This slide shows an intermediate cutoff value - somewhere in the middle between totally credulous and totally skeptical must lie a reasonable level of skepticism which maximizes your odds, so in this case the cutoff is represented by an intermediate size circle.
If you look at the selected therapies on the right side of the decision filter you will notice that some of them are bigger (less promising) than the cutoff. This is because you cannot know with certainty the exact effectiveness of any alternative therapy, you can only estimate it based on the evidence, and on how the plausibility of its rationale. Since your estimates will not be perfect, some therapies which are not as good as your minimum cutoff may slip by the filter as shown in this example. Still, even though the estimates are not perfect, this example filter has selected a reasonable set of therapies.
The ability to correctly estimate the promise of a therapy has, as you can imagine, a tremendous effect on how good your filter is, how good your decisions are. I call the ability to correctly estimate the promise of therapies the efficiency of your filter. The next three slides will explore the effect of varying efficiencies on the effectiveness of your decision filter.