These last two years, Penn and Teller have had the magician's dream job: Getting paid to watch other magicians and figure out how they perform their cleverest feats. The two seasons of Fool Us, a programme they produce with Jonathan Ross for the ITV network, has revealed a lot...from a magician's perspective, anyway...of how magicians think, and what it takes to fool one. Penn and Teller are surprisingly good at guessing the method of even the most elaborate effect, but it's perhaps even more surprising the sorts of things that have fooled them.
First, I should note that they are given certain advantages that the home viewer is not given. Namely, they get to watch every single move each performer makes, whereas the programme editors have a habit of cutting away from the performers exactly when they make their most critical moves. (On the other hand, they have, from time to time, showed the moves happening quite openly. I can make neither heads nor tails of their method for deciding what gets cut.) On the other hand, Penn and Teller get to watch it exactly once, while the home viewer can record it and watch it again and again until the method becomes apparent. Since the performer's goals are, first and foremost, to fool Penn and Teller and win a trip to Vegas, the difference this makes in which effects a magician should choose will be apparent.
The performer who best realized what would work and enacted it in exactly the best way is the one I will take as my case study, a Dorset-based professional magician and mentalist by the name of Graham Jolley. You can watch his performance (edited for TV) here.
End Clean
I'm going to start off with what I think is the number one factor in fooling Penn and Teller. It is a very simple concept, but one that is often difficult to achieve in the performance of magic. Namely, the effect should end clean. What I mean by this is that no gaffs or gimmicks should be in view of the audience at the moment the effect is over. Not every magician who has managed to fool Penn and Teller on this programme did this, but the one person I can think of who did not, Mathieu Bich, is the exception that makes the rule, and a following rule too: his gaff was so obviously a gaff, and was handled so openly and left out so plainly that Penn and Teller underestimated it. Moreover, he abused the rules somewhat, wherein Penn and Teller must guess correctly on the first try, by throwing in the most brazen red herring this show has seen. But more about red herrings later, right now, we're talking about ending clean.
In the clip of Graham Jolley, you will see that Penn and Teller ask to examine the deck the Graham used, and find nothing there to reveal the method. This is because Graham had the wisdom to select an effect which ended with absolutely nothing to hide. All he used was a deck of 54 ordinary Bicycle rider-back poker-sized playing cards, a favorite of magicians everywhere.
To further drive home this point, let's look at some other examples from the show:
John Archer: Penn and Teller spot his moves, but they have no idea what gaff he is using. At the end of the trick, he immediately tucks the incriminating evidence into his coat, making it clear that he refuses to produce it for inspection. The envelopes he hands out are clean. He successfully fools Penn and Teller.
Daniel Madison: Penn and Teller immediately examine the deck after he's done and find exactly what they expect to find: a duplicate royal flush in diamonds. Because this effect, so impossible looking to a lay audience, has so few possibilities as to the method, this confirmation is all Penn and Teller need to confirm their suspicions and shut him down.
Mark Shortland: He leaves the critical gaff right out on the table until the trick ends. Penn and Teller ask to see it, and easily work out the method. The funny thing is: he could have easily ended clean if he wanted. Just put the die box back under the table after the last die roll before the last envelope gets hammered. And keep a second ungimmicked die box under there to swap in if Penn asks to look at it. But hindsight is 20/20, and he didn't know how critical it would be to end clean. As a result, he won nothing. Well, I mean, other than the opportunity to perform live on national television and meet Penn and Teller, but really, what's that worth to a career?
Puzzles work better than Tricks
Every magician knows that an audience will tend to respond better to a genuine miracle, a feat that looks impossible, than something that is just a simple puzzle or illusion, what Graham rightfully calls a "minor miracle." And this is absolutely correct.
On the other hand, a magician is far more impressed by a minor miracle with a completely opaque method than ripping the heads off of birds when the method is obvious. This is something that all of the winners have in common: their methods have been either very cerebral, or at least two trains of thought removed from what the most obvious methods (to a magician) would be. The effects, on the other hand, look far more like puzzles than miracles.
Graham knew that he would be most likely to fool Penn and Teller with a self-working math trick that uses no gimmicks and no sleight of hand, because that was exactly what they wouldn't be expecting. He also knew that that wasn't going to entertain them very much, and he needed to have them on his side.
Win their Hearts
With perhaps the exception of Michael Vincent, every magician that has come on the show has done a single routine. Michael's strategy was to throw a bunch of techniques at Penn and Teller so that he might do something they hadn't yet seen, so he didn't know exactly which of his routines might fool them. Graham only the other hand did an entire routine /just for the sake of doing it/.
Yep, before he started the snooker ball routine, he said that he didn't expect it to fool them, and he repeated that again afterwards. So why did he do it? It's certainly not that he had to fill time. His two routines combined took as long as any of the longest routines on the show.
I think there are two reasons: One, it allows him to do something that actually astonishes the audience and establishes his claim that he is a mindreader. Second, the long list of jokes he injects into this and the way he gets the audience invested in him establish him as a quality entertainer. Together, these things accomplish his true purpose of getting Penn and Teller on his side before he has even started. They want good entertainers to open their show in Vegas, and so, he banks on them wanting, ever so slightly, to be fooled by his real effect.
It also does one more thing: it gives them the impression that the method to his second effect must be more elaborate than his first. It being that the first relies on a very intricate gimmick, the second would have to be very tricky indeed to fool them. And yet, it is actually less tricky in some sense. This is Graham's first red herring. (I want to point out that the method of the second effect, because it relies less on opaque gaffery, is much easier to work out just by watching it. However, I suspect if asked how the snooker ball trick was done, the best Penn and Teller could come up with is "the box records the order in which the balls were removed and you read that record when you close the lid." This might have been enough to keep him from winning the prize and so he didn't risk it, but it is clear this is hardly an explanation of the actual working of the gaff, since there are just so many different ways it could work, as Graham readily admits.)
As a way of backing up this argument, I would like to point out again what Penn said when John Archer (see link above) concluded his routine: "Wow! It's a funny routine! It's just silly! It's not really supposed to fool us!"
Did the fact that they genuinely believed it to be a funny routine get them on Archer's side enough to slightly bias them towards wanting to be fooled? As a third party observer, I couldn't say for sure. However, I also can't help but note that for no other act did they so readily admit the possibility of being fooled as soon as they were asked.
Red Herrings
Red herrings are something that magicians are supposed to be on the lookout for and aware of when deciphering another magician's act, but what is a red herring to a lay audience, and what is a red herring to a magician are two completely different things. As a result, if your goal is to fool a magician, you have to change up the sort of red herrings you throw out.
You see, the kind of red herrings that magicians use on their audiences are of the sort "I did it this way last time to make you think this is how it works, but I'm going to break the pattern or change the pattern now so you doubt that is actually what happened." Teller himself explains this much better himself in a lecture on the magic of consciousness.
A magician on the other hand is going to be burning your hands and looking for specific moments in your routine where they know something fishy has to happen. This knowledge is what led Mathieu Bich to include that box in his routine, a box that was entirely unnecessary to perform it. He knew that having a box would lead Penn and Teller down the primrose path of "he pulled the deck out after he got the card selected, maybe he has multiple decks?" Any magician not familiar with Bich's work would be misled the same way.
Graham did the same thing, but far more subtly. As already mentioned, he used the first routine to give the impression that the second routine used very sly and arcane trickery to hide the fact that it was a self-worker.
The next thing he did, which was cut from the show before it went to air, was proceed to pull out a deck of cards and false shuffle it a number of times, and then immediately and openly admit those shuffles were blind. Why? Besides showing off, he wanted Penn and Teller to know that he was a reasonably skilled card worker. As Teller is also a master of card handling, this would gain his respect especially and Win Their Hearts, but it also put them on a close watch for sleights. He wanted them to be so intent on looking for moves that they forgot to keep the bigger picture, the "awful lot of procedure" in mind. He did this also in two more ways: First, he sat them at opposite ends of the table so that maneuvering cards between them would give him the cover he might need to do the sleights he did not do. (This also reinforced the logic of the effect, whereupon spreading the deck, the joker closer to Penn was associated with him and similarly for Teller.) Secondly, he spent several seconds chatting while fiddling with the deck after they both had returned their cards to the deck. This second act warrants a bit more discussion.
You see, there are a number of deck manipulations and card controls that involve riffling and spinning the deck in the way that he did there, and he emphasized the possibility of those moves by openly indicating he is trying to misdirect Teller who, he says, is burning his hands. He knows that Penn and Teller only get one guess, and since his method requires no sleight of hand, he is doing everything he can to sell the possibility that he is using it. Immediately after he stops playing with the deck, he says "I have turned two cards face up in the deck--the two jokers." The most logically way to process this whole sequence of events is that all that (nonexistent) sleight of hand was used to somehow reverse the jokers and locate the selected cards.
Now, as soon as Penn and Teller are asked to explain the effect, the first thing they do is ask to examine the cards. I have already stressed the importance of ending clean, so you know that he really could care less if they looked at the cards. But there is a critical difference between fooling a magician and fooling an audience: you fool an audience eliminating all possibility of trickery, but you fool a magician who only gets one chance to guess your method by opening yourself up to every possible kind of trickery but the correct one. Thus, as soon as they ask to examine the deck, he consents and then says "I have, however, given you a clue..." Penn and Teller may or may not have heard this or paid attention to it, but Graham was fully aware of the fact that he was presenting a puzzle, not a trick, as he pretended to be nervous about them looking through his deck. Better that they believe there is something there that they are missing than that it was just a self-working trick with a regular deck of cards.
I conclude that the Graham did not win his spot opening for Penn and Teller in Las Vegas by presenting a spectacular, impossible bona fide miracle. He won it by taking a clean and clever puzzle and getting in their heads to present it in the most devious way possible. It was not his effect that fooled them, but his superior grasp of magician psychology.
EDIT (Jan. 27, 2012): Teller speaks about one more way to fool a magician: Let them underestimate you.