Whale Watching in Ethereum to Find Tokens to Invest In
As we noted in a previous post, our interest in cryptocurrencies as an investment opportunity stems largely from our experience in equities, in which 13F and related filings have proven to provide an informational edge that can be mined for outperformance. The blockchain associated with many cryptocurrencies seemed like a parallel, but even better — given that blockchains tend to provide a stream of transactions across all holders, and thus can be mined in greater detail to understand the buying and selling activity of the “smart money” behind a given currency.
This post marks our first foray in attempting to identify and analyze “crypto whales” — wallets that have large positions in cryptocurrencies, under the hypothesis that doing so may inform a portfolio strategy that results in outperformance. We started our exploration by observing a sample of Ethereum wallet addresses that hold at least $100,000 USD worth of ERC-20 tokens (excluding Ether). We excluded our list of identified multi-user wallets and exchanges, wallets known to be used for burning units of currency, and wallets that were identified to be the top-ranked holder for a given currency — as such wallets are often the default holder and issuer of said currency. Our sample consists of 22,866 wallets with positions in 736 ERC-20 tokens.
Below is the distribution of currencies the whales currently hold positions in.
As one might expect, there is a fairly strong correlation between the market capitalization of a given currency and the number of whales that hold it; we found the correlation to be .58, which is fairly strong (as a frame of reference, the correlation between human height and weight is approximately .71 — this is a very strong correlation, and is about 22% higher than the correlation we observed between whale count and market capitalization).
We then quickly developed a model to predict the number of whales based on a token’s market capitalization in US dollars. This model is not extremely accurate (r-squared of .48, suggesting almost half the variation in whale count can be explained by market capitalization), but we thought it may be worth exploring which tokens possess far fewer, or far greater, whales than expected. Those results are visualized below. Tokens for which the model’s predicted number of whales was more than 100% away from the actual number of whales are denoted in orange.
We are still conflicted as to what this might mean, and generally believe that further exploration is required. One hypothesis is that a less than expected number of whales may reveal an economy that is less subject to overconcentration and the manipulation and political instability that can result; this viewpoint may be more amenable to those who view cryptocurrencies as more akin to nation-state currencies in their value and behavior. Conversely, for those who prefer an equity-like view, one could argue that a lack of whales reflects a market with less sophisticated users that may be more susceptible to being shaken out of their positions, and may exhibit noisier price patterns as a result. And of course, a third explanation is that it means nothing. Nonetheless, we plan to explore this issue in a subsequent post, as we suspect something can be learned from the process of doing so.
The viewpoint we found most interesting can be found below. It summarizes the net change in currency held by whales since August 1 for each token observed whose market capitalization in US dollars as of September 10, 2018 was greater than $500,000. Green dots indicate where whales were net buyers, with the y-axis representing the percentage of total currency that the net change in whale holdings represents. The y-axis represents the approximate market capitalization as of September 10, 2018. If the “follow the whales” hypothesis is applicable to cryptocurrencies, instances in which whales are behaving in contrast to price may reveal opportunities for outperformance.
For more our analysis, follow our Medium blog. For summaries of articles shared by the most influential voices in the cryptocurrency world on Twitter created by our artificial intelligence, follow SixJupiterAI on Twitter. And for custom data analysis or access to our Ethereum One data set, visit the SixJupiter web site.