Before we begin scientific or philosophical analysis, we start out with intuitions and preconceived ideas.
However, we can be convinced that our prior beliefs were wrong and unjustified by using blind testing methods. Basically, we devise a test in which we separate the truth of our beliefs from how we feel about those beliefs. In other words, if we are interested in the truth, and eschew self-delusion, experimental controls are our friends.
Well, prayer and religious reflections are anti-controls. They are not just unscientific, they are anti-scientific. By focusing on particular outcomes, or outcomes that have particular meaning to us, we bias all our observations to a preconceived conclusion. We deliberately provide safe haven for any delusions we may have.
Suppose I pray for a good day. Any random event that occurs gets run through a filter. Is it neutral? If yes, then ignore the event. If the event is particularly good, then make a mental note. If the event is particularly bad, refocus the terms of our experiment and say that we were praying for something we shouldn't have prayed for. Or we say that God intended us to learn some other lesson. The result is that our experiences reinforce our superstitious beliefs, no matter what those beliefs happen to be. For example, a black cat crossing your path is said to be lucky in Britain, but said to be unlucky in the United States.
Superstition is a strange attractor indeed. And a very lucrative one for all the crackpot organizations out there that promote superstitions.
Of course, I cannot prove that one ought not delude oneself. I just feel strongly that one ought not do so. If you agree, then you ought to give up your superstitions, and quit pretending that prayers are answered, or that your magic spells are effective, or that your magic is better than placebo. At least not until you've performed some suitably scientific testing to verify your belief.