What is index structure of Multikey index for embedded attribute? Why indexsize of multikey index is less than simple index?

Working with Data
1

What is index structure of a Multikey index , which is automatically created for embedded attribute?

It is observed that indexSize of multikey index is less than simple index. It must because of its indexing structure.

For example, there is a collection R with 10^7 documents which has an attribute Rid, which has unique values. Another collection L which has 10^5 documents and each this document contain an array of on average 100 R documents. That means simple index on R.rid have 10^7 index entries as well as multikey index on L.rDocs.rid also points to 10^7 index entries, then Why index size of L.rDocs.rid is less than R.rid?

Sample L Documents : {Lid:1, rDocs:[{rid:1},{rid:2},{rid:3}]}, {Lid:2, rDocs:[{rid:1},{rid:2},{rid:3}]}
Sample R Documents :{ rid:1, rfield: 10}, {rid:2, rfield:20}

2

Even for unique rid, where index entries are there for each array element, Multikey index size is small than simple index.

For example,
Both L and R has six rid values and index entry available for each six key.
Sample L Documents : {Lid:1, rDocs:[{rid:1},{rid:2},{rid:3}]}, {Lid:2, rDocs:[{rid:4},{rid:5},{rid:6}]}
Sample R Documents :{ rid:1, rfield: 10}, {rid:2, rfield:20}, {rid:3, rfield:20}, {rid:4, rfield:20}, {rid:5, rfield:20}, {rid:6, rfield:20}