Analyzed: Penn & Teller’s “Cups and Balls” magic trick
The new open, online journal PeerJ has as one of their inaugural articles ” Perceptual elements in Penn & Teller’s “Cups and Balls” magic trick ” in which one of the famous magic couple’s tricks get the scrutiny:
Magic illusions provide the perceptual and cognitive scientist with a toolbox of experimental manipulations and testable hypotheses about the building blocks of conscious experience. Here we studied several sleight-of-hand manipulations in the performance of the classic “Cups and Balls” magic trick (where balls appear and disappear inside upside-down opaque cups).
We examined a version inspired by the entertainment duo Penn & Teller, conducted with three opaque and subsequently with three transparent cups. Magician Teller used his right hand to load (i.e. introduce surreptitiously) a small ball inside each of two upside-down cups, one at a time, while using his left hand to remove a different ball from the upside-down bottom of the cup.
The sleight at the third cup involved one of six manipulations: (a) standard maneuver, (b) standard maneuver without a third ball, (c) ball placed on the table, (d) ball lifted, (e) ball dropped to the floor, and (f) ball stuck to the cup.
Seven subjects watched the videos of the performances while reporting, via button press, whenever balls were removed from the cups/table (button “1”) or placed inside the cups/on the table (button “2”). Subjects’ perception was more accurate with transparent than with opaque cups. Perceptual performance was worse for the conditions where the ball was placed on the table, or stuck to the cup, than for the standard maneuver. The condition in which the ball was lifted displaced the subjects’ gaze position the most, whereas the condition in which there was no ball caused the smallest gaze displacement.
Training improved the subjects’ perceptual performance. Occlusion of the magician’s face did not affect the subjects’ perception, suggesting that gaze misdirection does not play a strong role in the Cups and Balls illusion. Our results have implications for how to optimize the performance of this classic magic trick, and for the types of hand and object motion that maximize magic misdirection.
It’s a interesting case of where the magician’s own beliefs on why the trick works get challenged:
Magician Teller devised this variation while fiddling with an empty water glass and wadded-up paper napkins for balls, at a Midwestern diner […]. He turned the glass upside down and put a ball on top, then tilted the glass so that the ball fell into his other hand. The falling ball was so compelling that it even drew his own attention away from his other hand, which was deftly and automatically loading a second ball under the glass (he was so well practiced that he no longer needed to consciously control his hands).
In fact, Teller found that the sleight happened so quickly he himself did not realize he had loaded the transparent cup. Teller further realized that all of this took place despite the fact that he should have been able to see the secret ball as it was loaded under the cup. Its image was on his retina, but he nevertheless missed it because his attention was so enthralled with the falling ball.
He surmised that if it worked for him with a transparent cup, it would work with an audience. The transparency of the cups would make the trick all the more magical to the audience. Penn & Teller claim that their version of the trick violates four rules of magic: don’t tell the audience how the trick is done, don’t perform the same trick twice, don’t show the audience the secret preparation, and never perform cups and balls with clear plastic cups.
As a side note - this is one of the most interesting insights into humans and relates directly to the idea of “eyewitnesses” in deciding truth. Physically an image can be projected upon our retina but we miss it. Many tests have been done on this phenomenon over the years, and it is one of the important realizations to gain when studying human-ness.
In the discussions the authors point out that Teller’s belief, as to why the trick works like it does, fails when tested:
We investigated the potential contribution of several perceptual elements in Penn & Teller’s version of the classic “Cups and balls” magic trick. We measured the perceptual performance and gaze behavior of naive observers as Teller surreptitiously introduced balls inside opaque and transparent upside down plastic cups.
Contrary to the magician’s intuition, a gravity-driven drop of a ball into his hand (or to the floor) caused less misdirection, both in terms of gaze displacement and impaired perception, than alternative manipulations such as lifting the ball, or attempting to drop a ball that is stuck to the cup. Thus, perception of (the effects of gravity on) falling objects does not enhance magic misdirection, at least in the performance of this particular sleight-of-hand trick.
The contradiction between our results and the magician’s original perception may have been caused by one or more of several possible sources. One possibility is that performing the trick in a new way may have drawn his attention towards the new element (the ball dropping), and away from the common element (the loading of the cup). Successive, non-controlled repetitions of the procedure could have given the impression of a worse detection of the loading because of confirmation bias.
Our results confirm that controlled experiments give valuable insight to reject (Cui et al., 2011) or accept (Otero-Millan et al., 2011) intuitive judgments about attention and misdirection formulated by magicians. Further, the three consecutive sleight-of-hand manipulations (actual or simulated loads) were presented in isolation, rather than as part of a complete “Cups and balls” magic routine (an arrangement of tricks organized in logical fashion as part of a magic performance). Finally, because an actual magician (i.e. rather than a cartoon or computer simulation) performed all maneuvers, motion features such as timing, duration, etc. could not be exactly equated across all experimental conditions. Future research using computer simulations of the magician’s sleight-of-hand movements should be conducted with the goal of replicating and generalizing the current findings to other sleights-of-hand and magic tricks.
This reminds me of the old maxim that it is easiest to sell a salesman.
As described in the article, the viewers of the trick learned upon repeat viewing what was going on and were fooled less and less as they were trained. Another important lesson that I propose we can glean, to apply to the rest of our life as we deal with the avalanche of advertising and PR spiels-of-the-day.
The researchers included videos of the magician doing variations of the trick, that were used in some tests:
The “standard” routine:
The “No ball” routine:
The “Lift” routine, which is particularly effective:
And the “Drop” routine:
Which variant diverted your gaze the best?