Yes, if you're happy with HTTP protocol based driven load, you're correct, eValid is overkill.
However, if you engine is serving page, a single HTTP POST does not a web page retrieval make.
eValid, when doing a document-get action (e.g. with a FollowLink command) gets not only the base page but automatically collects all of the other needed pages that it refers to.
It depends on your page, but in many cases the size of the base page is less than 10% of the total data volume delivered over the wire for a single complete page. If so, then your measurements for 500 hits will be off by a whole order of magnitude...
The quantity to look for is the total data volume downloaded.