rpietila Wall Observer - the Quality TA Thread ;)

[AnonyMint]

How can you assert that a fit with one model is lesser fit than a fit with another model? Define 'lesser'?

The best fit is when you have a better R-squared value than any other fits. Excel calculates the best fits for every model automatically, so you can just conclude that a log-linear model has a better fit (0.94) than log-logistic (0.73).

If I am not mistaken, the best R-squared (least error from the data points) would be an N-degree polynomial for N data points such that the curve passes through every point.

Thus 'best fit' may have no correlation to predictive power.

Surely you of all people understand the concept of overfit...

Exponential growth is not some "arbitrary function."  It is the solution to a very simple--and very meaningful--differential equation.  It occurs whenever the growth rate of something is proportional to the size of the thing that's growing: e.g., the population of bunny rabbits in a park, bacteria in a petri dish, or users of a social networks.

My mathematical point to Risto is that the relatively lesser fit of 0.73 for the log-logistic exponential model is entirely meaningless because precisely the choice of model is what matters. So aminorex and Peter R have supported my argument.

P.S. kudos to Peter R's generative essence abstraction.

[1714468153]

1714468153

[1714468153]

1714468153

[1714468153]

1714468153

[aminorex]

My mathematical point to Risto is that the relatively lesser fit of 0.73 for the log-logistic exponential model is entirely meaningless because precisely the choice of model is what matters. So aminorex and Peter R have supported my argument.

Not only implicitly but explicitly as well, yes:  A bad fit of a structurally representative model is far more predictive than a good fit of an unrepresentative model, because it is a compressive attractor.  There are some statistical tests you can use to discriminate, but knowing structure beforehand gives you a lot more leverage than structural estimation after the fact, because noise.

There are certainly some intuitions to support log-logistic, but they aren't compelling to me. Still have not taken time to think this one through properly, after all this time.  Summer has been crazytown.  And I've been indulging myself with distractions.

Yet, events in the world will never be entirely modeled by anything so simple.  COIN will be a liquidity event at a magnitude far larger, relative to current market cap, than was the advent of exchanges - which precipitated the superbubble.

[AnonyMint]

while I agree that number of users is increasing exponentially (so far), the number of transactions is NOT, which was posted earlier.
Possible explanation-the absence of accurate statistics, but then the whole argument is mute.

If it is true the "number of users is increasing exponentially (so far), the number of transactions is NOT" this would be congruent with my thesis that Bitcoin is abnormally skewed towards investors and not sufficiently balanced to (transaction) adopters.

Today it appears that a "concave down" [log-logistic] function would fit better than a straight line, but late last November people were arguing for a concave up (super exponential) model.

We have more history now so what we see now is more supported (by data) than what they saw historically. That doesn't necessarily make it more predictive, as it depends on our characterization of possible noise and other supporting logic for which model should be predictive.

IMO, the growth rate for bitcoin would need to decline strongly (or retreat over a long period) to truly invalidate the exponential growth model.

Your subjective bias is fooling you here. Mathematically false because there was never a mathematical confirmation that the predictive model was correct in the first place. We are simply picking straws blindfolded unless we have historically complete examples and or some other well supported logic from which to argue which predictive model should be correct.

Here's a model that fixes bitcoin's market cap at inception at $500,000.  The "rationale" is that Satoshi spent approximately 2 years building it, and the market-value for Satoshi-level talent is $250,000 / year.  

I argue this proposed model is invalidated because it is internally inconsistent. It is based on the notional of investment value of market cap as a starting point, yet you are modeling the Metcalf Law adoption value which is based not on investment adoption but rather on transaction adoption. In other words there is a category error here.

P.S. thanks so much for applying the effort for the examples and sharing. It really helps to discuss from your effort as I think it helps raise understanding.

[Peter R]

IMO, the growth rate for bitcoin would need to decline strongly (or retreat over a long period) to truly invalidate the exponential growth model.

Your subjective bias is fooling you here. Mathematically false because there was never a mathematical confirmation that the model was correct in the first place. We are simply picking straws blindfolded unless we have historically complete examples and or some other well supported logic from which to argue which model should be correct.
...
In other words there is a category error here.

Yes, I see your point and you might be right.  There's probably some more advanced ways to analyze the question of "what is the most accurate growth model?" objectively, as well as the question of "what would it take to invalidate the exponential growth model?"  But I don't have the skillset.  

For example, perhaps we could assume the adoption process is exponential Brownian motion [like the plot I showed earlier], estimate the hidden values for mean drift and volatility from the number-of-TXs-per-day time series, and calculate some sort of p-value for the fit.  But I'm way out of my depth now.  I wonder if there's anyone here who has a deep understanding of SDEs and things like the Black Scholes equation, etc.  

[AnonyMint]

That is far out of my depth (and available time to dig) too. I defer to animorex perhaps. Or perhaps SlipperySlope.

[AnonyMint]

COIN will be a liquidity event at a magnitude far larger, relative to current market cap, than was the advent of exchanges - which precipitated the superbubble.

I reiterate my thesis that the larger this COIN event, the more coins that get locked up in a vault and never transact. All the buying and selling churn for COIN will occur off-chain. Thus it removes transactions from Peter R's Metcalf Law model of the market cap. Whether you believe that model is predictive is another issue. In short, COIN could make a lot of investors happy and boastful, but it has a cost to upside growth. It is expedient but destructive. It is top-down not decentralized.

My thesis is fairly simple. The network effects value of Bitcoin is the decentralized potential for every user to transact with any other user. For example, imagine two off-chain entities that refuse to interopt (we are then going backwards, i.e. analogous to paywalls on internet urls, e.g. subscription magazines). As that is taken away with too much investment focus and top-down off-chain activity, the headroom of Bitcoin's market cap is declining, i.e. log-logistic. The fascist bastardization of the internet is well underway too. We hackers have a lot of work to.

For me it is common sense. Bitcoin is being sought primarily as an investment, not as a tool. The internet was sought primarily as tool.

I think we need a crypto-currency that is sought by the majority of the population not as an investment, but rather as a tool. The investors will certainly follow, just they did for the internet or how Warren Buffet admires rapper Jay-Z. We shouldn't totally dismiss what was learned from Dogecoin.

I hope readers appreciate I have shared my secrets to a large extent. We will never know who copied and ran with them. Was it me? You will never know but I will.   lol. Peace. (it also helps to preserve my life I suppose, don't cha think 'Satoshi' is reading this)

P.S. In short, fuck the SEC! (and every other alphabet soup top-down molasses)

[SlipperySlope]

Here is a graph of Bitcoin Hash Rate growth for the last couple of months as presented by Bitcoin Wisdom. I believe from news announcements that mining is not only now industrialized, but becoming more vertically integrated as manufactures operate their own miners during that crucial period where return on capital investment is possible.

There appears to be little correlation this year between bitcoin price and the mining difficulty.

[AnonyMint]

There appears to be little correlation this year between bitcoin price and the mining difficulty.

Mathematically there shouldn't be unless at least Moore's Law died. You want to correlate mHash/$ and mHash/Watt, i.e. those are both increasing faster than Moore's Law I think?

http://www.marketoracle.co.uk/Article47164.html

There was a move down on BTC-e, visible even from the long-term perspective. Bitcoin went below $500 in a first decisive move in days. The magnitude of the slide itself is, however, only secondary as far as the importance of the move was concerned. Why is that?

A look at the spike in volume is pretty much self-explanatory here. While 1,164 bitcoins were traded on Saturday, Sunday saw 41,590 (!!!) bitcoins change hands. A 3,473% increase, if you will. This activity has been blamed on a trading bot gone wild. This might be the case, but the most important short-term clue for investors here is that the price hasn’t actually completely recovered following the spike in volume, whatever the reason. Because of that, it seems that we’ve actually witnessed an important change in the short-term outlook.

The move below $500 on BitStamp and BTC-e, the volume and the fact that we haven’t seen a recovery suggests that the short-term outlook has just deteriorated. Consequently, we don’t support any short-term speculative positions in the Bitcoin market at this time.

The appreciation we’ve seen today doesn’t alleviate the concerns that Bitcoin might be on the brink of a new decline following a move below $500 which is now more than visible. The volume levels we’ve seen today don’t inspire too much confidence and project and image of a weak corrective move to the upside.

At present, it seems that unless we see a move above $500, the expected direction for the next move is down. We’re close to a situation when opening short speculative positions might be the way to go but we’d prefer to see where Bitcoin closes today before opting for that.

Summing up, in our opinion no speculative positions should be kept in the Bitcoin market now.

Trading position (short-term, our opinion): no positions.

[RoadStress]

while I agree that number of users is increasing exponentially (so far), the number of transactions is NOT, which was posted earlier.
Possible explanation-the absence of accurate statistics, but then the whole argument is mute.

What would you prefer on a personal level as a miner in the short and long term? An increase in transactions or an increase in exchange rate? What about for bitcoin overall?

[SlipperySlope]

There appears to be little correlation this year between bitcoin price and the mining difficulty.

Mathematically there shouldn't be unless at least Moore's Law died. You want to correlate mHash/$ and mHash/Watt, i.e. the those are both increasing faster than Moore's Law I think?

Agreed. I keep a column in my data series for daily mining reward and currently it is $1.7 million at Bitstamp prices. Miners are likely selling most of this to pay for the upgrades Moore's Law provides.

I will be speaking on a mining algorithm panel at the Hasher's United Conference in Las Vegas in October, where I will briefly explain to an audience of proof-of-work miners my alternative in development. Vitalik Buterin and Charles Lee will also be there to explain Ethereum and Litecoin algorithms respectively.

[AnonyMint]

Conceivably Armstrong might.  Although, I am very skeptical of his technical claims, having seen powerpoints attributed to 1997(?) but authored with Office 2003.

Please provide me evidence and I will email him. He reads my email. He has 80,000+ readers. Someone would be making such proofs available on the internet if it has any credibility.

So what he opened an old office file in a newer version to create a blog post? Does that refute all the attendees none who have refuted his claim as far as I know? Shoud I advise him to use instead open orfice in the future.

http://armstrongeconomics.com/2014/09/03/the-september-start-of-war/

QUESTION: Mr. Armstrong; The fact that you gave us all warning of the turn in the Cycle of War coming in 2014 back at the 1998 conference and at 2011 conference, was truly amazing to watch how this has all worked out. But the stunning realization that you pinpoint Ukraine as the flash point and then said it will turn up in September which has arrived with the Russians openly invading Ukraine, I just do not know how anyone from the goldbugs to government cannot appreciate what you have done with analyzing time. They expose their own stupidity and bias by trying to criticism you on a personal level when you have shown us this is not opinion or just outright ignore you and hope you go away. People should be supporting your work and the sooner you go public the better.

Do you have any idea what makes this all work so precisely? The ECM picks events to the day, but you said this war in Ukraine would turn up starting in September. Your timing is truly amazing.

Thank you so much for showing us a new way to look at everything. I use to attend various lectures on cyclical analysis. But they seemed primitive at best and only concerned about trading some single market thinking they have conquered the world. Nobody has ever comprehended the global connectivity that you show us all the time. I really do hope the government would just for once care about society and listen to what your models really project. So many lives would be saved not to mention money from chasing absurd ideas like global warming that is obviously a natural cycle. It is a shame the analytical community does not stand up and support you rather than trying to compete with you when they cannot even understand the depth to which you have taken cyclical analysis.

Thank you once again for showing us the light.

HRK

ANSWER: Thank you. I am very glad that some people are starting to see that this is something very important that could really change the future is implemented. We are trying to take this public to protect everything and then hopefully the technology can be applied to many fields. Ego keeps the industry fragmented. Only when we crash and burn does it appear that there would be any type of unity in the analysis community. They argued Keynes was nuts until the Great Depression hit. This is just how it functions. As for a market forecasters, generally all they are doing is trying to make a buck. They are not normally inspired by lofty goals. They are content trying to develop just trading systems to sell.

[findftp]

http://armstrongeconomics.com/2014/09/03/the-september-start-of-war/

Wow, thanks for the link.
Finally someone who writes what I already been thinking all along.
Bookmarked it.

[hyperdimension]

Question for rpietila and friends, i just got my first BTC. Once you reach 1 BTC, what would be the best strategy to start working towards doubling it and so on? I've thought starting by investing 0.2 and try to double that,  then invest these 0.4 and so on and so on.
I got like 50 XMR as well.

[aminorex]

Question for rpietila and friends, i just got my first BTC. Once you reach 1 BTC, what would be the best strategy to start working towards doubling it and so on? I've thought starting by investing 0.2 and try to double that,  then invest these 0.4 and so on and so on.
I got like 50 XMR as well.

The likelihood of daytrading your way to a large stash is very low.   Selling tops and buying bottoms is very slow work, if you want to keep risk managable.  Doing it fast with managed risk requires talent training and experience.  Almost certainly you are better off just buying it gradually, in constant fiat amounts, small and frequent.  Various savings plans involving sales to reduce exposure at highs, have been published.  Risto has one which is well considered.  Look for that thread.  David Latapie has one which he tuned for Monero.  Check his posts, probably during June.

[Capt Drake]

Question for rpietila and friends, i just got my first BTC. Once you reach 1 BTC, what would be the best strategy to start working towards doubling it and so on? I've thought starting by investing 0.2 and try to double that,  then invest these 0.4 and so on and so on.
I got like 50 XMR as well.

The likelihood of daytrading your way to a large stash is very low.   Selling tops and buying bottoms is very slow work, if you want to keep risk managable.  Doing it fast with managed risk requires talent training and experience.  Almost certainly you are better off just buying it gradually, in constant fiat amounts, small and frequent.  Various savings plans involving sales to reduce exposure at highs, have been published.  Risto has one which is well considered.  Look for that thread.  David Latapie has one which he tuned for Monero.  Check his posts, probably during June.

1 - Buy some coin that's very cheap that offers anonymity
2 - Wait for some weaks....
3 -
4 - PROFIT!

[BitCoinNutJob]

Im looking for the pic of rpietila's castle, i think its in this thread, looks a bit gothic, which page is it?

[Capt Drake]

Im looking for the pic of rpietila's castle, i think its in this thread, looks a bit gothic, which page is it?

http://www.bizforum.org/Journal/www_journalJVP016.htm

[Biodom]

I agree re waiting, but by December it should be clear (at least short term), don't you agree?

Not at all.  There is a great deal of psychological research that shows how our human minds try to find patterns that aren't really there.  A fantastic book on this topic is "Thinking Fast and Slow" by Daniel Kahneman.

IMO, the growth rate for bitcoin would need to decline strongly (or retreat over a long period) to truly invalidate the exponential growth model.  Here's a model that fixes bitcoin's market cap at inception at $500,000.  The "rationale" is that Satoshi spent approximately 2 years building it, and the market-value for Satoshi-level talent is $250,000 / year.  

https://i.imgur.com/OiKgYfu.png

I'm not arguing for this model, just pointing out that if growth slows to a more modest (but still exponential level), arguments could still be made that we are on trend.  IMO it would take a failure to reach a new ATH by 2017, or a sustained (1 year+) fall below $250, for me to say that "bitcoin growth has halted."  

...

It's so easy to be fooled by randomness.  Here's several simulations of the same underlying exponential growth model.  But instead of solving a regular differential equation to get a smooth exponential curve, I'm solving a stochastic differential equation that adds process noise.  The people in Alternate Universe #1 who get to ride the upper purple curve will think bitcoin is the most fantastic thing!  The people riding the bottom blue curve are going to make up story after story about how it's failing.  But in all cases, the only difference was randomness.  

Interesting..
However, if bitcoin will go through it's "plaque" years and then will be glorious by 2070, it wouldn't matter much to me. I am of the opinion that each generation has to build their own wealth, not ride on savings of their folks, so it does not look really appealing if some of my descendents will be very wealthy because of my activities in the first half of this decade.

[rpietila]

I agree re waiting, but by December it should be clear (at least short term), don't you agree?

Not at all.  There is a great deal of psychological research that shows how our human minds try to find patterns that aren't really there.  A fantastic book on this topic is "Thinking Fast and Slow" by Daniel Kahneman.

IMO, the growth rate for bitcoin would need to decline strongly (or retreat over a long period) to truly invalidate the exponential growth model.  Here's a model that fixes bitcoin's market cap at inception at $500,000.  The "rationale" is that Satoshi spent approximately 2 years building it, and the market-value for Satoshi-level talent is $250,000 / year.  

https://i.imgur.com/OiKgYfu.png

I'm not arguing for this model, just pointing out that if growth slows to a more modest (but still exponential level), arguments could still be made that we are on trend.  IMO it would take a failure to reach a new ATH by 2017, or a sustained (1 year+) fall below $250, for me to say that "bitcoin growth has halted."  

...

It's so easy to be fooled by randomness.  Here's several simulations of the same underlying exponential growth model.  But instead of solving a regular differential equation to get a smooth exponential curve, I'm solving a stochastic differential equation that adds process noise.  The people in Alternate Universe #1 who get to ride the upper purple curve will think bitcoin is the most fantastic thing!  The people riding the bottom blue curve are going to make up story after story about how it's failing.  But in all cases, the only difference was randomness.  

Interesting..
However, if bitcoin will go though it's "plaque" years and then will be glorious by 2070, it wouldn't matter much to me. I am of the opinion that each generation has to build their own wealth, not ride on savings of their folks, so it does not look really appealing if some of my descendents will be very wealthy because of my activities in the first half of this decade.

Thank You Peter R! I will add this model to the toolkit when I try to explain to people why sometimes (2013) it rises 100x and sometimes (2012) hardly at all. Randomness.

[AnonyMint]

Thank You Peter R! I will add this model to the toolkit when I try to explain to people why sometimes (2013) it rises 100x and sometimes (2012) hardly at all. Randomness.

Randomness is not the only possible explanation. Log-logistic adoption is another possibility for a declining rate of price appreciation given  Peter R's model correlating adoption to market cap and given the market cap is growing.

So it is not sufficient to say, "the model must be my fit, and randomness explains why the fit is off this year".