Fools ignore complexity. Pragmatists suffer it. Some can avoid it. Geniuses remove it.

Alan Perlis

In Self-Working Coin Magic, Karl Fulves poses the following puzzle: Can you arrange six pennies on the table so that each coin is touching four others?

It’s not a particularly difficult puzzle. A mathematician might note “Ah, you’re describing an octahedral graph. So maybe we can model an octahedron in coins by flattening it!” They’ll pretty quickly come up with one of a couple of very simple solutions:

So, what was Fulves’ intended solution? Behold this abomination:

If you try to make this, you’ll find it’s impossible to get pennies to stand up like that without sliding down, which is likely why Fulves wrote “You may have to hold them in place to keep them propped up.” Is Fulves’ solution wrong? No. It does meet the conditions of the puzzle as laid out. But it’s…clearly not quite right either. No sensible person would come up with it. You almost have to wonder whether he included it as a joke.

Alan Perlis seems to think that the most obvious approach to a complex problem is a complex solution, and it takes a few more brain wrinkles than most have to achieve a simple one. And as someone who also knows a thing or two about software development, I can see how he landed on that perspective.

Simplicity does not precede complexity, but follows it.

Alan Perlis

But I would like to note for the record that no one but Fulves himself would come up with that awkward structure as a first solution to his coin puzzle. No, I fundamentally dispute Perlis’s opinion that the road to simplicity must pass through complexity. Inventiveness is not purely subtractive. It often takes misguided humans and their tendency to way overthink a problem to solve a problem in the most convoluted, inconvenient and unreliable way possible.

Don’t worry, this post is more than tangentially related to magic. Consider the field of pen-etrations. There have been dozens of methods to stick a pencil or pen through some kind of paper (bill or card) and then “heal” it since just the 1980s. Someone them require expensive gimmicks, some require cheap gimmicks, some require DIY gimmicks, and some are completely impromptu. Some kind be handled freely and done surrounded, others are quite sensitive to angles and lighting. Some methods let you clearly see a hole before it is healed, others simply imply it. Some allow everything to be immediately examined and others require something to be switched out. Some let anything be pierced while others only work with US currency. Some allow the pen to be moved through the paper as if it were merely water and others require it to come out the same way it went in. It’s clear magic creators have been doing everything they can to be able to achieve X with the effect for any X. But it’s clear what the perfect version of the effect would be if it were possible:

• Impromptu with borrowed items
• Angle-proof
• Moment of penetration is clearly seen, as is the hole
• Can be done in bright daylight
• Pen(cil) can be moved around in all three dimensions while it is inside the bill/card
• Objects can be immediately handed out and examined. No evidence of tampering can be found in the pen(cil) or pierced object.
• Hands are empty before and after–nothing to ditch.

It is impossible to achieve all of these things at once, and one must take care not to sacrifice too many of them—most importantly, not sacrificing simplicity—in order to achieve the ones you are aiming for. But Lloyd Barnes has elected to sacrifice simplicity (as well as angle-proofness, lighting-insensitivity, visibility of moment of penetration, and immediate examinability) just to achieve the full 3-dimensional movement:

I’m not trying to beat up on Lloyd. He’s a brilliant creator and has a lot of great work on pencil penetration plots. But I think maybe his desire to focus his creativity on DIY card gimmicks that anyone can put together has led, in this case, to something that abandons the KISS principle entirely. To be sure, the illusion this creates is very convincing—but only under certain lighting conditions from a certain angle. In short, it’s great for social media magic, but it’s impractical. The fatal flaw is that you have to introduce your own cards and then you can’t immediately hand them out at the end. People are going to want to look at the thing that just had a hole in it. That’s the most important piece for examinability in this effect. And it doesn’t really justify why there are two cards involved at any point.

In short, it’s not the thing that someone trying to solve this problem in a way they would use in the real world would come up with. It’s what someone who has to drive more viewers to their channel by posting a new tutorial every week would come up with because they don’t have more time to think about it. Under those conditions, of course there are going to be a few flops among the bangers. But for the rest of you, assuming you don’t find yourself under that kind of time pressure, there are guidelines you can follow when creating in order to avoid these situations. I’ll lay them out in a format that the Python programmers will find most familiar.

The Zen of Magic

• Open is better than covered.
• Visible is better than implied.
• Simple is better than complex.
• Complex is better than complicated.
• Angle-proof is better than angle-sensitive.
• Fail-proof is better than risky.
• Lighting insensitivity is better than controlled lighting.
• Pocket space counts.
• Deviations from the expected deserve an explanation.
• Although obvious implicit explanations beat actually explaining.
• Affordances requiring awkward handling should be revisited.
• Unless awkward handling would be expected.
• There should be one best way to achieve the desired goal given the tradeoffs you’re willing to make.
• Although which ideals to sacrifice may not be obvious unless you’re Canadian.
• Ungimmicked is better than gimmicked.
• Impromptu is better than carefully set up.
• Although a quick set-up on the fly may be just as good as impromptu.
• If the method is hard to explain or the gimmick is hard to build, it may be a bad idea.
• One Ahead is one honking great idea — let’s do more of those!
• ______

In case you’re wondering, yes, the one best way to achieve the visual effect Lloyd’s gaff is designed to achieve is something like John Cornelius’s Perfect Pen. Angle-proof, lighting-proof, the moment of penetration is completely open, the penetrated object can be borrowed, and everything is immediately examinable. Sure, you could ask for more, but it’s right at the point where the added complexity will result in a loss of reliability. If you start adding more, you’re probably overthinking it, and Alan Perlis will shake his head in disapproval.

Every single time I thought I found an extremely sophisticated, “genius” solution to a simple problem, I was mistaken.

GM Noel Studer

No post next week. See you the week after.