(Image Modified from Pixabay)
About a month ago I made two especially strong claims. I argued both (1) that Bitcoin would see choppy returns for several weeks going forward, and (2) that Bitcoin would fare well longer term, i.e., over the next twelve or so months.
The analysis not only linked my claim to specific price movement, but I did so in a specific time frame. And I was right as far as can be tested (news for claim (2) will have to wait a year).
As a bonus, I also argued that proxy plays, such as Overstock (OSTK) would not fare well without some sustained Bitcoin price movement. I was right about this too, as OSTK has declined from $15.09 to $9.86 at the time of writing.
To make such strong claims, and to be right about them, you need a strategy that works and comes with reasons to exit. This is the “being proven otherwise” part that’s relevant to the title.
In the present article, using the same three indicators as my previous, I outline why I think the situation has improved from about a month ago. The article closes with a note on why proxy plays, such as Overstock (NASDAQ:OSTK) but not Grayscale Bitcoin Trust (OTCQX:GBTC), fare so poorly in the present environment.
I begin with the conceptually simplest of these indicators: sentiment.
The idea behind sentiment in the case of Bitcoin is straightforward: buzz drove the previous bubbles, so buzz (as an indicator of investor sentiment) will probably drive this one.
Predictive or fast responding sentiment indicators are difficult to develop, and often have inconsistent results. Richard Peterson, who developed the TRMI-Index (for trading stocks by sentiment indicators from social media sites) eventually closed his own hedge fund and decided to just sell his data as a product to other hedge funds in part because of uneven returns (also managing a hedge fund involves quite a bit more than making mathematical models). You can read about it in his book Trading on Sentiment: The Power of Minds Over Markets.
Slow sentiment indicators, the sort that you get by looking at Google Trends, by contrast, can be useful to confirm a broader interest. You can augment Google Trend analysis with some wider searches (I find looking at sub-Reddit trends to be more informative than Twitter), but when I’ve done this for my own purposes, I’ve not seen any materially different outcome. Below, then, is an image of Bitcoin sentiment as indicated by Google search traffic.
What’s noticeable is that for the first time since the crash, Bitcoin search traffic has hit the level that marked the beginning of the 2017 run up. That means Bitcoin is buzzing again.
This is undoubtedly a positive indication for Bitcoin, so let’s look at the second indicator.
Network (Metcalfe) Value
The idea behind using (the first) of Metcalfe’s Laws is to see if it might be possible to construct something like a Price to Earnings (P/E) ratio for Bitcoin. Scaling price to something like earnings would enable one to determine whether that price was higher than or below Bitcoin’s actual value (if there is such a thing for Bitcoin).
An intuitive way to do that is to suppose that the value of a network is equivalent to the square of its nodes, scaled by some constant k that is determined through regression analysis (i.e., Value = kN^2). Basically, imagine a phone network with just two people who could call each other. If you added another person, so that you can call two rather than just one person, then the new network value should be exponentially (by a square) more useful.
When adapting the model to Bitcoin, the problem is that it just isn’t clear how to measure nodes. Two common proxies are unique daily addresses, and average daily transactions, but data acquisition for both is a non-trivial task. The first of these is likely to be affected by the development of the Lightning network that allows for off-chain transactions, so that its numbers are too low. The latter is likely to have distorted numbers because of the existence of mixers that create artificial transactions, so that its numbers are too high.
This is exactly what you see in the charts below (from Blockchain.com). The first of these shows the peak of unique addresses at the end of 2017 (the big spike) and where levels sit today.
The next chart shows the same timeframe, but measures by average daily transactions.
The story is clearly much better using average daily transactions, but since this data is likely to be corrupted multiple ways, I would in general have less confidence in it.
The best data likely available (and they’re not cheap) is that from Coinmetrics (shout out to Non-Correlating Stock Ideas for letting me know about this!). The following image is their “cleaned” image for daily transactions, smoothed for a 7-day moving average.
This chart confirms the general uptick in daily transactions that one gets from the Blockchain.com info. So this story looks good.
The following chart gives you an intuitive idea of what’s going on with Bitcoin. Cleaned active addresses are the red line, and the pink line indicates the price per coin of Bitcoin. You’ll notice that active addresses have genuinely outpaced the rise in Bitcoin price … meaning that Bitcoin is in pretty good shape.
Where does that leave us? In a somewhat better place than the analysis showed early May. If pricing Bitcoin relative to the size of its existing network makes sense to you, then the current price to network value (proxied by active addresses), or P/MV, is indicating fair value.
Finally, let’s look at momentum. The alpha of 200-day momentum of almost any asset testable has proven remarkably well-founded (read “Triple Momentum Investing: Even Higher Returns at Lower Risk” for a review of the scientific literature surrounding momentum investing).
Looking at the chart below (black line is the 200 day simple moving average or SMA), you should have bought Bitcoin.
I ran a backtest using the 200 day SMA, starting out with a nice 42.9% decline after the 2011 Bubble, and used the following as my strategy parameters:
If Price of Bitcoin > 200-day SMA, then buy or hold Bitcoin.
Else Sell or hold cash.
Even with that terrible start, the strategy returned too much money (it turned $10,000 which then declined 42.9% after its first trade into $11,537,295.27). It had exactly an 8/16 win ratio. That means it has an enormous average return per trade, making the indicator significant.
Don’t expect those returns going forward. Do expect that so long as Bitcoin remains the volatile “asset” we know it is, that this is a strong indicator in its favor.
So unlike the earlier part of May, when I rightly predicted that Bitcoin would hit some choppy waters for the next couple of weeks, Bitcoin looks to be in a more steady position now.
I nevertheless retain the caveat that should price action drop below the 200 day simple moving average (200 SMA), then the situation has deteriorated materially.
Bonus Analysis: Overstock
I argued before that price movement in proxy plays for Bitcoin were likely to be muted, and one should stay away for the moment. This turned out exactly right. See the chart below.
The reason Bitcoin price increases don’t result in Overstock price increases is double. First, Bitcoin makes up only a portion of OSTK’s value. Second, as a stock, OSTK is subject to the general worries affecting the broader market. It’s not, in short, a wholly independent asset class yet … though that might signal a mispriced buying opportunity other things being equal.
That’s all for now, and I look forward to your comments as always!
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.
Additional disclosure: I do own Bitcoin.