Forecasting Gold Prices!
![SHINY.jpg](https://steemitimages.com/640x0/https://s16.postimg.org/dqmovtk2d/SHINY.jpg)
This piece is dedicated for my GOLD bug friends out there, I am going to forecast the gold spot price (XAU/USD) in this article. In my previous article I have forecasted crude oil, which is a very popular traded asset, so to be fair, I am going to forecast gold too for my gold bugs out there. Here is the oil forecast if you are interested:
https://steemit.com/money/@profitgenerator/the-oil-market-3-3-forecasting-crude-oil-prices
I have almost 40 years of historical data in daily timeframe, so it should be as accurate as it gets. Well it could be more accurate on tick based historical data, but that would cost like 3000$ for high-quality resources like that, so unless somebody donates 3000$ or a tick based historical data, I am going with daily timeframe.
Also remember that the forecast doesn't tells us the exact future price, but rather the future estimated mean of the price and a statistical boundary between the price should move in the future. There is no way to forecast the price accurately, otherwise I'd be a trillionaire
, but this forecast could still be useful so to know what to expect in the future.
FORECAST
![GOLD.png](https://steemitimages.com/640x0/https://s16.postimg.org/eakrp0g5x/GOLD.png)
This is how the gold market looks like over almost 40 years, it had quite a ride since the 400's, it has a median of 400.20$
so it means that the price had hovered mostly around that level, and going up until 1888.7$
and with a minimum of 253.00$
, now being at the 1200$ levels, let's see what the future holds!
I am using ARIMA model to forecast, as it is a very popular quant tool, and pretty robust. I am not going to bore you with details so I did the calculations myself, the best ARIMA model for this data is ARIMA(2,2,2) with constant
with a Mean Absolute Error of 5.0921
, multiple models could be correct, and there could be an even better model, but this is what I have found.
![GOLDRES.png](https://steemitimages.com/640x0/https://s3.postimg.org/z8hqvs8wz/GOLDRES.png)
The residuals conform very well to the mean with deviation of only 0.80075%, so it will be probably as accurate forecast as my crude oil forecast.
![80-20 FORECAST.png](https://steemitimages.com/640x0/https://s11.postimg.org/3x4kg6rpv/80_20_FORECAST.png)
Then I like to forecast the end 20% of the data from the previous 80% just to make sure that the model is Ok. Well it's pretty fair, no model will give you a more accurate mean when there is a turning point like that at the 1800$ levels. The Mean Absolute Error is 282.33
which is pretty big, but it shows the uptrend well, and as I said, no model can predict turning points like that.
![365 FORECAST.png](https://steemitimages.com/640x0/https://s17.postimg.org/5ze5ampqn/365_FORECAST.png)
Then we forecast 365 days into the future, and we get this, it predicts a price increase. In 365 days the price should have a mean of:
1355.64$
(With 95% confidence between 1022.44$ - 1688.84$)
So that is the gold market for you, the research calls a bullish year, and let's hope it will be true as I have invested in gold!
Disclaimer: The information provided on this page might be incorrect. I am not responsible if you lose money using the information on this page! This is not an investment advice, just my opinion and analysis for educational purposes.
How do you resolve this comment:
Satoshi Nakamoto: “The price of any commodity tends to gravitate toward the production cost.”
I don't know ,I don't really keep up with commodities lately, but I don't think its entirely true.
It is true that producers who have a monopoly (oil) wont sell their commodity below production cost, but if there is no intrinsic demand for it, then it may very well fall below that.
Gold & Silver is pretty speculative, and it could very well fall below production cost, once cryptocurrencies replace them (as their industrial use is not that huge) compared to Oil that has a huge demand (Cars, Heating, Plastic, Medicine, etc)...
Besides the production cost is always rising due to inflation, and the commodity is inflating too,so it is hard to tell. It has to be researched.
For things like grain the production cost went down massively since the ancient times, yet the price stayed pretty much the same. A price of bread hardly changes except in inflation, shortage or famine.
heyy nice charts!! you make it in R ?
i am interested abouy historical data ....
meybe we can shared..
greatings
Hello, no I have my python scrips for market analysis.
But I am very busy these days so I only have longterm positions now (3-6 months). This year I will be very busy. I am getting married soon, so maybe in autumn I will have more time to talk about charts and markets, I am now just basically coming to steemit just to relax and talk about other stuff. I am too busy now to be stressed about short term investments.