by Matthew Cobb
The first of these videos of cats playing the shell game popped up in my Tw*tter feed yesterday (forgotten from whom). I then went mooching on YouTube and found the others. Mine is a life well spent.
The ‘shell game’ (aka the Three Card Trick or Find the Lady) is a classic game used to by con men to fleece hapless observers. You see a group of men standing around on a street corner, shouting animatedly. One man is moving three shells about, or playing cards face-down. You have to guess which shell conceals a marble (or whatever), or where the Queen is. If you get it right, you win the bet. It’s very difficult to follow, but one man in the audience is able to work it out, and he wins money! So you decide to have a go. And fail dismally, losing your money. The guy who won was a stooge. I have seen this happening, although I’m wise enough (and no gambler) so I’ve never lost any money myself.
Anyway, these videos suggest that if you want to beat the system, take your cat down to the street corner. Each of these cats is pretty good at spotting where the marble is. These videos raise a number of questions.
First, how do they do it? Sound would seem the obvious cue (I bet the cats couldn’t do it with cards), although there may be subtle cues unconsciously emitted by the human – the ‘Clever Hans’ effect, named after a horse in Germany at the beginning of the last century who could apparently add numbers, but was in fact picking up tiny behavioural signals from his (honest) owner.
Second (and more interestingly), why would the cat want to bat the right shell? They don’t get any reward except for praise. It’s interesting that they all seem to be thinking about the problem, and then make their choice. Without wishing to spoil a nice story, I suspect they aren’t thinking about it at all, they’re listening. They know where the marble is (in most cases), having tracked it with their ears. They’re now listening to see if their potential prey is making a noise. If it’s quiet, they then biff the shell, just like they’d hit a mouse’s hiding place. Of course, you may have a different interpretation.
[Note from Jerry: if you can, try this with your cat and report the results below. It’s CITIZEN SCIENCE!]
Fascinating! The cats don’t seem to be paying that much attention but they still get it right, though of course we don’t know how many failures didn’t make it to You Tube. I will have to see if Chairman Meow can do it.
The second cat seems the most into it & seems to take it really seriously.
I do not know how they do it, but maybe they can hear the object rolling under the cup?
Actually, you can do the trick, too. Just watch the one that it was put under in the first place and ignore the others.
doh. And sub.
Yes, that is how I do it, but I do not think the cat would do that.
I think that is exactly what the cat would do.
Not sure how to determine which of us is right. We could ask the cats.
“Hey cat, explain if you are not just paying attention to the shell with object.”
To test this, one could presumably block the cat’s view for a portion of the shuffling. A cat detecting the ball by hearing (or smell, or by recognizing imperfections in the cups, or whatever) wouldn’t be disadvantaged by this; a cat keeping track would be.
Controlling for the “Clever Hans” effect could be done by double-blinding – have the human that performs the shuffling not themselves know which cup the ball is under. We can’t see much of the humans in any of these videos, so we don’t know how much they’re providing feedback.
I don’t think it’s sound. Some of the objects in the videos above — such as a crumpled-up chewing gum wrapper — aren’t going to make a significant sound.
Lots of opportunity for experimentation, here, with different cups, balls, and surfaces.
I’d also like to see Randi work the cups; he’ll know all the different misdirection techniques that work on humans and be in a good position to figure out which are more or less effective against cats.
b&
But, I do notice that, in the first video, when the cat has an object and type of cup material that will make a decided sound (the bell hitting the metal cups), it succeeds every time, while the cat in the second one, who has what looks like walnut sheels and a die, which won’t make such a distinguisable noise each time, has a slightly lower batting average. Also, that cat isn’t watching on some of its successful attempts.
So, maybe try these two types of object-and-cup materials, those that make distinctive sounds every time they’re moved, vs. those that don’t, and keep score? To see if that matters in the cats’ success rate.
They’re probably using multiple methods and enjoying our amazement, the tricksters.
Cats have good directional hearing. Seems like that could be a factor,
I am skeptical. My initial assessment is that the cat is just being a cat, batting stuff in front of it. The “correct” choice is explained by the file drawer effect.
Yeah, until we know how many errors never got put on Youtube, we don’t really have anything to explain.
That’s cute, but it would be interesting to see the cats against a con man or magician. That could narrow down the number of hypothesis.
How so?
Can you give an example of how that would help?
Trying not to give too much away,let’s just say the “tricks of the trade” could help determine if the cat is tracking the ball regardless of which cup it is under; or tracking the cup which initially covered the ball.
In case of a con man .. the ball isn’t under ANY of the cups! That’s how the trick works.
Nobody can win against a con-man. That’s kinda the point.
What might be interesting is the cat’s reaction to the ball being in the “wrong” place. Would it get annoyed?
I mean .. I can see how changing the ‘cup operator’ will eliminate something. (If the cat can’t pick the right cup anymore, then we’ll have to concentrate on the original operator).
However, introducing a con man only adds new variables. After all, a con man doesn’t play the same game! (When a con man is done shuffling the cups .. there’s NO ball under ANY of the cups .. that’s why they always win).
Thats not really true. The operator of the shell game usually transfers the pea (it’s traditionally done with walnut shells and a pea, for practical reasons related to the shape of a walnut shell) from one shell to another, but rarely removes it from the table all together, except in demonstrations for entertainment purposes. So yes a magician who uses the shell game in his act might end up with the pea missing entirely, but on the street that’s neither necessary nor practical. If someone does get frustrated and grab the shells, a magician wants them to see it’s not there at all and realize how effectively they were tricked. The con man needs them to believe it was there all along and they just lost track of it. If this backfires and the person has to be payed off, it’s better to take the loss than be exposed. And the crew can just follow the winner and get their money back later in some dark alley if the loss was too big.
Still, I think there would be great benefit to doing a little sleight of hand for these cats. For one thing, a true shell game would have much less of an aural element, because of the materials and because it is accomplished very differently than the cups and balls are. I do think that it would be possible in this way to determine whether the cat is following the shell with their eyes or following the pea using some other sensory input.
Interestingly, my few experiments with magic on animals have convinced me that people play a huge part in fooling themselves, and that if the audience doesn’t understand that they are about to be fooled it is much harder to direct their attention to the place where they think the sneaky stuff is going to happen, while you do it somewhere else. In other words, people are looking for the trick when they should be watching the pea, which actually makes them very easy to trick.
Source: I’m a professional magician pursuing a theater history degree with emphasis on the performance, representation, and demonstration of magical powers.
My first thought was that cats have evolved a finely honed ability to follow the motion of small things but then I thought ‘What good would that do for a predator of small animals?’
Heh!
I’m going with theory that are a lot of “failed” videos that never made it to youtube.
I do so wish I had a cat to experiment on though!
We’ll have to wait for the results of the citizen science.
I think the first one (2 times with 3 cups, 1 time with 4 cups) is the most impressive.
IF I can get my cat to put a paw on one of the cups after I’ve ‘shuffled’ the cups, then I think I can pull of the same feat with my cat.
Just video a little ‘shell game’ session every day. If I’m not mistaken, the odds to get the right cups in this scenario are 1 in 36.
So in a month or two I could expect to have a ‘winning’ sequence.
Alternatively, cats pursue small and highly-mobile prey who are, for obvious evolutionary advantage, pretty good at evasive manoeuvres that rely on distraction and confusion.
How would we test that hypothesis?
There are some interesting questions about cognition in it: this is an experiment worth doing.
Here’s an idea for an iPad app. Mice that run around on the screen and interact with a cat who’s playing the game.