Article contributed by Avran Wong
Follow him on Medium
“Is the market cycle over?” Amid emotion-inducing hype content promising quick riches in the world of cryptocurrencies, how can we tell what kinds of price movements are realistic for the asset class? What are some ways we can value a cryptocurrency based on its previous price action?
To me, crypto doesn’t always have to be a “buy and hope it goes up” kind of investment. I believe that historical movements in price can indicate what could theoretically be possible. Using basic data science techniques, we can look through the noise to get a sense of where we are in the market cycle.
Since the inception of Bitcoin in 2009, the cryptocurrency asset class is known to go through various boom and bust cycles, with price declines as much as 80–90% and gains in the thousands as well. The models we are going to discuss are referenced from Benjamin Cowen’s content, though the words and code are originally written by myself.
We will be going through three models
- Deviation from Bitcoin’s “fair value”
- Market cycle Returns
- Theoretical logarithmic upper and lower bounds of price
That said, while Bitcoin has been around for more than a decade, there is still relatively little data in terms of market cycles. All models are wrong to an extent and they should be used cautiously along with other indicators.
The data shown in the charts are as of 28th December 2021.
Thoughts on diminishing returns and lengthening cycles
This is the chart of bitcoin’s price against time on a logarithmic scale. As we can see, though the price of bitcoin tends to trend up with time, the rate of price appreciation seems to be slowly decreasing.
The models here are rooted in theories of diminishing returns and lengthening cycles. This means that as an asset matures and grows in market capitalisation (market cap), it will be more and more improbable that the price will appreciate as much as before, simply because it takes way more money to move the price of bitcoin by 10% now, as compared to 10% in 2012. It also stands to reason that market cycles will lengthen as the market cap increases as well. Even though Bitcoin is still volatile, we don’t see as rapid and as dramatic (relatively) price movements anymore, we simply need more time than before to change the overall momentum of Bitcoin given its larger market cap (~900 billion as of Dec 2021)
Deviation from Bitcoin’s “fair value”
In this model, we will be charting Bitcoin’s percentage deviation from its estimated fair value. The fair value of bitcoin is estimated using the 20-week simple moving average (20W SMA) and the 21-week exponential moving average (21W EMA). These moving averages are historically shown to be a decent indicator of bullish/bearish sentiments.
By charting the difference between Bitcoin’s value and the estimated fair value, we can visualise previous peaks and bottoms against its fair value.
*Deviation = (Price/ Fair Value) x 100%. The graph is using a logarithmic scale
This graph is the percentage deviation of Bitcoin’s price relative to its estimated fair value. We can see that the deviation decreased both on the upside and downside over the years. We can extrapolate the previous peaks to find a linear trend line. While it has only been fitted to the previous cycle peaks, it has been successful in calling the pullback in May 2021. However, I do expect this model to be wrong at some point simply because the trend line will eventually cross the fair value. The point of this model is to give us an idea of what the next peak can reasonably be, allowing us to manage our emotions during euphoric phases.
Market Cycle ROI
The next model we will be looking at is to compare the market cycle returns of investment (ROI) across the history of Bitcoin. This can give us a sense of where we are in the current cycle (assuming lengthening cycles and diminishing returns)
*The market cycle is defined from the bottom of the previous peak to the next peak. Take note that this is on a log scale as well.
Cycle 1: Length: 245 days, ROI: 58,333%
Cycle 2: Length: 744 days, ROI: 50,262%
Cycle 3: Length: 1068 days, ROI: 11,047%
Cycle 4 (Current cycle): Length: 1109 days, ROI: 2,083%
The data shows that not only does each cycle’s ROI get lower and lower, but the length of each cycle also increases. While we are technically already experiencing a lengthening cycle, the data suggest that not only does the cycle usually lengthen, it extends for hundreds of days more at a time. If lengthening cycles were to play out, this model suggests that we are still a ways to go from the peak, but we definitely should not be expecting another 50,000% ROI anytime soon.
Another popular way to compare market cycles is to measure from Bitcoin’s halving, though this is not covered in this article, it does suggest similar trends as well.
Theoretical logarithmic upper and lower bounds of price
The last model I will be presenting is the theoretical log upper and lower bounds.
*The peak data consist of the peak price in 2011, 2013 and 2018. While the “bear market data” is defined as the bottom of the previous peak to the last time the price reached its fair value.
In this model, the data set is split into “peak data” and “bear market data”. By splitting the data set based on peak market fear and greed, we can study theoretical tops and bottoms. Using a logarithmic regression function, we can predict the theoretical next market cycle peak and next bear market.
While we still have some room before we reach the upper bounds, we are closer to the peak than the bottom. The good thing about Bitcoin is that you don’t need to time the perfect bottom or top to get a good return, by understanding the possible upper and lower bounds of Bitcoin, we can better measure our possible risk and reward.
The models presented here are solely based on historical data, at the end of the day we are just dealing with probabilities and nobody knows for sure what is going to happen. All models are wrong to a certain extent, but to say that history has no impact on what is possible is a bit too far-fetched to me.
I find that it’s useful to have an open mind, while there are similarities between cycles, no two cycles will be the same, thinking this cycle will have to play out the exact way it did previously is a more or less guaranteed way to be wrong. By understanding what kinds of returns are possible within what kind of time frames, we can manage our expectations and be mentally prepared for the volatility. After all, Bitcoin is one of the best-performing assets in the decade, I argue that as long we stay the course and manage our emotions, we are all going to make it.
Bitcoin market price data from quandl: https://data.nasdaq.com/data/BCHAIN/MKPRU-bitcoin-market-price-usd
Benjamin Cowen’s youtube channel: https://www.youtube.com/channel/UCRvqjQPSeaWn-uEx-w0XOIgThe original notebook can be found on github here: https://github.com/AvranWong/bitcoin_historical_movements