In the research, the causal relationships between Bitcoin, gold and oil prices were examined. The data of the research covers the period from 2015 to July 2020 and consists of daily price values. Augmented Dickey-Fuller Unit Root Test was used to see whether the stochastic process changes with time. Bitcoin and gold series do not contain a unit root since the oil series is stationary at the level while the difference is stationary. The reason why the series containing unit roots are not stationary is due to structural breaks or not, was investigated by Bai-Perron Unit Root Test with Multiple Structural Breaks. According to the test, it was determined that the Bitcoin series has one break and two regimes, while the gold series has two structural breaks and three different regimes. Whether the research series are cointegrated or not was investigated with the Gregory and Hansen test. The causality between the series was examined with the Toda-Yamamoto causality test, which is based on the VAR (Vector Autoregression) model and examines the causality in the series regardless of the unit root. A two-way causality relationship was determined between the eight lag-long Gold series and the Bitcoin series. In other cases, a causal relationship has not been established. As a result, we give an evidence that Bitcoin and gold prices series followed a parallel pattern while with oil not. Therefore, investors can add Bitcoin into their portfolios to make balance of the risk and return.