Running an aggregation pipeline composed by the following $stdDevPop and $stdDevSamp does not produce the expected output.
If we have a collection where documents contain individual "age" field and we want to calculate the standard deviation ($stdDevPop) for a given selector (city) the pipeline should be the following:

db.ages.aggregate([ {$group: { ages: {$push:"$age"}, _id:"$city"  }}, {$group: { _id: null,  std: {$avg: "$ages"}  }}   ] )

but the output I'm currently getting is the following:

{
  "waitedMS": NumberLong("0"),
  "result": [
    {
      "_id": null,
      "std": null
    }
  ],
  "ok": 1
}

Running the same operation but using an intermidiate collection does perform the expected result:

db.ages.aggregate([ {$group: { ages: {$push:"$age"}, _id:"$city",  }}, {$out:'aaa'}] )
db.aaa.aggregate( {$project: {_id: "$_id", s: { $stdDevPop: "$ages"}}})
{
  "waitedMS": NumberLong("0"),
  "result": [
    {
      "_id": "NYC",
      "s": 29.037802755871343
    }
  ],
  "ok": 1
}

Also tested with $avg with the same issue.

Seems like there is some issue with pipelined $group operations