Pelman08


Total Posts: 6 
Joined: Aug 2019 


Hello I have been working on a Markov switching GARCH model my intention is to use it to trade options volatility . I have created a Markov switching garch model using the MSGARCH package in R and in the example below I have used it to generate a TGARCH model with a normal distribution(I know that is not what I should be using this is just an example ) to model the daily returns of the QQQ index .
``` library(quantomod) library(MSGARCH) getSymbols("QQQ") qqq_rets=dailyReturn(QQQ$QQQ.close, type="log") mods= CreateSpec(variance.spec = list(model = "tGARCH", distribution.spec = list(distribution = "norm"))) as=FitMCMC(mods, qqq_rets) ```
I am trying to figure out how to access how good my model is I have seen VAR used as a measure of how good a Garch model is I Ould like to know why it is used and if there are any alternative measures to evaluate how good my garch model is and if the kind folks on nuclearphynance can also provide me with some code that would be very helpful




doomanx


Total Posts: 89 
Joined: Jul 2018 


Compare it to random, then compare it to the simplest nonrandom model you can think of. That should get you started. 
did you use VWAP or triplereinforced GAN execution?

