# Analysis of cryptocurrency market trends (on the example of Bitcoin)

Currently, there are an infinite number of publications analyzing the cryptocurrency market from experts of all stripes and ranks. The topic is so popular that only lazy and not sophisticated did not make its review. Analysts' forecasts are sometimes so different that you can’t decide what to do next: look for a bank that you can lay an apartment in and invest in another “-coin”, or train an aggressive “I said!” In front of the mirror and portray it with infernal laughter the ghost of Leni Golubkov.

This article does not set itself the goal of persuading the reader in favor of investing in the cryptocurrency market, or, on the contrary, disowning the cryptocurrency with the Orthodox dollar. This essay proposes to evaluate the prevailing patterns in terms of numbers and formulas, omitting the permanent fluctuations caused by fundamental factors or the strength of habit (yes, the psychological factor and market sentiment can sometimes significantly shift the current state of things; at the end of the article we find out what caused collapse in currency prices in January 2018).

Well, let's do an analysis of the main indicators of the cryptocurrency market together with the example of the mastodon industry - Bitcoin. However, I dare to assure that the analysis of the behavior of most top currencies gives similar results (Litecoin, Dash, Ethereum, Monero, Z-cash, etc.). Undoubtedly, the most important and at the same time interdependent indicators of the cryptocurrency market are the price of the currency itself (we will consider the dollar expression of value), as well as the degree of involvement of computing power in the process of “mining” coins - the network hashrate.

What is the relationship between cryptocurrency price and hash? Firstly, it’s absolutely obvious that by involving your own equipment in the process of confirming transactions in the cryptocurrency network (“mining”), a person or company does not only expect to return the cost of the invested equipment (and its replacement is necessary every 2-3 years due to physical wear and tear and obsolescence ), but also to cover the infrastructure costs of its maintenance, as well as to obtain economic benefits in the form of profit no less than alternative sources of profitability; secondly, the issue fee for participating in cryptocurrency mining falls on the secondary market only indirectly at the time of sale of the currency by the miner (the owner of the mining equipment). Thus, an artificial asset shortage is created in the market,

On the other hand, with the increase in the price of cryptocurrency, mining profitability also grows, which encourages to involve more and more capacities in the cryptocurrency “mining” process, thereby increasing the network’s total hashrate.

Let's go directly to the analysis of indicators. All historical parameter data was downloaded from bitinfocharts.com.

Consider the graph of Bitcoin price dynamics for the period from July 17, 2010 to January 24, 2018 (the moment of writing the article)

From the graph, it can be seen that the dynamics of the price change is clearly non-linear in nature, and a power-law time-dependent. We resort to the logarithm of the data in order to move from the obviously power-law dependence of the function on the argument (price versus time) to the expected linear one and see how this will affect our graph.

Now the graph represents a curve linearly dependent on time with a confidence level of 84%.

We will figure out where this degree of reliability came from. The curve linearly dependent on time means that the graph of our function should tend to a straight line (shown by a dotted line on the graph). This line can be obtained by resorting to a regression analysis apparatus for finding linear regression coefficients. Further, it is important to evaluate how much our initial function is approximable to the obtained straight line (that is, how much the behavior of the initial curve can be predicted by the obtained straight line). The determination coefficient R2 can help in this - a statistical characteristic that describes how much the variance (spread) of the source data is explained by the variance of the model. Simply put, the closer the original curve is to the simulated straight line, the higher the value of R2 (varies from 0 to 1), and the higher the probability that the future behavior of the indicator will correspond to the trend of the resulting model. In our case, the determination coefficient was 0.8432.

It is clearly seen that the curve of the dependence of the logarithm of price on time has local extremes and trends, in other words, at different points in time, the price is either ahead of the global trend, it is somewhat behind it, but with 84% confidence it followed. Looking at the graph, it would be logical to assume that by changing the analysis horizon, it is possible to achieve a more “accurate” superposition of the curve on the linear trend. Below are charts with statistics of R2 for the last 2 years, 1 year and 6 months, respectively.

As you can see from the graphs, the value of statistics R2 is from 0.88 to 0.95. Obviously, if the calculated linear trend (dashed line on the chart) is extended into the future, then it is possible to obtain estimated levels of indicator values for a given date (data extrapolation).

Let's mentally go back half a year ago and test the hypothesis about the possibility of predicting cryptocurrency price trends. So if we were to evaluate the current value of the price ($ 10,000) of Bitcoin six months ago (the price as of July 27, 2017 was $ 2,500), we would get the following price levels depending on the analyzed horizon of historical data:

2-year trend (07.28.2015 - 07.27.2017) - the price is $ 4,400;

one-year trend (07/27/2016 - 07/27/2017) - the price is $ 5,800;

9-month trend (10.26.2016 - 07.27.2017) - price $ 6,900;

6-month trend (January 25, 2017 - July 27, 2017) - the price is $ 8,900.

It can be seen that the price behavior changes depending on the analysis horizon and tends to accelerate growth, and the real current price range is even slightly higher than the most optimistic forecast. Thus, it is advisable to evaluate potential future price levels relative to the current date.

So, using the obtained linear regression price trends, we have that the average forecast of the price level after 3 months is $ 13,600, and after 6 months - $ 16,700.

An analysis of the involvement of computing power in the process of “mining” cryptocurrencies gives similar results with even greater significance of statistical reliability (real data are located closer to the approximating line). Below are graphs of the dependence of the logarithm of the total hash of the network on time for different time ranges.

If we talk about quantitative estimates of extrapolated data, then the average forecast for trends after 3 months is 26.3 EH / s (26.3 * 1018), and after 6 months 34.6 EH / s at the current level of 20.5 EH / s.

Thus, it is clear that the predicted relative increase in the main indicators (price and hashrate) is almost identical, and, therefore, tells us about the expected conditionally constant income from mining. And if we take into account the fact that with the growth of the cryptocurrency exchange rate against the dollar, the infrastructure costs in terms of cryptocurrency decrease, then investing in its “production” is very attractive despite the steady increase in the aggregate computing power.

And finally, the promised story about the power of habit and the cryptocurrency market crash. I, as many readers believe, am used to looking at regular statistical and analytical data on the same information resources, be it weather forecasts, sports results or the dynamics of the cryptocurrency market. So, there is some very popular resource that analyzes the prices of various cryptocurrencies and the capitalization of their markets (coinmarketcap.com), while it is a fairly convenient source of operational market statistics. As you know, the prices of cryptocurrencies on different exchanges vary, and sometimes significantly, and such analytical resources average these prices and use aggregated data in their calculations. In the crypto community, the fact that that South Korean traders artificially inflate the value of cryptocurrency and the quotes of these exchanges distort the picture. The owner and manager of CoinMarketCap, a certain Brandon Chaz, decided to correct this market injustice and excluded the quotes of the Korean market from the calculation of average cryptocurrency prices, but forgot to inform customers about it, as a result of which the market capitalization fell sharply. Few outside of CoinMarketCap understood the true reason for this “fall”, which was essentially just a numbers game. However, many traders trying to take profits began to actively sell cryptocurrencies, which significantly affected their rates like a snowball. As a result, the capitalization of the cryptocurrency market as a whole lost about $ 100 billion. The owner and manager of CoinMarketCap, a certain Brandon Chaz, decided to correct this market injustice and excluded the quotes of the Korean market from the calculation of average cryptocurrency prices, but forgot to inform customers about it, as a result of which the market capitalization fell sharply. Few outside of CoinMarketCap understood the true reason for this “fall”, which was essentially just a numbers game. However, many traders trying to take profits began to actively sell cryptocurrencies, which significantly affected their rates like a snowball. As a result, the capitalization of the cryptocurrency market as a whole lost about $ 100 billion. The owner and manager of CoinMarketCap, a certain Brandon Chaz, decided to correct this market injustice and excluded the quotes of the Korean market from the calculation of average cryptocurrency prices, but forgot to inform customers about it, as a result of which the market capitalization fell sharply. Few outside of CoinMarketCap understood the true reason for this “fall”, which was essentially just a numbers game. However, many traders trying to take profits began to actively sell cryptocurrencies, which significantly affected their rates like a snowball. As a result, the capitalization of the cryptocurrency market as a whole lost about $ 100 billion. Few outside of CoinMarketCap understood the true reason for this “fall”, which was essentially just a numbers game. However, many traders trying to take profits began to actively sell cryptocurrencies, which significantly affected their rates like a snowball. As a result, the capitalization of the cryptocurrency market as a whole lost about $ 100 billion. Few outside of CoinMarketCap understood the true reason for this “fall”, which was essentially just a numbers game. However, many traders trying to take profits began to actively sell cryptocurrencies, which significantly affected their rates like a snowball. As a result, the capitalization of the cryptocurrency market as a whole lost about $ 100 billion.

With this example, I would like to emphasize once again the fundamental impossibility of accurate forecasting of the cryptocurrency price at any given time, however, you can evaluate the internal trends of the main market characteristics and understand price levels-benchmarks that you can rely on in the decision-making process.

This article does not set itself the goal of persuading the reader in favor of investing in the cryptocurrency market, or, on the contrary, disowning the cryptocurrency with the Orthodox dollar. This essay proposes to evaluate the prevailing patterns in terms of numbers and formulas, omitting the permanent fluctuations caused by fundamental factors or the strength of habit (yes, the psychological factor and market sentiment can sometimes significantly shift the current state of things; at the end of the article we find out what caused collapse in currency prices in January 2018).

Well, let's do an analysis of the main indicators of the cryptocurrency market together with the example of the mastodon industry - Bitcoin. However, I dare to assure that the analysis of the behavior of most top currencies gives similar results (Litecoin, Dash, Ethereum, Monero, Z-cash, etc.). Undoubtedly, the most important and at the same time interdependent indicators of the cryptocurrency market are the price of the currency itself (we will consider the dollar expression of value), as well as the degree of involvement of computing power in the process of “mining” coins - the network hashrate.

What is the relationship between cryptocurrency price and hash? Firstly, it’s absolutely obvious that by involving your own equipment in the process of confirming transactions in the cryptocurrency network (“mining”), a person or company does not only expect to return the cost of the invested equipment (and its replacement is necessary every 2-3 years due to physical wear and tear and obsolescence ), but also to cover the infrastructure costs of its maintenance, as well as to obtain economic benefits in the form of profit no less than alternative sources of profitability; secondly, the issue fee for participating in cryptocurrency mining falls on the secondary market only indirectly at the time of sale of the currency by the miner (the owner of the mining equipment). Thus, an artificial asset shortage is created in the market,

On the other hand, with the increase in the price of cryptocurrency, mining profitability also grows, which encourages to involve more and more capacities in the cryptocurrency “mining” process, thereby increasing the network’s total hashrate.

Let's go directly to the analysis of indicators. All historical parameter data was downloaded from bitinfocharts.com.

### Bitcoin Price Trend Analysis

Consider the graph of Bitcoin price dynamics for the period from July 17, 2010 to January 24, 2018 (the moment of writing the article)

From the graph, it can be seen that the dynamics of the price change is clearly non-linear in nature, and a power-law time-dependent. We resort to the logarithm of the data in order to move from the obviously power-law dependence of the function on the argument (price versus time) to the expected linear one and see how this will affect our graph.

Now the graph represents a curve linearly dependent on time with a confidence level of 84%.

We will figure out where this degree of reliability came from. The curve linearly dependent on time means that the graph of our function should tend to a straight line (shown by a dotted line on the graph). This line can be obtained by resorting to a regression analysis apparatus for finding linear regression coefficients. Further, it is important to evaluate how much our initial function is approximable to the obtained straight line (that is, how much the behavior of the initial curve can be predicted by the obtained straight line). The determination coefficient R2 can help in this - a statistical characteristic that describes how much the variance (spread) of the source data is explained by the variance of the model. Simply put, the closer the original curve is to the simulated straight line, the higher the value of R2 (varies from 0 to 1), and the higher the probability that the future behavior of the indicator will correspond to the trend of the resulting model. In our case, the determination coefficient was 0.8432.

It is clearly seen that the curve of the dependence of the logarithm of price on time has local extremes and trends, in other words, at different points in time, the price is either ahead of the global trend, it is somewhat behind it, but with 84% confidence it followed. Looking at the graph, it would be logical to assume that by changing the analysis horizon, it is possible to achieve a more “accurate” superposition of the curve on the linear trend. Below are charts with statistics of R2 for the last 2 years, 1 year and 6 months, respectively.

As you can see from the graphs, the value of statistics R2 is from 0.88 to 0.95. Obviously, if the calculated linear trend (dashed line on the chart) is extended into the future, then it is possible to obtain estimated levels of indicator values for a given date (data extrapolation).

Let's mentally go back half a year ago and test the hypothesis about the possibility of predicting cryptocurrency price trends. So if we were to evaluate the current value of the price ($ 10,000) of Bitcoin six months ago (the price as of July 27, 2017 was $ 2,500), we would get the following price levels depending on the analyzed horizon of historical data:

2-year trend (07.28.2015 - 07.27.2017) - the price is $ 4,400;

one-year trend (07/27/2016 - 07/27/2017) - the price is $ 5,800;

9-month trend (10.26.2016 - 07.27.2017) - price $ 6,900;

6-month trend (January 25, 2017 - July 27, 2017) - the price is $ 8,900.

It can be seen that the price behavior changes depending on the analysis horizon and tends to accelerate growth, and the real current price range is even slightly higher than the most optimistic forecast. Thus, it is advisable to evaluate potential future price levels relative to the current date.

So, using the obtained linear regression price trends, we have that the average forecast of the price level after 3 months is $ 13,600, and after 6 months - $ 16,700.

### Bitcoin Hash Trend Analysis

An analysis of the involvement of computing power in the process of “mining” cryptocurrencies gives similar results with even greater significance of statistical reliability (real data are located closer to the approximating line). Below are graphs of the dependence of the logarithm of the total hash of the network on time for different time ranges.

If we talk about quantitative estimates of extrapolated data, then the average forecast for trends after 3 months is 26.3 EH / s (26.3 * 1018), and after 6 months 34.6 EH / s at the current level of 20.5 EH / s.

Thus, it is clear that the predicted relative increase in the main indicators (price and hashrate) is almost identical, and, therefore, tells us about the expected conditionally constant income from mining. And if we take into account the fact that with the growth of the cryptocurrency exchange rate against the dollar, the infrastructure costs in terms of cryptocurrency decrease, then investing in its “production” is very attractive despite the steady increase in the aggregate computing power.

And finally, the promised story about the power of habit and the cryptocurrency market crash. I, as many readers believe, am used to looking at regular statistical and analytical data on the same information resources, be it weather forecasts, sports results or the dynamics of the cryptocurrency market. So, there is some very popular resource that analyzes the prices of various cryptocurrencies and the capitalization of their markets (coinmarketcap.com), while it is a fairly convenient source of operational market statistics. As you know, the prices of cryptocurrencies on different exchanges vary, and sometimes significantly, and such analytical resources average these prices and use aggregated data in their calculations. In the crypto community, the fact that that South Korean traders artificially inflate the value of cryptocurrency and the quotes of these exchanges distort the picture. The owner and manager of CoinMarketCap, a certain Brandon Chaz, decided to correct this market injustice and excluded the quotes of the Korean market from the calculation of average cryptocurrency prices, but forgot to inform customers about it, as a result of which the market capitalization fell sharply. Few outside of CoinMarketCap understood the true reason for this “fall”, which was essentially just a numbers game. However, many traders trying to take profits began to actively sell cryptocurrencies, which significantly affected their rates like a snowball. As a result, the capitalization of the cryptocurrency market as a whole lost about $ 100 billion. The owner and manager of CoinMarketCap, a certain Brandon Chaz, decided to correct this market injustice and excluded the quotes of the Korean market from the calculation of average cryptocurrency prices, but forgot to inform customers about it, as a result of which the market capitalization fell sharply. Few outside of CoinMarketCap understood the true reason for this “fall”, which was essentially just a numbers game. However, many traders trying to take profits began to actively sell cryptocurrencies, which significantly affected their rates like a snowball. As a result, the capitalization of the cryptocurrency market as a whole lost about $ 100 billion. The owner and manager of CoinMarketCap, a certain Brandon Chaz, decided to correct this market injustice and excluded the quotes of the Korean market from the calculation of average cryptocurrency prices, but forgot to inform customers about it, as a result of which the market capitalization fell sharply. Few outside of CoinMarketCap understood the true reason for this “fall”, which was essentially just a numbers game. However, many traders trying to take profits began to actively sell cryptocurrencies, which significantly affected their rates like a snowball. As a result, the capitalization of the cryptocurrency market as a whole lost about $ 100 billion. Few outside of CoinMarketCap understood the true reason for this “fall”, which was essentially just a numbers game. However, many traders trying to take profits began to actively sell cryptocurrencies, which significantly affected their rates like a snowball. As a result, the capitalization of the cryptocurrency market as a whole lost about $ 100 billion. Few outside of CoinMarketCap understood the true reason for this “fall”, which was essentially just a numbers game. However, many traders trying to take profits began to actively sell cryptocurrencies, which significantly affected their rates like a snowball. As a result, the capitalization of the cryptocurrency market as a whole lost about $ 100 billion.

With this example, I would like to emphasize once again the fundamental impossibility of accurate forecasting of the cryptocurrency price at any given time, however, you can evaluate the internal trends of the main market characteristics and understand price levels-benchmarks that you can rely on in the decision-making process.