Disliked{quote} Curious to hear how you are planning to use it. From what I see in the marketplace, it is a valuable skill to have for sure!Ignored
Self-sufficiency is the greatest of all wealth. - Epicurus
- #42
- Edited 9:51pm Sep 8, 2019 9:39pm | Edited 9:51pm
I came across an interesting script to view at seasonal returns using violin charts. Violin charts work like this: 
- Length of violin represents the max and min returns during the period.
- Width of violin represents the frequency of returns. If the widest point of the violin is around 3%, then you know 3% has historically been the most common return.
- There's a black dot in the violin that represents the median return.
I like this visualization a lot because instead of just showing the average monthly return, it also shows the distribution of returns.
I decided to run it on GLD and SLV to see how they look. I'm still getting used to this software so bare with me on the graphics, but here are the plots:
Self-sufficiency is the greatest of all wealth. - Epicurus
One interest of mine has been volatility indexes like GVZ. Used correctly, I think studying volatility is useful and can offer some predictive value. I routinely use the VIX index to enter into trades so I'm going to see what I can uncover using metals volatility indexes. 
First, I did a study to see if there was any predictive value when you compare gold's volatility with silver's volatility. In theory I think an extreme reading might be indicative of a trend change. Unfortunately, comparing gold and silver volatilities didn't yield any valuable insights.
The below charts show gold at the top and silver volatility / gold volatility at the bottom. 2017 and 2019 were kind of interesting, but the other years not so much.
First, I did a study to see if there was any predictive value when you compare gold's volatility with silver's volatility. In theory I think an extreme reading might be indicative of a trend change. Unfortunately, comparing gold and silver volatilities didn't yield any valuable insights.
The below charts show gold at the top and silver volatility / gold volatility at the bottom. 2017 and 2019 were kind of interesting, but the other years not so much.
Self-sufficiency is the greatest of all wealth. - Epicurus
I was going through an R course tonight and saw something pretty neat. With just a few lines of code you can easily calculate and plot drawdowns for any asset or trading strategy. Here's BTC for example:
I knew you could calculate this in R, but figured it would take a lot more code. What a great language for finance.
I knew you could calculate this in R, but figured it would take a lot more code. What a great language for finance.
Inserted Code
library(quantmod)
library(PerformanceAnalytics)
btc <- getSymbols("BTC-USD", auto.assign = FALSE)
btc_ret <- periodReturn(btc, period = "monthly")
chart.Drawdown(btc_ret, main = "BTC Drawdowns")
table.Drawdowns(btc_ret) Self-sufficiency is the greatest of all wealth. - Epicurus
- #45
- Edited 10:42pm Feb 12, 2020 10:00pm | Edited 10:42pm
I was playing around with some forecasting functions in R tonight and made a quick forecast for the US Unemployment Rate using ses():
The naive() function produces very similar results so I'll probably use both in the future.
The holt() function looks even better than ses() and naive(). Here's what it's forecasting for New Home Sales for the next 12 months:
The naive() function produces very similar results so I'll probably use both in the future.
Inserted Code
library(quantmod)
library(forecast)
UNRATE <- getSymbols("UNRATE", src = "FRED", auto.assign = FALSE)
forecast <- ses(UNRATE["2000/"], h = 12)
autoplot(forecast) The holt() function looks even better than ses() and naive(). Here's what it's forecasting for New Home Sales for the next 12 months:
Inserted Code
library(quantmod)
library(forecast)
NHS <- getSymbols("HSN1F", src = "FRED", auto.assign = FALSE)
forecast <- holt(NHS["2000/"], h = 12, damped = TRUE)
autoplot(forecast) Self-sufficiency is the greatest of all wealth. - Epicurus
- #46
- Last Post: Feb 15, 2020 1:25pm Feb 15, 2020 1:25pm
I was studying more forecasting with R and came across something neat. Using the ets() function, R will pick the best model for you automatically. 
Here's the code I'll use to forecast next week's Unemployment Claims data:
Here's the code I'll use to forecast next week's Unemployment Claims data:
Inserted Code
library(quantmod)
library(forecast)
ICSA <- getSymbols("ICSA", src = "FRED", auto.assign = FALSE)
ets <- ets(ICSA)
autoplot(forecast(ets, h = 12))
forecast(ets, h = 1) Self-sufficiency is the greatest of all wealth. - Epicurus