Yeah it's a performance trade-off. But to calculate max, min and avg, we do need to access all documents to calculate these values for this particular query at runtime