Getting rid of transient state


I'm doing some simulations. I have some time series data. The thing is that, when I plot the moving average graph, there is a transient period, and I wish to eliminate those initial values from my output.

I know there is no exact method of determining how many data points to eliminate. However, what is a good approach to select how many of them to delete? I'm looking for something better than just looking at the graph and guessing some numbers (because that is not working quite well). However, I'm not looking for a professional or detailed method either. I'm looking for something in between. Any ideas?

Someone told me about envelopes around the mean, what is that?

Thanks a lot!!! :)


TS Contributor
What I've done in the past is to simply not start the moving average plot until the transient state has gone by. In other words, if you're using, for example, a 5-point moving average, don't start the moving average plot until the 6th point....