Quote from: AnonyMint on July 19, 2014, 02:05:31 AM

Quote from: rpietila on July 18, 2014, 07:06:00 AM

Quote from: AnonyMint on July 18, 2014, 01:46:07 AM

As Bitcoin has become more centralized, they will strangle it, which we can see happening if we look at any metrics that measure the rate of growth, e.g. the market cap rate of growth has slowed which is what I showed that it is log-logistic not logistic as should be normal for technology adoption. (note Risto doesn't accept or agree with my curve fit, so we will have to wait to see if it was correct or not)[/b]

It is so tiring (and wasting my time which takes away from doing real work that could help us) to respond to slander by ignorant posters due to their ignorance.

Your log-logistic fit would have been perfect in late 2012, showing the slowing of growth and maturation of the technology. So having this analysis, you would have sold all your bitcoins at $13 or at least never bought any. Right?

How do you calculate that my fit predicted that in 2012? Here is the crude fit graph I did (notably after 2012). It seems that even if I did that fit before the run-up in 2013, it would be impossible to fit it such that it wouldn't have shown the price was far too low before the run-up in price in 2013. This is because of that spike up in the price in the early stage (which is characteristic of log-logistic).

I didn't try to make the data to fit my model. I was trying to find a math model that fit that data. Logistic doesn't fit, go compare to SlipperySlope's graphs.

Quote from: AnonyMint on April 02, 2014, 09:52:15 AM

Quote from: AnonyMint on April 02, 2014, 04:59:30 AM

Quote from: rpietila on July 18, 2014, 07:06:00 AM

So one of us now holds bitcoins at $623, because his model was correct and already has an astounding 94.5% Rsq-coefficient (far higher than any log-logistic fit to the bitcoin 5-year price data).

Please show me the math you claim, because I am lazy (lacking time) to go figure it out myself?

Quote from: rpietila on July 18, 2014, 07:06:00 AM

Bad models have real world consequences to those who choose to follow them. The weakness of yours is that it is obvious that the saturation point of bitcoin is in 100s of millions or billions of users, and you are interpreting random noise to mean a secular slowing of trend at the point when the actual slowing of the adoption rate is still 3 orders of magnitude away!

Bitcoin might indeed reach a few 100 million (but only via some Thiel service such as Bitpay, Coinbase, Facebook WhatsApp, etc), but that doesn't mean its adoption rate wouldn't slow on the way there. You seem to be assuming the adoption rate wouldn't slow until that point, and then suddenly so. Continuous curves can't do that. Either you have S logistic curve or you don't.

The global population is 6+ billion and will be 7 - 8 billion by 2020. So Bitcoin scaling to 1/10 or 1/50 of the global population is not the same penetration as cell phones, computers, and internet which will likely reach 80% (8/10) of the population.

Please realize that WhatsApp is more important to most users than Bitcoin. Thiel can do a money transfer service in WhatsApp to address most of Bitcoin's thrust for most users (who don't need some crazy volatile technological thing). He is merely pulling Bitcoin into the fold of his overall control over money transfer globally with his also Paypal in his stable.

Trying to curve-fit "adoption" is really silly at this point. It's probably silly for most tech phenomena as they all have very different, often unpredictable, dynamics that are never going to follow any nice curve, so the fit is really just an exercise in self-indulgence. Furthermore, the playing field is always changing, metrics are very fuzzy, and statistical significance likely goes out the window.

Essentially, if you could run this earth sim a thousand times and drop bitcoin into the world and see what happens, you'd probably get pretty different graphs. Likely the similar t=large end-state most of the time, but the path is tumultuous, *especially* in the beginning (which this likely still is (either that or the end, of course)).