trying to find the probability of a certain price in the oil market being touched within a time frame

#1
I have built a model that predicts closing prices of the oil market. My goal is to make a model that can help find the probability of reaching a certain price within a certain time frame but I am lost. For example, if oil is at 60.00 a barrel, I would like to know how to find the probability of it reaching 63.00 within 14 days. I am not sure which probability model to use for this. I have thought about using a logistic regression model; however, that does not seem to work because of the binary choices that I have for the dependent variable. I thought about drawing trendlines on oil prices and calculating the probability of a reversal once a support/resistance line is touched by a candlestick. I would just identify the candlestick reversal pattern, among other momentum oscillators, and give it a 1 or 0 based if it reversed. The problem with my current time series model is it only gives the probability on a certain date of reaching a price when I need the probability of it occurring within a time frame. I thought about the exponential probability distribution but that doesn't seem practical.
Any help to identify what type of model or prob/stat method would help me with this.
 

hlsmith

Less is more. Stay pure. Stay poor.
#2
Some type of bayesian time series or Monte Carlo simulation would allow you to see how often forecasts reach that threshold, which could be used to calculate a probability via the frequency of the occurrence.

Logistic isn't traditionally used in time series, which is what you are describing.
 
#3
Thank you for your reply.

Yes, I am aware that logistic is a regression concept and not a time series concept. I was going to use the logistic regression model to calculate the probability of a reversal at support/resistance lines since I was having trouble trying to find the probability of the price reaching a certain time frame within a certain time frame.

I will certainly look into a bayesian time series model; thank you
 
#4
How accurate is the model that predicts closing prices on the oil market? How does it hold up with unknown unknowns e.g. war breakouts, pandemics, subprime mortgage collapses?