Dear contributors, dear Alexios Galanos, as suggested by Alexios I'm using the var_cp_test from the tsmodels/tstests package over the VaRTest in the rugarch package to backtest a calculated value of risk vector against my log returns.
This works well in most cases. However, in some situations, the test returns an subscript out of bounds error. I found an example where the test returns no error for the full vector of VaRs and log returns but when I perform the test without the first value in the vector, it returns an error. There is no obvious reason why. Data for the example can be found here: https://drive.google.com/file/d/1na5w-15U70RtRA6Mh2jpytDDCafacYHG/view?usp=sharing This is my code: ``` if(!("tstests" %in% installed.packages())){ remotes::install_github("tsmodels/tstests", dependencies = TRUE) } library(tstests) ## load the data current_path = rstudioapi::getActiveDocumentContext()$path project_folder = dirname(current_path) # define the project folder path setwd(project_folder) # change current working directory filename = paste("report_bug_var_cp_test_subscript_out_of_bounds.csv",sep="") df_bug = read.csv(file = filename) rownames(df_bug) = df_bug[,1] df_bug[,1] = NULL ## select forecast and actuals forecast = df_bug[1:nrow(df_bug),"VaR"] actual = df_bug[1:nrow(df_bug),"log_returns"] ## run the var_cp_test with full data var_cp_test( actual=as.numeric(actual), forecast=as.numeric(forecast), alpha=(1-0.99), ) ## run the var_cp_test with all data but first values forecast = df_bug[2:nrow(df_bug),"VaR"] actual = df_bug[2:nrow(df_bug),"log_returns"] var_cp_test( actual=as.numeric(actual), forecast=as.numeric(forecast), alpha=(1-0.99), ) ## error message: # Error in `[.default`(tab, 2, 2) : subscript out of bounds ``` Thanks a lot in advance! Best, Ayla [[alternative HTML version deleted]] _______________________________________________ R-SIG-Finance@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-sig-finance -- Subscriber-posting only. If you want to post, subscribe first. -- Also note that this is not the r-help list where general R questions should go.