James Stanley

Using Google Trends data to speculate on cryptocurrencies

Sun 27 August 2017

I did a little analysis this evening to try to find cryptocurrencies that are under-valued relative to Bitcoin, based on Google Trends data, with the intention of buying whatever is under-valued and waiting for it to go up. I found some possibilities and I spent some money.

(Note: you may have seen a previous version of this article that said every cryptocurrency is over-valued relative to Bitcoin. My maths was wrong. I've fixed it now. The difference was big enough that I actually spent some money, so this post is what that post should have been).


The hypothesis here is that if the Google Trends score for (e.g.) "buy ethereum" is 50% of the score for "buy bitcoin" (i.e. about 50% as many people are looking to buy Ethereum as are looking to buy Bitcoin), then the Ethereum market cap should be approximately 50% that of Bitcoin. If the market cap is too low then it is under-valued and I should buy some Ethereum.


We're looking for cryptocurrencies with a "market cap over-valued factor" of less than 1, i.e. under-valued. Here's a table of my data:

CoinGoogle Trends scoreScore relative to BitcoinMarket capMarket cap relative to BitcoinMarket cap over-valued factor
Bitcoin Cash80.11$10,200,000,0000.141.34

(I excluded Ethereum Classic and ZCash because they scored 0 on Google Trends, which would mean the market cap is over-valued by a factor of infinity. Which might be accurate... but this isn't the post for that kind of discussion).


So the promising contenders here are Ripple, Litecoin, Dash, IOTA, and Dogecoin. With a factor of 1.02, Monero appears to be priced about right (as an aside: at my previous employment I wrote a series of internal analyses in which I said Monero looked under-valued; it has since increased by a factor of ~3 over the past month). Bitcoin Cash appears a little over-valued, and Ethereum appears to be heading for a substantial drop.

Since this is all relative to Bitcoin, it doesn't give us a figure for how much Bitcoin might be over-valued. But we can get a rudimentary estimate by comparing it to the average of the others: the mean "market cap over-valued factor" is 0.94, which might suggest Bitcoin is over-valued by ~6%. Weighted by market cap, the average is 1.28, which might suggest Bitcoin is under-valued by ~28%. Both of those are close enough to 1.00 that I'm not going to worry about it.

It is also quite possible that these "market cap over-valued factors" will correct towards 1.00 not by the price changing, but by the Google Trends scores changing. I'm also not going to worry about that.


Just in case this analysis might work, let's allocate a small portfolio to it and see how it performs over the next couple of months. I don't actually have any faith at all in the fundamental proposition of Ripple, Dash, or Dogecoin, and I have some doubts about IOTA. But I'm not here to talk about fundamentals, I'm here to talk about Google Trends data, and I'm here to speculate, so we'll include them regardless.

I spent $100 on each of those cryptocurrencies, and, after trading fees, transaction fees, and unfavourable exchange rates, the portfolio looks like this:

IOTA107 M$98.44
Total-$495.62 (-0.9%)

If everything in this post is correct (a big if!), then we can expect each of those coins to increase in value such that the "market cap over-valued factor" is 1.00. If that happened, and the relative Google Trends scores remained the same, then the portfolio would be up by ~120% (or ~50% if you don't count Dogecoin). If it comes anywhere close to that, I'll be delighted.

I intend to post updates on this whenever there is anything interesting to say. Value goes up, value goes down, Google Trends scores change and I change the coins in the portfolio, that sort of thing.

Also, for the avoidance of doubt: this post is not investment advice. Don't do as I do, don't do as I say, and don't complain when you lose all your money.

If you like my blog, please consider subscribing to the RSS feed or the mailing list: