The following point is of interest primarily to the OP and is orthogonal to the OP’s question.
I’d give p>.33 for success, and thus expect it to work on at least one of the three people I’m writing about.
You should maybe spend some time looking at the foundation of your rationality, as this statement rings some alarm bells.
Probability estimates should be numbers, not ranges, unless you’re doing something nonstandard. I can understand saying something like “I don’t want to commit to saying anything about the probability of event A beyond 0.25 < P(A) < 0.3 because I don’t trust my brain’s probability-assigning hardware/software”. But your range is really wide, and includes probability 1! I don’t think you believe that you are certain that you can convert people, so it looks like you are not clearly reporting your probability judgment.
I’m addressing this in your comment, which I’ve ignored in a lot of other comments, because it looks like you’re doing this towards a self-serving end. The conclusion you’re reaching for is that you’ll convert someone, so you claim a lower bound for your probability estimate that let’s you assert this. (Incidentally, a conversion probability of 33% gives you a (1 − 0.33)^3 = 30% probability of converting none of the three people.)
The following point is of interest primarily to the OP and is orthogonal to the OP’s question.
You should maybe spend some time looking at the foundation of your rationality, as this statement rings some alarm bells.
Probability estimates should be numbers, not ranges, unless you’re doing something nonstandard. I can understand saying something like “I don’t want to commit to saying anything about the probability of event A beyond 0.25 < P(A) < 0.3 because I don’t trust my brain’s probability-assigning hardware/software”. But your range is really wide, and includes probability 1! I don’t think you believe that you are certain that you can convert people, so it looks like you are not clearly reporting your probability judgment.
I’m addressing this in your comment, which I’ve ignored in a lot of other comments, because it looks like you’re doing this towards a self-serving end. The conclusion you’re reaching for is that you’ll convert someone, so you claim a lower bound for your probability estimate that let’s you assert this. (Incidentally, a conversion probability of 33% gives you a (1 − 0.33)^3 = 30% probability of converting none of the three people.)