This is not investment advice. Bitcoin is highly volatile. Past performance of back-tested models is no assurance of future performance. Only invest what you can afford to lose. You must decide how much of your investment capital you are willing to risk with Bitcoin. No warranties are expressed or implied.
Figure 1: Bitcoin’s death cross in mid-June 2021. Bloomberg chart.
What’s special about 200 days?
Recently there has been a lot of news around a “death cross” on the Bitcoin price chart. This is the crossover of the 50 day moving average below the 200 day average, that has occurred as Bitcoin moved down from a peak of over $60,000 in mid-April to the low $30,000s today.
Here’s one view of death crosses and moving averages in the stock market: https://www.forbes.com/sites/michaelcannivet/2019/06/06/if-you-want-to-time-the-market-ignore-moving-averages/?sh=3a793b527f48 A five-year backtest of long the S&P500 when above the 200 day and short otherwise from mid-2014 to mid-2019 would have lost you money. And the long side only position averaged just 1% per year.
What’s so special about 200 days? Well in the stock market it corresponds to a little over 40 weeks of trading once you account for holidays, about the human gestation period of nine months. Pardon me for that comparison, my wife and I just had a baby boy this month.
But consider Bitcoin and cryptocurrencies. They trade 7 days a weeks, so 200 days for Bitcoin is 28-1/2 weeks, just over half a year. There is nothing special about 50, 100, or 200 day intervals unless other traders believe in them, and then you might want to front run by using 40, 80, 160 days instead.
And this is all before noting that the US stock market trades 32.5 hours a week, just 19.3% of the 168 hours a week that Bitcoin can be bought and sold.
How 19th century is that? If we talked about 6.5 hour days, there would be a month of stock market equivalent trading days every week for Bitcoin.
In Bitcoin’s case it looks like you are better off accumulating when it falls below its 200-day moving average, not selling. Even more interesting in terms of long term support for Bitcoin’s price is the 200-week average. We’ll get to that, but first…
Bitcoin’s natural calendar
Bitcoin has its own natural rhythm. It is driven by three time scales defined in the Nakamoto consensus algorithm: block length, difficulty adjustments, and Halvings.
Block length averages 10 minutes, implying around 144 blocks per day. We define a Block day as 144 blocks and it could be longer or shorter than a Gregorian calendar day, but is typically somewhat shorter.
The difficulty changes how much hash rate miners have to deploy in order to win the Bitcoin mining block subsidy, and the change happens about every two weeks. The Bitcoin fortnight is 2016 blocks in length, or 14 Block days.
The importance of the difficulty adjustment is very clear now, when hash rate has dropped from around 160 Exahashes/s to 110 EH/s due to the Great Chinese Mining Shutdown this month. This loss of 1/3 of the hash rate is mitigated by the difficulty adjustment, with the next one due next week as I write. The next difficulty adjustment downward in about a week will help start to restore hash rate, since marginal older mining rigs become profitable if difficulty drops -15% as estimated. Difficulty is simply a scaled average of the prior fortnight’s global hash rate.
In practice, Chinese Bitcoin pools are quickly relocating their ASIC mining rigs outside of the country, or selling them to non-Chinese mining pools. So much or most of the lost hash rate will come back on line.
Then we have the all-important Halvings. Each 210,000 blocks the block subsidy reward is cut in half. That occurs a little faster than each four years. So we can define a Block Year as 52,500 blocks and it contains 12 Block Months of equal length: 4375 blocks each.
These definitions give us our Block calendar, that in practice has years slightly shorter than regular calendar years. As I write this the latest block height is 688,554. Divide by 52,500 and you see that 13.115 Block years have elapsed. Indeed, 12 Block years elapsed by 5/11/20 when the Third Halving occurred at block height 630,000.
And since then we have 58,554 additional blocks since the Third Halving, so a 13th Block year plus 1.384 Block months have elapsed since then. (We are now in the second month of Anno Satoshi 14 if you will).
Figure 2. Bitcoin chart from 2020, with 200 week moving average. From PlanB.
Long term support at 48 Block Months average
PlanB (Twitter @100trillionUSD) and others have noted that Bitcoin’s 200 weekly moving average has acted as a firm price floor. The price touched a few times in 2015 and at the start of 2019 but never fell below.
What does 200 weeks correspond to? Well since recent Block years have been averaging 50 calendar weeks, it corresponds almost exactly to one Halving cycle or four Block years.
To this end, I have built a long term chart using Block monthly prices and four Block years = 48 Block monthly moving averages. Making this choice allows us to match precisely to Bitcoin’s natural, fundamental, long-term rhythm.
Because Bitcoin is so dynamic, with high volatility, we use the log of prices, log10 for ease of interpretation. Note that over any four Block year period that the price is substantially higher at the end of the period, as evidenced by the steady upward slope of the green line.
Figure 3. A 48 Block month (one Halving cycle) moving average chart, running from 6 Block years elapsed until the present 13.1 Block years elapsed. We do our analysis on log10 price. The gold line is the log10 of price, and the green line is the 48 Block month moving average of the log10 prices. So 3.0 on the y-axis is $1000 and 5.0 is the $100,000 price point. The point 12 Block years corresponds to May, 2020, and the COVID-19 dip just prior to the Third Halving is apparent. The blue line shows the Z value (number of standard deviations away from the green line). Note that the blue line never touches zero; Bitcoin is highly skewed positively (convex). All three curves are read off from the ordinate scale on the left side.
I have also calculated the 48 Block month standard deviations of log10 of price and the corresponding Z-scores on the log prices, that is the number of standard deviations that the gold line is away from the moving average green line.
Table 1. The number of Block monthly data points above a given Z-level. Z below 1 is a favorable buying zone. Below 2 is a good HODL for the long term zone. And above Z of 2.5 one could consider trimming one’s Bitcoin position if uncomfortably large.
In the Table above I have tabulated how many data points lie above a given value of Z in our seven Block year data set. This allows us to identify favorable zones for accumulation and distribution.
Fully half of the points are found below Z = 1, which represents a very favorable zone for accumulation, as one can also see from visual examination of Figure 3. The zone from Z = 1 to 2 could be considered a hold (‘HODL’) zone or mild accumulation zone.
And above Z = 2.5 one might considering trading around your core position, lightening up somewhat. Reaching Z = 2.7 happens less than 10% of the time.
You might do this in conjunction with maintaining an optimal full Kelly or partial Kelly allocation of Bitcoin within your investment portfolio. See details here: https://stephenperrenod.substack.com/p/kelly-capital-growth-criterion-for .
However, please note that it is very hard to outperform a regular dollar cost averaging buy and hold strategy, given Bitcoin’s rapid price rises to date as I showed here: https://stephenperrenod.substack.com/p/outperforming-buy-and-hold .
Note that unlike the stock-to-flow model and my own Future Supply, Lindy, and S-curve (Weibull distribution) models presented in other Substack and Medium articles, this method assumes no model, just price history, regularly updated.
The current most recent month boundary Z-score is 1.74, in the hold zone. Note that the support level is well below the price, and is even below $10,000 at $9333. However, based on the one year slope of the moving average line, it would be projected to more than double to $19,950 a year from now.
And the standard deviation is 0.32 in log10 terms or a factor of 2.08, so a year from now a typical Z =1 score would correspond to $41,600 and a rather elevated Z=2 score would correspond to an $86,600 price level.
When the long term price trend doubles in a year and increases over an order of magnitude each four years, do you really want to be using leverage and taking on the risk of being liquidated? Do you want to lose your position? Or do you want to hold patiently without leverage for substantial long-term gain, also avoiding taxes along the way?