I've got 50 documents (for trial purposes) more or less like this one:

```
"_id" : ObjectId("5461c8f0426f727f16010000"),
"f" : [
{
"t" : "ry",
"v" : 1972
},
{
"t" : "g",
"v" : [
"Crime",
"Drama"
]
},
{
"t" : "ml",
"v" : "English"
},
{
"t" : "k",
"v" : "movie"
},
{
"t" : "ai",
"v" : 972353
},
{
"t" : "ec",
"v" : 46
},
{
"t" : "rel",
"v" : true
}
]
}
```

I made the following index:

```
{
"f.t" : 1,
"f.v" : 1
}
```

So I make the following query:

```
db.items.find(
{
"$and": [
{"f": {$elemMatch: {"t": "ry", "v": {$gt: 1980, $lt: 2000}}}},
{"f": {$elemMatch: {"t": "g", "v": {$in: ["Drama"]}}}},
{"f": {$elemMatch: {"t": "k", "v": "movie"}}}
]
}
).explain()
```

Explain document:

```
{
"cursor" : "BtreeCursor f.t_1_f.v_1",
"isMultiKey" : true,
"n" : 6,
"nscannedObjects" : 50,
"nscanned" : 50,
"nscannedObjectsAllPlans" : 50,
"nscannedAllPlans" : 50,
"scanAndOrder" : false,
"indexOnly" : false,
"nYields" : 0,
"nChunkSkips" : 0,
"millis" : 0,
"indexBounds" : {
"f.t" : [
[
"k",
"k"
]
],
"f.v" : [
[
"movie",
"movie"
]
]
},
"server" : "b009115.mongolab.com:42700",
"filterSet" : false,
"stats" : {
"type" : "KEEP_MUTATIONS",
"works" : 51,
"yields" : 0,
"unyields" : 0,
"invalidates" : 0,
"advanced" : 6,
"needTime" : 44,
"needFetch" : 0,
"isEOF" : 1,
"children" : [
{
"type" : "FETCH",
"works" : 51,
"yields" : 0,
"unyields" : 0,
"invalidates" : 0,
"advanced" : 6,
"needTime" : 44,
"needFetch" : 0,
"isEOF" : 1,
"alreadyHasObj" : 0,
"forcedFetches" : 0,
"matchTested" : 6,
"children" : [
{
"type" : "IXSCAN",
"works" : 51,
"yields" : 0,
"unyields" : 0,
"invalidates" : 0,
"advanced" : 50,
"needTime" : 0,
"needFetch" : 0,
"isEOF" : 1,
"keyPattern" : "{ f.t: 1.0, f.v: 1.0 }",
"isMultiKey" : 1,
"boundsVerbose" : "field #0['f.t']: [\"k\", \"k\"], field #1['f.v']: [\"movie\", \"movie\"]",
"yieldMovedCursor" : 0,
"dupsTested" : 50,
"dupsDropped" : 0,
"seenInvalidated" : 0,
"matchTested" : 0,
"keysExamined" : 50,
"children" : []
}
]
}
]
}
}
```

As I see, it has scanned all documents (50) when only six are the matched documents. Why the scanned/matched document ratio is so poor (is it?). Is there a better way to improve it?

Thanks!