# In a parable about a million US dollars

The first step for the modelling is to create a fit based on the time between 17 July and 29 November 2017. As a function we assume the following:

Course = K0*(1+gradient)^(t-t0)

The function that best fits the price trend would be one where the distance between this model and the actual price is smallest. One then obtains the following relationship, represented logarithmically:

The data of the fit model can now be used to determine a value for 31.12.2020. And indeed! The best fitting fit, extrapolated by the end of 2020, even leads to 23 million US dollars. So far one could say that John McAfee’s forecast is very conservative.

The problem with such a view is that we focused at the top on the development during a bull run. If the data are taken into account from July 2017 to the present day, “only” half a million people are left.

## This study could be carried out in more detail

How does the forecast for 31.12.2017 develop depending on the starting time? For the next analysis, the fit is extrapolated to 31.12.2017, whereby the fit is based on data between a variable starting point and today’s date.

Of course, the data can only be considered to a limited extent; from the beginning of 2018, the statistics are too small, which means that forecasts fluctuate considerably. But before that the graph looks like this:

As one can see, the forecast “one million US dollars per bit coin” is currently not tenable at least over a fit based on an exponential price development. If only data since the beginning of 2017 are taken into account, the forecast is tenable, but after that the forecast price drops dramatically. In this respect, it is to be feared that McAfee’s claim was primarily based on the bull market.

## Watch out, McAfee: Only a small probability for a million!

Another approach was introduced a few months ago: Using random walks and the Monte Carlo method, you can model estimates of the Bitcoin price. In this article, the aim was to obtain a price statement for only one year, but the method can also be used for larger time forecasts. Let us therefore consider the distribution of simulated price developments up to 31.12.2020. The histogram shows the probability of achieving certain prices over the same period:

If you take the last 2,000 prices into account, there is actually a certain probability that a price of one million euros will be reached at the end of 2020. How high is this? Unfortunately only at 3 percent.

Time will tell whether John McAfee should worry about his crown jewels. What is certain is that his bold forecast is not impossible, but quite unlikely.