Moving window impact for features calculation.

[apalaqa]

Hello, I try to reproduce your work but I faced an issue in features calculation.
• StdRushOrders and AvgRushOrders: Moving standard
deviation and average of volume of rush orders in each
chunk of the moving window.
Please correct me if I'm wrong.
To calculate these features should I use only .std() and .mean() on the chunk of data? I can't understand the impact of moving window here. Where do we use this window?
• StdTrades: Moving stardard deviation of the number of
trades, both buy and sell.
To calculate this feature should I use .std() on the chunk or on the whole window?

Many thanks in advance