian, what do you think would happen when the current state Bitcoin network ON-chain processes halve of the FX-Market (2,5 Trillion USD per day). Big transactions and really high fees. So high, that normal people could not afford to use the network. You could never ever move you money again. On September the 6th in 2019 a whale moved 1billion USD for $700 fees, and today we have seen a 1.1 billion USD transaction for $83 . This is cheap, they easily would pay $6000. Supply and Demand.
The Scaleability-Trilemma is real!
Lets summarize so far: despite bubbles, trends are real and can be verified. Networks have intrinsic value. Since the value is dependent on adoption, technical hurdles can play a significant role. Even if the problem is something “simple” as store-of-value.
13) The Statistics Part
we know the story: PlanB came with a model fit of r²=.95, Nick came with a cointegration model, supporting the “predicition”
Now people come and mix up many different things.
Longterm predictability in markets is of course never possible. NEVER!
Markets are in the Domain of fat tailed Distributions
Statistics is based on the law of large numbers. The law of large numbers says that our sample mean (the statistical observation) converges to the real mean in probability.
For some classes of distribution, convergence to the real-mean (to reality) is so slow, that one will never ever observe convergence in real time!
In mathematical terms: If there is a sum of random variables X1,…,Xn with finite mean m, that is E(X) < ∞ , then:
…this term above, which describes your sample mean, converges to m (the real mean aka reality) in probability, as n → ∞ (as your sample size aka necessary amount of observation, years of market data, what ever “gets bigger” or more correct: approaches infinity).
Here is why Stock-to-Flow is probably right anyways -
The Black Swan Stuff
You cant predict markets, you should not try BUT trends none the less are real! Two different things hard to comprehend. Trends can be real, but you cant predict. Well, even if you can, you really cant.
In math it is called a ruin problem: even if you can predict most of the time, you should not try to >>rely<< on the prediction. As time moves on, the ruin probability for a single individual (time-perspective) goes to 100%. (Peters and Gell-Mann 2015, Taleb 2014)
Even if the chance of ruin is very small, by trying to predict, you will hit a absorption barrier, you will go bust -100%
Imagine it rings at the door, you open and you see a cloud of dust. Suddenly a Grunman F-14 Tomcat appears…
it carries a M61 Vulcan machine gun but only one bullet, 10000:1 Chances of not to die. How many rounds would you play? The longer you are exposed to the game, the lower your chances to survive. Even if it is the same very low probability for each round, sooner or later you will hit an “uncle point”, an absorption barrier, go bust, extinct, “die”. Everyday-in-the-market means serial risk-taking
Ruin problems like this are “non-ergodic” processes. In the ensemble perspective (the group) chances of 10000:1 for the player, look great. Go with your friends to a casino and most will not “die”. Since you cant assess other people, since your capacity as one person is n=1, it is a completely different game.
Even IF you have a model which is predictive 9999 out of 10000 times, hitting the bullet @round two → will end your life.
And there are many possible bullets. This is a really fu__ing problem. You can handle Bitcoin as a asymmetric bet, low exposure high possible gain…but for most of us 10000000% of nothing is still nothing. We need high exposure to make a difference. We don´t want to swallow the Warren Buffet pill. We know that most of the out-layer guys started their career with high risk-high reward investments (“Dempster Windmill deal” huh Buffet you ´member? ´Member the “wringing out businesses” style?). All this compound-interest sh*t is sensitive to the starting conditions.
Enlightenment — Nothing here is financial advice!
The Co-integration Data of User Nick is interesting, no it is excellent work! but I would not consider it a verifiable fact. The efficient Market Hypothesis states, that all available and verifiable information is already priced in. Verifiable is the number of active users and other network parameters, but projections aka the future of non-linear systems is not verifiable.
Lets assume that S2F is real: since it is not verifiable, → it cant be priced in, as long as another verifiable non-proxy is not verified yet. In the very moment users (miners, banks, private users) adapt to the change in scarcity, which is a multimodal and multifactorial process, the efficient market prices those facts in, an efficient market ignores any change in Bitcoins Protocol behavior but accept any change in the physical “substance” of the network.
Even if markets are populated by efficient/rational agents all the time, still the dumb money covers the efficiency with huge piles of sh*t.
So to summarize the fallacy of S2F in one sentence: Price does not increase because Bitcoin gets more scarce… Bitcoin gets more scarce → this attracts more users and speculators and incentivises miners to keep on mining, hence the value increases (because by matching-law participants adapt to the new verifiable state)... ergo the price increases, since this is a causal-relationship, of course it is ex-post “pre”-dective…I would rather call it post-dictive.
Literature:
[2] Zhang, X., Liu, J. & Xu, Z. Tencent and Facebook Data Validate Metcalfe’s Law. J. Comput. Sci. Technol. 30, 246–251 (2015) doi:10.1007/s11390–015–1518–1
[3] Ken Alabi, (2017) Digital blockchain networks appear to be following Metcalfe’s Law, Electronic Commerce Research and Applications, Volume 24,
Pages 23–29, https://doi.org/10.1016/j.elerap.2017.06.003.
[4] Peterson, Timothy, Metcalfe’s Law as a Model for Bitcoin’s Value (January 22, 2018). Alternative Investment Analyst Review, Q2 2018, Vol. 7, №2, 9–18.. Available at SSRN: https://ssrn.com/abstract=3078248 or http://dx.doi.org/10.2139/ssrn.3078248
[5] Civitarese, Jamil, Does Metcalfe’s Law Explain Bitcoin Prices? A Time Series Analysis (January 23, 2018). Available at SSRN: https://ssrn.com/abstract=3107895 or http://dx.doi.org/10.2139/ssrn.3107895
[6] Zhang, XZ., Liu, JJ. & Xu, ZW. J. Comput. Sci. Technol. (2015) 30: 246. https://doi.org/10.1007/s11390-015-1518-1
[7] Peterson, Timothy, Bitcoin Spreads Like a Virus (March 20, 2019). Available at SSRN: https://ssrn.com/abstract=3356098 or http://dx.doi.org/10.2139/ssrn.3356098
[8] K. D.Ide (UCLA), D. Sornette. (2001): Oscillatory Finite-Time Singularities in Finance, Population and Rupture.
10.1016/S0378–4371(01)00585–4
[9] Sornette, D. (2009): Dragon-Kings, Black Swans and the Prediction of Crises. arXiv:0907.4290
[10] Spencer Wheatley, Didier Sornette, Tobias Huber, Max Reppen, Robert N. Gantner. (2018): Are Bitcoin Bubbles Predictable? Combining a Generalized Metcalfe’s Law and the LPPLS Model. https://arxiv.org/abs/1803.05663
[12] Buterin, Vitalik. (2019): Hard Problems in Cryptocurrency: Five Years Later. Link.
