the three-point test for credible prayers

During my wanderings through the WWW not too long ago, I made a piercing comment about prayer. Someone replied with the dare to just give prayer a chance. This entry is addressed to you, wherever you are.

On one hand, this isn't a feasible suggestion: is it sensible for me to attempt an action before I've heard an acceptable rationale for it? I suppose that I could try almost anything, such as pyramid power, but surely I may at least demand beforehand that the suggestion has a plausible explanation to back it up?

On the other hand, it's obviously too strict to require that a comprehensive theory must always precede experimentation. The precise details of the supernatural domain don't need to be coherent and complete in order to check a concrete procedure such as prayer. So, fellow mysterious WWW participant, let's do this.

Of course, for the test of a request for supernatural intervention to be credible, it needs to fulfill three easy points.

1. significance: The success condition must not be due to natural chance. Supernatural intervention cannot be mistaken for a commonplace event. Let's say that the typical probability of the success condition must be 5% or less.
2. falsification: As for the failure condition, it must be obvious and unambiguous and undeniable. If a failure of a test "isn't really a failure", then the test isn't really a test either. 
3. repetition: Ideally, the test must be repeated many times, and the count of all successes must be compared to be count of all failures. It's dishonest to disregard the result of any correctly-performed repetition of a test.

There have been statistical medical studies on the effect of prayer. Of the studies that fulfill these three points, using large-scale populations and clear interpretations, the final conclusion doesn't strongly support prayer's medical effectiveness. Certainly not in comparison to usual medical interventions such as surgery, therapy, pharmaceuticals, etc.

But it's difficult for you or me to carry out complicated studies like these. How about we settle on an alternative that's more personal and quick? While we're in a sober and alert state (no hallucinogens or hypnotic suggestions or group rituals!), each of us will pray for an audible supernatural message. The message (test pattern?) shall happen at normal conversational volume and be encoded in words. 

Note that this isn't a formidable request. Objective sound waves are nothing more than air vibrations, and presumably soulful supernatural entities are capable of some form of language. If the supernatural domain "refuses" to respond, the refusal is indistinguishable from a supernatural that doesn't exist at all. Therefore anything supernatural that refuses to respond is at minimum unconcerned with whether or not you or I think that it exists.

This test fulfills the three points. Based on experience, a message with no known natural source is generally unlikely (#1). An absence of a comfortably-loud stream of words is itself unambiguous (#2). The request for a message can be repeated and counted as many times as needed (#3).

...did it work for you? 

You may object that the prayer didn't work for me because I didn't want it to work, but if that was always the case then you shouldn't have made the original suggestion that I give prayer a chance. After all, if a prayer test only works for those who don't feel the need to test it, then it's not a good evangelism strategy...