Executive Summary: Nope. Not even close. --- As an offshoot of another thread about entropy, I offered an idea that cards could be "shuffled", yet still come out with an order to them. (I put shuffled in quotes because we weren't talking about a random shuffle). As an example of it being plausible, I offered the idea that a magician (or was it a croupier?) claimed he could tell a card from its weight alone. The idea is that the more ink, the heavier the card, so Ace through 10 should have increasing weights, with face cards weighing the most. (It wasn't meant to be an assertion, merely a thought experiment, so magician/croupier's claim was just an example of plausibility). Several people challenged whether there was sufficient data to back this up. I decided to put my money where my mouth is, and do some good science on it. I went to my brother's lab - where he sciences the heck out of microbes - and availed myself of a microgram scale. Here are my results. Sorry about the tiny text and graphics. There's a lot of data. Also sorry that it's not a proper statistical analysis. Never studied Standard Deviations & such. --- First things first. Do I even have the equipment to test this? I test the accuracy (precision?) of the scale: Please Register or Log in to view the hidden image! Yup. This will do very nicely. --- Now to correlate the weight of the cards with their value: Please Register or Log in to view the hidden image! So that's pretty conclusive. --- I did some further tests while I was there, including seeing if it was correlated with the suits: Please Register or Log in to view the hidden image! (This is so weird, it makes me suspect my data got wonky.) So yeah, the magician/croupier was blowin' smoke out his butt. And yeah, my "plausible" idea did not stand up in the face of hard data. --- (I also tested a separate deck of cards, to see about variance across production. I didn't bother graphing it though, since the above results sort of render it all moot.) Happy to share the raw data, if anyone wants to do some fancy stuff with it.
technically... looking at it from a simple human science perspective the human mind can do the math to follow the numbers of the cards being shuffled to be constant probabilistic formula presenting known variables at any given time. quantum computers might be capable of this at some stage. it is a part of the human brain that is not understood you see it in savant autism as a clearly identifiable measurable pattern. like the person who could count the number of matches in a box between it being knocked off a table and the full box landing open spilling on the floor. what you are suggesting is something casinos would spend millions investing in to prevent. however, gamblers don't shuffle, so there is no need to prevent it for the ones taking 93 cents in every dollar that walks through their front door. thats 93% tax folks that's real earning power ! the human stats would run on probable weights to assign to values of probable ques of which end of the number set is most likely given the most detectable weight. it would be near to impossible to industrialize such a skill to make hundreds of millions off it, just like electric cars powered by the national grid. while it may save the world, the rich people don't want to be saved.
So, iow - you’re saying that the magician was likely lying and couldn’t tell the weight differential among the cards?
Actually no. I must confess: In retrospect, it simply never occurred to me that it wasn't true. It seemed impressive, but didn't seem outlandishly so.
Those graphs seem to indicate that you can sort playing cards by value/weight - just not in the value order presumed.
Good point. Although the final graph, showing that it's not even consistent across suits seems to blow that idea up. I'd need to do the same test on a different deck of the same brand - to see if it is consistent i.e. - I mean, if a 2 of spades is always the heaviest and a queen of spdes always the lightest, etc.
As you handle the cards you're also bound to leave finger prints, oils etc. on the surfaces. It would be interesting to see whether, after handling, they tend to get heavier.
Indeed, that was proposed by my brother, as he shoulder-surfed my experimentation. So he brought me some gloves. Also, I began holding the cards by their edge, so as to minimize transfer.
I'm kind of fascinated now, by the data, and am tempted to return with another deck of cards. Anyone got any ideas of some interesting things to test?
I'd suggest that the variation in the weights of the cards is probably due in large part to inaccuracies in the manufacturing process. I would assume that the cards are cut from larger sheets. The cuts would not necessarily result in cards of identical dimensions. Also, the cards are most likely coated with a plastic protective layer, whose thickness (and hence mass) would be variable across different cards. The hypothesis that the cards should be identical prior to having their faces printed onto them is probably not reflected in practice. I also wonder about the accuracy of the scales you're using, especially when there is a significant time gap between measurements. You're trying to measure differences in mass amounting to a few milligrams. Also, what's the environment of the scale like in terms of the air-conditioning and air circulation in the room? Is the scale itself fixed in place and secure between measurements? You say it's a microgram scale. Do you know its manufacturer-rated accuracy (which I assume would be in the instruction manual for an instrument like that)?
I addressed the plausibility of this earlier. I assumed that, given cards need to be used in mechanical shufflers in Vegas, their thickness should have low tolerance. This may be a bad assumption, at least for non-pro cards. I'm thinking about a follow up experiment with multiple decks of the same brand. The scale is professional grade. This is in a microbiology lab. It has a little door you close before taking measurements. I waited until the reading stabilized before recording the value. The scale displays down to 0.0001g but the last digit should be ignored (or at least used only for rounding). However, it did so well that I recorded it. If I had ignored the last digit, the the error margins would be ... exactly zero - and I'd have three perfectly overlapped curves. Here is all three passes in one graph: Please Register or Log in to view the hidden image! The difference between successive measurements of the same card (series 1,2,3) was on the order of 0.0005g - two orders of magnitude smaller than measurements of different cards - on the order of 0.05g. The ostensible "errors" were extremely consistent across an entire suit The deviation of series 2 always showed a mass less than series 1, and series 3 always less than series 2 This suggests strongly to me that the scale (and technique) is extremely accurate, it is reading a real phenomenon (likely humidity loss), it is far more accurate than needed to measure the effect I was studying
