Establish minimum term count in aggregation, then do more work

I have a bunch of log entries in the form of {"id":"abc123", "delta":7}. I want to do a terms aggregation on id, and then get the lowest standard deviation values where the count IDs is still at least N in length.
So if I have

{
	"aggs": {
		"idcount": {
			"terms": {
				"field": "id",
				"order": {
					"stats.std_deviation": "asc"
				}
			},
			"aggs": {
				"stats": {
					"extended_stats": {
						"field": "delta"
					}
				}
			}
		}
	}
}

I get back the lowest standard deviation of 0, for IDs that occur only once. I want to be able to say "give me the lowest standard deviation where aggregations.idcount.doc_count>30, but still order by the ascending std_deviation calculated in the nested stats agg". Any hints?