Excerpt:
Magician Rick Johnsson proposed the idea in 1970, stating that if a trick is too perfect – seemingly impossible, with no apparent flaws over the course of its conjuring – it mght paradoxically become less impressive and make it easier for the audience to guess how it was performed. To back his contention, he argued that, in an era dominated by science, audiences no longer attribute magic tricks to supernatural abilities, but rather to the magician's skills. Moreover, onlookers now actively seek to explain how a trick was done. If a trick is too grandiose, it might seem more staged and leave room for only one obvious explanation, thus reducing the audience's enjoyment.
"For instance, a trick seen on television that appears ‘too perfect'—such as a magician correctly guessing all five words that a spectator is thinking of—might be explained in terms of camera tricks, or stooging," a team of University of London psychologists wrote in a new paper published to PeerJ. "However, if the performer introduces an error and misses one of the words, the performance might be more likely to be interpreted as ‘true’ mind reading skills, leaving the audience therefore being more impressed as well."
Johnsson's "Too Perfect Theory" has been contentious since its publication. Magician Tommy Wonder called it ‘‘one of the worst concepts to appear in magic in a very long time’’, while John Carney constructively critiqued that magicians shouldn't introduce flaws into their tricks, but rather subtly hint at plausible – but wrong – ways their tricks are performed, misleading the audience.
Despite the longstanding debate, no attempt has been made to scientifically test the "Too Perfect Theory". So, in two different experiments, the University of London psychologists did just that.
Link to the scientific paper: https://peerj.com/articles/13449/