Predicting the price of any cryptocurrency is difficult. But if you’re watching the live charts and looking for signs, it helps to know what kinds of data points are likely to correlate with price, and what kinds of data points are not. Interestingly, this varies quite a bit depending on which coin you’re talking about.
Consider, for example, the chart below, which uses Coin Metrics data to visualize Pearson correlations for the top 20 cryptocurrencies on CoinMarketCap as of January 13 (with the exception of stablecoins and Cosmos, for which Coin Metrics does not have data).
On the scale below, a score of 1 would reflect a perfect positive correlation (the movements of the two variables over time always correlate) and a score of -1 would reflect a perfect negative correlation (the movements of the two variables never correlate).
In this data we can see that Bitcoin, for example, behaves much as we might expect. Both the number of transactions and the number of active addresses appear to correlate fairly positively with its price. This makes sense, of course — the more people are using and transacting with Bitcoin, the higher we might logically expect the price to be. More people and more transactions means greater demand, while Bitcoin’s supply is relatively static.
But it’s interesting to note that this isn’t the case for all coins. For Bitcoin Cash and Bitcoin SV, for example, there seems to be no correlation between the token’s price and network transaction activity. A few tokens, such as Binance Coin, TRON, and Tezos, seem to have a mild correlation between increased transaction counts and price drops.
And while active addresses are at least somewhat correlated with price for most tokens, others — particularly Bitcoin SV, TRON, Tezos, and Huobi token — don’t seem to show any correlation at all between price and active address count. Monero has no active address count due to the way its privacy-focused blockchain handles transactions.
Since each coin’s full history on Coin Metrics is reflected in the chart, comparing them to each other is, admittedly, a little unfair. The chart below, which displays the same thing using data from Jan 1, 2018 through Jan 12, 2020, makes a more direct comparison (although some of these tokens have been around fewer than two years) at the price of reduced historical accuracy for some of the older tokens:
Of course, this doesn’t mean that you should rush to buy BNB, for example, just because you saw a live active address count jump. Correlation doesn’t necessarily reflect causation, after all, and an analysis of historical correlation patterns may not have any predictive power for the future.
Still, it’s interesting to see which token prices seem to move more closely with these fundamental network metrics, and which tokens are more divorced from them.