In the past year between me and colleagues, a dispute arose regarding the relationship between bitcoin and ether. “Almost all crypto-currencies rise and fall synchronously and cointegration are violated only a few times a year,” I said. Whether so it actually?
In this blog post will present the results of analysis of the presence of cointegration between bitcoin and ether. Recall that cointegration is essentially a regression of a nonstationary time series. I was asked to investigate two time period — at 01.12.2017 30.01.2018 and 07.02.2018 for 03.04.2018 for couples USDT_BTC and USDT_ETH on the exchange Poloniex. Also I will give the results of the analysis for the whole of 2018.
Start with the period between 01.12.2017 and 30.01.2018. Chart of bitcoin:
Schedule of broadcast:
Now draw the spread of two couples: USDT_BTC — 14,7609 * USDT_ETH.
As you can see from the last graph, there is no linear relationship between bitcoin and ether no. If such a linear dependence has been the spread would fluctuate around zero, but here he just falls down trend. This does not mean that between bitcoin and ether no dependencies. Perhaps it is, but what is it, and how you will trade?
Looking at the spread chart, you can’t just take to get 1 bitcoin in position 14 esters in the opposite position and catch inefficiencies in the market. So it’s just not working. You will lose money.
Let’s now look at charts with 07.02.2018 for 03.04.2018. Bitcoin:
And their spread USDT_BTC — 12,5207 * USDT_ETH:
Again we see that spread out far from zero. This means that between bitcoin and ether no linear relationship. The air is falling faster than bitcoin, so it is stupid to take in long 1 bitcoin and short esters 13, hoping for a mean reversion. Again, I’m not saying that between bitcoin and ether there is generally no relationship, I’m just saying that simply arbitrarily plus will not work here.
And finally, let’s look at the graphics for the whole of 2018. Bitcoin:
Their spread USDT_BTC — 13,3542 * USDT_ETH:
In short, be careful. In order for this pair was successful and just arbitrarily, spreads should fluctuate around zero. The graphs (not only graphs, but also when testing statistical tests for stationarity and cointegration), we see a far from ideal picture. So far, I would not recommend you without training arbitraging bitcoin and ether.
Cryptocurrency can really rise and fall synchronously, but they often rise and fall at different speeds, and this difference in speeds may render you bad service.
The message How to trade bitcoin and ether appeared first on crypto.by.