Not overcommitting is Windows's default and only behavior

A memory allocator can implement overcommit, because you can separate reserving virtual memory and having it backed by physical memory into two different system calls. But from the point of view of the kernel, any time it promises to give you physical memory that memory is backed either by RAM or by space reserved in the swap file