Patch operation taking more RU's than replace. What gives?

Question

I have the following document in a cosmos DB (running emulator).

        {  
            "partitionKey": "3be5653d-c500-4a20-9f30-ce9df876c397",  
            "NumberOfRatings": 0,  
            "AverageRating": -1,  
            "Views": 2,  
            "Favorites": 0  
        }  

The above gets updated many times, and according to my reading patching a document should be more optimal than replacing an entire document. I have run some tests with the following code. The first block patches a document the second replaces it. They both increment the Views count in the above document by 1:

    //  Gets a user in the database   
    UserInfo userInfo = GetUser();  

    //  This operation costed 10,79 RU's  
    ItemResponse<UserInfo> response = await container.PatchItemAsync<UserInfo>(  
        id: userInfo.Id.ToString(),  
        partitionKey: new PartitionKey(userInfo.UserId.ToString()),  
        patchOperations: new[] {  
                            PatchOperation.Increment("/Views", 1)  
        }  
    );  
      
    //  This operation costed 10,67 RU's  
    response = await container.ReadItemAsync<UserInfo>(userInfo.Id.ToString(), new PartitionKey(userInfo.UserId.ToString()));  
    UserInfo itemBody = response.Resource;  
    itemBody.Views++;                          
    response = await container.ReplaceItemAsync<UserInfo>(itemBody, itemBody.Id.ToString(), new PartitionKey(itemBody.UserId.ToString()));  

The patching costed 10,79 RU's while the replacing costed 10,67 RU's. I would like to therefore ask the following two questions:

1. Is this normal?
2. If its normal, what the purpose of patching if it does not improve RU cost, rather makes it worse?

Answer 1

Partial updates main advantage is not saving R/U for transactions.
They will make update process easier. (You don't need send the whole document to Cosmos DB to update an item)
Some of the advantages are :
Also, it reduces network bandwidth since you don't send the whole document anymore.
lower end-toend latency
saves CPU cycles on the Azure Cosmos DB SDK client hosts.

Answer 2

I am glad this helped @Bubba Jones
For monitoring, I can reference the following article since it is a free way to do this.
https://stackify.com/w3wp-high-cpu-usage/