Our study will focus on the role of rare variations in the performance of bitcoin (BTC). This exclusive study shows amazing findings. Indeed, it appears in particular that 7% of the time explains up to 90% of the performance of bitcoin (BTC). Therefore, this is obviously out of proportion with what one would expect. This is also in line with the conclusions of our previous article. Our study will aim here to detect the variations “rare” that explain the performance of bitcoin (BTC).
Distribution of daily changes in bitcoin (BTC)
We first selected the daily data set for bitcoin between May 2018 and May 2023. That is to say over a period of 5 years that we consider sufficiently representative, with more than 1800 days. We then compiled all these variations according to their range to obtain the following distribution.
Thus, we can specify that the average daily change of bitcoin (BTC) is exactly +0.128%. Moreover, one observes rather intuitively that the variations “extreme”those located at the extremities of the bell curve, are “abnormally” many. The two bounds in red represented on the graph illustrate the bounds that include a daily variation of +7% (right) or -7% (left). This will help us in our reasoning.
In addition, we will specify that the proportion of bearish days is greater than the proportion of bullish days. Similarly, the share of bearish days is greater than the share of bearish weeks, months, or years on the price of bitcoin (BTC). Indeed, according to another study we conducted in August 2021, “between 2016 and 2021, out of more than 2,000 daily variations, the risk of loss is 63.9% of the time”.
How to measure the importance of extreme variations?
Our study first requires a brief explanation of the method used. From the more than 1800 data we have, we establish the next steps.
- First of all we rewrite the data set by excluding the most extreme data (beyond the threshold we have set ourselves). For example, we will replace all variations above 8% and below 8% with zero performance. The number of days that are also beyond this threshold are counted in parallel.
- Then we averages all the daily variations that we have rewritten.
- Finally, we get a new average daily performance which can be compared with the global average daily performance stated above (+0.128% per day). We also compare in parallel the number of days “extreme” compared to the total number of days (1800).
In other words, we measure the daily performance deficit induced by the absence of the significant variations that we have eliminated. As a result, this allows us to deduce the performance induced (respectively the share of performance absent from the initial total performance) by the variations “extreme”.
7% of the time explains up to 90% of performance!
The heart of our work is summarized in the graphic below. We studied separately the average daily performance of bitcoin in the case where we crowd out variations beyond + / – 3%, then 4%, then 5%, and this, up to 11%… For example, if we exclude variations above 3% or below -3%, then we deduce that variations beyond this threshold represent 73% of all the average daily performance of bitcoin (BTC). Similarly, the number of days on which the daily variations exceeded + / – 3% was calculated. Consequently, it appears that 27.6% of the days have a daily variation greater than 3% or less than -3%.
The graph above shows in dotted lines the proportion of time that includes variations above (and below) the percentage determined on the abscissa. Thus, more than 27% of the time we notice variations of + / – 3% on bitcoin (BTC). In addition, the continuous curve in black represents the share of bitcoin’s performance induced by variations beyond the threshold on the abscissa. The greater the gap between the two curves, the more a small part of the time concentrates most of the performance.
Major observations
For example, variations above 10% or below -10% represent 2.3% of the time. Nevertheless, they explain more than 30% of the average daily performance of bitcoin (BTC). But the disproportion is most noticeable for extreme variations of + / – 7%. Variations below -7% or above +7% represent 6.5% of the time. However, they represent nearly 90% of the average daily performance of bitcoin (BTC).
This can be seen visually on the distribution graph shown above. Beyond the red marker on the right, positive “extreme” variations are over-represented between +7.5% and +8.5%. In other words a trader who misses most of these daily variations would see his average performance greatly reduced. Finally, we notice that extreme swings of +/- 11% have a smaller impact on explaining bitcoin performance.
Deduction on the behavior of bitcoin (BTC)
An antinomy seems to appear. Extreme variations at 7% are more influential on the performance of bitcoin than extreme variations at 10%. This means that extreme variations of +/- 10% are less likely to make the market perform better and are more often followed by at least partial recovery in the opposite direction.
In the case of extreme variations at 7%, this means that variations of this order are more rarely followed by a reverse catch-up movement. The market is more impulsive and less hesitant and realizes on these variations most of its opposites, and consequently, of its performances.
The importance of extreme variations
We have already discussed, in previous publications, the importance of extreme variations. Indeed, it appears that most trend reversals appear around these variations “extreme”. We said above that extreme variations of + / – 10% represent 2.3% of the time since 2018. That is to say that in a year of 365 days, approximately 8 to 9 days will concentrate these variations. Which means that such a variation ” extreme “ takes effect on average every 6 to 7 weeks. Finally, it is relevant to recall that Bitcoin is a volatile asset and its behavior strongly depends on the temporality of its variations. We have already written about this (fractals): Study of fractals on bitcoin (BTC) – Tremplin.io.
The graph above illustrates with vertical lines the times of appearance of these extreme variations. Bullish or bearish reversals are usually driven by such movements. The recurrence of extreme variations reflects a strong change in investor psychology and often provides information on the state of the trend. Finally, these variations are most often frequent around the peaks. Indeed, bitcoin volatility is usually greater in uptrends.
Bad news for short-term trading?
We have seen that a minority of the variations, namely the strongest variations, explain almost all of the performance of bitcoin (BTC). This is not without consequence on how to approach their strategy for investors and traders. Indeed, how to generate bullish performances if the major part of these performances are concentrated on a few days?
An investor who would be positioned throughout the period studied would benefit from this phenomenon. He wouldn’t miss any extreme market moves, so he wouldn’t miss most of the market performance. The risk for the trader who makes many back and forths is to stay away during these few days which make the performance of the market.
But the conclusions we have drawn can be interpreted in another way. A trader who would be able to position himself on these few exceptional days would be able, in theory at least, to outperform the market. However, this hypothesis remains difficult to support insofar as it would require great capacities for anticipation. Most traders react with volatility but not necessarily in a good way and not early enough. It’s no surprise that most cryptocurrency traders lose in the long run.
Conclusion
Our study looked at a large data set from 2018 to 2023. The sample studied clearly shows that a minority of the daily variations of bitcoin (BTC) have a major influence on the overall performance of the latter. The most eloquent figure comes with the extreme variations at 7%, that is to say the variations below -7% and above +7%. While these variations represent only 7% of the time, they explain up to 90% of the performance observed on the price of bitcoin (BTC). We have a lot to do, according to the statistics, at a “Pareto’s law” exacerbated!
There is a clear over-representation of exceptional variations in bitcoin’s performance. In addition, it is important to remember that bitcoin (BTC) is by nature a volatile asset. We therefore already knew that extreme variations played a central and abnormally frequent role. But this study confirms to us that these days “exceptional” clearly have an even greater influence than one could imagine. The influence of an extreme variation is all the greater in that it can drive clear trends, without turning back.
In this context, an investor (or a trader) who would miss most of these variations “extreme” would underperform the market. This observation is also common to most markets and reminds us once again of the inefficiency of certain traditional models, and of certain behaviors, be ” phase “ with the market.
Receive a digest of news in the world of cryptocurrencies by subscribing to our new service of daily and weekly so you don’t miss any of the essential Tremplin.io!
