https://www.cnbc.com/2018/08/08/yal...ategy-for-buying-bitcoin-ripple-ethereum.html
The paper's authors, economics professor Aleh Tsyvinski and economics Ph.D. candidate Yukun Liu, sought to "formulate and investigate potential predictors for cryptocurrency returns," according to the paper, and analyzed years worth of past price data for Bitcoin, Ripple and Ethereum. (The prices studied for bitcoin span from 2011 to 2018, while Ripple's XRP and Ethereum's ether data begins at the newer currencies' inceptions in 2012 and 2015.)
Historical data is not a guarantee for an investment's future performance, and Tsyvinski and Liu aren't giving financial advice, but their research reveals two factors that can be meaningful pricing tools for predicting bitcoin's next move.
1. The "momentum effect"
The first significant factor is momentum: The report found that if the price of bitcoin increased sharply over a week, it would be likely to continue to increase for the following week, Tsyvinski tells CNBC Make It.
"Momentum is actually something simple," he says. "If things go up, they continue to go up on average, and if things go down, they continue to go down," at least in the short run. Momentum has been documented in mainstream assets like stocks, bonds and currencies, and Tsyvinski says the pattern holds true in cryptocurrencies.
In fact, the best historical strategy would have been to buy bitcoin after its price already had a sharp increase — 20 percent in a single week — and sell just seven days after buying, Tsyvinski and Liu concluded.
Following that strategy, "the investor would have made an 11 percent [return] during the periods we looked at," Tsyvinski explains. The momentum effect was stronger for bitcoin than for ether or XRP, although still statistically significant, according to the report.
2. The "investor attention effect"
Second, Tsyvinski and Liu found the amount of interest and hype around cryptocurrencies, measured by investors searching and posting online, was a significant predictor of price movements.
Looking at analysis of searches on Google for bitcoin, the report concludes, "for weekly returns, the Google search proxy statistically significantly predicts 1-week and 2-week ahead returns." That means more searching online about bitcoin was a leading indicator that the price of bitcoin would increase in the coming weeks.
For Ripple, "the Google search proxy statistically significantly predicts 1-week ahead returns," and for Ethereum, "the Google search proxy statistically significantly predicts 1-week, 3-week, and 6-week ahead returns."
