CREATE VIEW dbo.vw_BatchItems
SELECT
    BatchID,
    FieldA,
    FieldB,
    CAST(ISNULL(ROW_NUMBER() OVER (PARTITION BY BatchID ORDER BY DateEntered), 0) AS INT) AS CheckOrder
FROM <myDatabase>.<mySchema>.<myTable>;

CREATE VIEW dbo.usp_sel_BatchItems (@BatchID INT)
AS

SELECT *
FROM <myDatabase>.<mySchema>.vw_BatchItems
WHERE BatchID = @BatchID;

SQL Server 2008R2 build 4266

USE AdventureWorks
go
CREATE VIEW test1 as
SELECT
    ProductID,
    CarrierTrackingNumber,
    UnitPrice,
    CAST(ISNULL(ROW_NUMBER() OVER (PARTITION BY ProductID ORDER BY ModifiedDate), 0) AS INT) AS rownum
FROM Sales.SalesOrderDetail

2 records, INDEX seek, KEY lookup AS expected
SELECT *
FROM test1
WHERE ProductID = 897 --707

3083 records, table scan AS expected
SELECT *
FROM test1
WHERE ProductID = 707


Why are you doing the CAST ISNULL thing?  Seems unnecessary to me, and this view avoids that "unnecessary" ComputeScalar in the plan:

CREATE VIEW test2 as
SELECT
    ProductID,
    CarrierTrackingNumber,
    UnitPrice,
    ROW_NUMBER() OVER (PARTITION BY ProductID ORDER BY ModifiedDate) AS rownum
FROM Sales.SalesOrderDetail

SELECT *
FROM <myDatabase>.<mySchema>.vw_BatchItems
WHERE BatchID = @BatchID
OPTION ( RECOMPILE );