The problem
 The solution

 Contents
 Questions
 Anonymous hubs

The Problem

When you test the speed of an anonymous hub, GhostSurf forces the hub to download a test file twice, and averages the speed. Although this helps smooth out the performance numbers for the hub, giving you a more accurate view of its speed than a single download would provide, the test is still not as accurate as one would hope; sometimes, a slow hub will look good.

To combat this, the Scheduler uses a little trick that you too can use if you're setting up hubs for yourself. This trick is described below.

The Solution

The Scheduler uses a trick to improve its performance measurements for anonymous hubs. The trick is this: it retests hubs that worked the first time, and then deletes slow or broken hubs from this list, then tests again from this smaller list, then prunes the list again, and so on. Any hubs that are left after this process have survived multiple tests. And the repeats aren't very time consuming, because after the first round only the few hubs that worked the first time will be on the list.

To mimic this process yourself, open the anonymous hubs window, and click on the "Download hubs" button. This will fill the anonymous hubs list, if it is not already filled. Then, speed test the hubs by clicking on the "Calculate hub speed" button. Then, prune the list by clicking the "Delete broken hubs" button. Now you're ready to repeat; click the "Calculate hub speed" button again. This next speed test won't take too long, because you're working from a much shorter list. Some of the hubs that worked last time will be broken this time. These hubs are unreliable; it is exactly for this reason that the trick really helps improve your connection speed. Now click "Delete broken hubs" again to remove these, and repeat the two-step process until you're satisfied. The Scheduler repeats this twice; we don't recommend doing it more than three times, because you're unlikely to improve your results beyond that point.