Predicting Short-Term Bitcoin Price Fluctuations from Buy and Sell Orders
The purpose of this paper is to analyze whether order book (buy and sell) data can be used as a short-term predictor of Bitcoin (BTC) volatility against the US Dollar. A temporal mixture model is used to capture the dynamic effect the order book has on the price volatility of BTC. This model is tested against more traditional models such as time series or ensemble models and is shown to provide more robust results. Order data from over a year-long period is obtained from one of the largest bitcoin trading offices from 2016-2017. This study also provides insight into how specific features of the order book, such as spread and volume, affect short-term volatility.
This paper considers hourly price volatility of BTC prices which refers to the standard deviation of minute returns over an hourly time frame. The data set contains time series of hourly BTC volatility data for over one year. Order book data is obtained from one of the largest digital asset trading platforms, which does about 39% of all BTC trading volume and is therefore considered representative of the market. Standard time series models have a limited ability to combine volatility and order book features, so a temporal mixture model is used. This predictive model incorporates volatility series and order book features into forecasting future price movements. Mixture models are commonly used in data science as they allow for flexibility in the specification of each component model.
Traditional time series models are tested however they are unfit to represent the dynamical, time-varying relationship between order book data and volatility, and therefore do not produce straightforward results. The mixture model does however capture the effect of order book features on the volatility of BTC and provides straightforward results. The results indicates there is an importance and interaction of the order book features for high volatility regimes, therefore it implies that order book features can be an indication of short-term price volatility. Read the full study by Tian Guo and Nino Antulov-Fantulin here.
Start Algo Trading Today!
Get your free token to connect to the REST API and start trading. To generate your token:
Risk Warning: The FXCM Group does not guarantee accuracy and will not accept liability for any loss or damage which arise directly or indirectly from use of or reliance on information contained within the webinars. The FXCM Group may provide general commentary which is not intended as investment advice and must not be construed as such. FX/CFD trading carries a risk of losses in excess of your deposited funds and may not be suitable for all investors. Please ensure that you fully understand the risks involved.