backtesting and out of sample test of portfolio optimization

Hello I am new on the forum, I wanted to ask you a clarification on the topic without go into much details just if i am on the right path.

So i have a dataset from 1st jan 2016 till 31st dec 2018 of daily return for 30 stocks, i calculate my portfolio optimization with a sample that goes from 1/1/16-31/12/2017, thus i find the optimal weights. Now what i would like to do is backtesting or insample analysis, (are those different?), and i would like to do it for 1 year 1/1/2017-31/12/2017. Then for the out of sample i would like to use a rolling window of 1year and estimate the new optimal weights every day. does it make sense to you? is there some better options i could take in order to validate my findings?