I was wondering how built-in "dev" function in Pine is calculated so I made a little research.
I examined 7 samples: 0) "dev" function itself 1) "dev" according to its description: series - sma(series) 2) Mean Absolute Deviation 3) ratio of the absolute difference from 1) divided by period 4) ratio of the difference from 1) divided by period 5) Median Absolute Deviation 6) tricky for-loop to calculate Mean Absolute Deviation
The results of the null and sixth samples are identical.
So, TV built-in "dev" function represents Mean Absolute Deviation and it's description is incorrect.
Where it is used? For example: Commodity Channel Index. You can check its original formula and if you used simple standard deviation instead of MAD in your CCIs - well guys, you were wrong.