I would like to share my brief research in oil market from fundamental and price-action perspectives. First at all, I will mention principal, but non-exclusive, factors which affect Oil prices. Finally, I will introduce a rolling estimation of oil prices based on the assumption of log normal daily return distribution.
Factors that influenced the evolution of Oil Price
It is stated that Crude Oil Prices are affected by many physical markets factors as well as those related to trading and financial markets. Also, supply and demand stability drives spot prices in regards of what OPEC and Non-OPEC countries perform.
Forecast Oil Price
For forecasting WIT, I assumed Log normal return distribution. I also use a rolling moving average of 10 days and a rolling standard deviation of 10 days to calculate returns and thus, estimate WTI prices:
From the chart above, price estimation seems to be accurate at some point, as long it is restarted the forecast each two weeks. It exploits time series properties such as stationarity and log normal distribution (assumption) of returns. Rolling analysis presents some advantages rather than evaluate all-data time series giving the notion of predicting some information in the future based on past information.
Looking for factors that affect WIT prices I took one that is constantly highlighted in news headlines, U.S. Crude Oil Production. Back in 2012, the fracking revolution in the US started to boost its economy since they started to produce enough crude oil to supply their demand. That is why there was an imbalance which ended up in oil surplus and low prices. It is observable that increasing crude oil production in the US, make surplus be larger and consequently decrease WIT prices.
Further analysis should be made to confirm two results resented: First, rolling windows and other distributions could be used to enhance forecasting results and exploiting them for trade signals. Second, correlation and cointegration analysis could be performed to check the relationship between WIT price and U.S. Crude Oil Time Series in a more rigorous approach.