Yes, you are correct. Apparently the issue for me is the server used in Greensboro or something here or between here and there that causes some problems in getting even a semi-accurate reading. Using Speedtest and the recommended server city, I received results that were often much too high to be true and that ranged wildly upon repeats.
Now that I re-do the testing, choosing the Jacksonville or Atlanta server at Speedtest, I get more consistent results, which also match up well with what Speakeasy reports.
Speedtest gives me 6788 / 357 (Atlanta)
Speakeasy gives me 6697 / 371 (Atlanta)
My apologies for the blanket statement about Speedtest. I jumped the gun with that comment.
