@@player-8740 because there's literally no reason to do it. It would take more time to get rid of the pattern while shuffling the next time you want to play, and you need to shuffle them anyway.
Haha yes, probably. But that's ok with me, the "cheating" part was mostly just for fun to see if I could, I never plan on actually using it. In fact, I was going to add a red LED that illuminated whenever it was running a cheat routine to make sure it's very obvious, but that didn't make it in the final prototype.
What ya do is have 2 decks worth, a play deck, and a shuffling deck. While one is getting played, the other is being shuffled, and then at the end of a hand, they get swapped.
@@billmight5662 It doesn't have to be on the table to distract or maybe noise could be reduced. Also maybe with a smooth shell casing on the outside and noise reduction it would become viable? Definitely a cool product though :)
Back in 6th form when ever we played cards we had 2 decks and what would happen is while the game is going on the dealer would shuffle the cards for the next go. We also had it where the dealer must make everyone a stack of cards in front of them and people would grab a stack to prevent the dealer from dodgy deals and the dealer would have to wait until the other stacks where collected before collecting theirs.
one of the most important things done with real gears is to use helical gears instead of spur gears. Spur gears are more efficient, but all auto transmissions and tons of machinery use helical gears, just to help reduce noise.
@@OooohReally Not for anything directly touching the cards (you wouldn't want to grease anything that touches the cards either). Just for the drive gears (not the toothed pulleys for the belts either)
True, but there actually are some deck configurations in which, even after a random cut, you will still always win. The stack depends on number of people playing and you can't decide exactly what you get, but still an interesting concept!
@@opfromthestart3645 go to jail for a few months... 3 max thats all it took me to learn every fun card "mechanic" "*MAGIC*" trick for only entertainment purposes
“I’ll see yah in a few more years” 😂😂 gosh I hope not. Amazing video Andrew! This channel is one of my favorites and I can’t wait to see what you do next!
Let me know if you need some eyedrops. And thanks, I have been keeping it up! My newer videos have way better production quality, though I will admit some of them were better ideas in my head than in practice XD
Nice that it can start from an unordered deck with card recognition; quite a project! In some competitions (bridge, for instance) all tables need to play the exact same hands. They use devices to create a series of identical decks. But building your own is of couse way more satisfying, and a nice project indeed!
Evil! Subbed. Two ideas. First for the feeder, have a look at how printers run their rollers. They almost invariably have three: pick, feed, and separation. On cheaper machines the sep roller is just a pad. Anyways ideally the pick roller (the one that sits on the stack of paper) rotates along with the feed roller, and the sep roller is either stationary or even turning slightly backwards. That's what prevents double feeds. The pick and feed rollers are usually on sprag clutches so they can freewheel. After those three, you'll have two driven pinch rollers that actually deliver the paper. Second...there should be an even more evil and harder to detect algo in the evil shuffler: make player x really lucky. Don't specify any one hand, just make sure that x always has a better hand than everyone else. Maybe they'd even occasionally lose a hand just so nobody gets suspicious.
If you enjoyed this project, please check out some of my other videos and consider subscribing. I'm releasing new project videos every month! ruclips.net/user/3DprintedLife
Bro, did you consider tagging your video under the section of 'horror' ..This is the stuff that fuels nightmares. These casino machines will flat out put people in the streets.
Love the video! But there is a cut by the dealer once it's done. There are even ways to set up the deck where cutting it anywhere will still result in a rigged result.
This is a great project and video! Glad I found this channel. You could maybe add a physical button on the device itself to do the regular random shuffle so that mode can be used without having to open up a browser (if that's the most common use I mean haha)
Hi, great vid. Just had an idea for the dispenser to make it more reliable. Where you have the pair of rollers spitting the cards out, the top one could briefly rotate in the opposite direction, so if there is a second card on top it gets put back in the deck while the bottom one continues but if there is just one card it will get sucked slightly in then spat back out again. Alternate idea, put the roller on a moving arm that goes on top of the deck so it only pushes the topmost card out at a time, if you springload the arm so the downwards force (and move the tray up a bit for each card dispensed) is minimal the top card wont have enough friction with the one under it to drag it along. Have the arm mechanized to move out of the way (idk with a servo or something) so when the shuffled cards are being inserted it can get out of the way.
Great suggestions! I actually considered something similar to that first concept for V2, but got sidetracked by other projects. But I may end up doing just this once I get back to it!
Make an algorithm that makes the interface more simple. Like: which player will win and then it just makes a random situation in which that player will win. Also, you could add a little slide thingy to the end which can rotate. This then automatically distributes the cards to the different players / to the middle. You could add so many cool features in the code.
Really neat ideas, I may try those in V2! I've also been interested in calculating certain ways of stacking the deck such that no matter where it is cut, you can still always (or with high probability) select who should win the hand. I know some combos exist but I would love to write my own algorithm to identify all the options for different numbers of players and who's dealing, etc!
I'm late to this party, but I think the mechanics can be simplified quite a bit. The movement of the dispenser could replace the movement of the slots module - The slots module doesn't have to move at all and the ejection bar for the slots module can unload all of the slots at once if the dispenser is at the very bottom. This would make the sorting mechanism quicker and would allow you to fit more slots in. If the dispenser was double sided you could have slots on both sides as well, allowing even more slots within a compact space.
Nah the party never ends. I eventually want to build a V2 but I have a few other large projects to finish up first. I appreciate all the feedback though, definitely some things to consider!
Given the ability to set the deck order, and read the deck order, you could actually shuffle in software, and then just build the deck to that shuffle. No need to run through excess shuffles. I'm guessing only 2 would be necessary.
You are right, this should be able to sort into any generic order of 52 cards in two cycles with the following algorithm: 1) Give every card in the arbitrary order an index of 1 to 52 2) Look at the first card to sort. Assuming 8 bins in the sorter, if the card index is 1,9,18, 25, 33 40, 49, 57 (ie start at 1 and count up in 8s) then put the card in bin 1. If the card index is in the set 2,10,19… it goes in bin 2, If the card index is in the set 3,11,20… it goes into bin 3. 3) Reassemble the deck, ensuring bin 1 is at the bottom, bin 2 is next etc 4) Check if the card is in index range 1-8, if so place in the first bin. If in range 9-18 place in second bin and so on up to 52. Do this with all 52 cards. 5) Reassemble starting with bin 1. This should give you a perfect arbitrary sort - because in the first phase you will have done a pre-sort which puts the cards in an order where the second sort will put it into a perfect order. Because bin 1 in the pre sort contains cards 1,9,10 etc, these will be the first cards sorted in the second sorter pass. In the second sorter pass you effectively have pre-sorted chunks of 8 cards, and the first 8 cards must contain index 1 for bin 1, and index 1 will go to the bottom of bin 1 pile etc. Hope I’ve explained that enough!
@@AshHartwell I watched this at like, 2AM and didn't want to actually figure out the algo. Thanks for pointing out that Its literally just Radix Sort, base {number of buckets}, in this case 8. And radix sort takes Log (n) sorts, so for 52 cards, you only need 2.
Great project. And the reason that most state gaming regulations required all cards to hand cut is to disable any chance that an ASM can be hacked and used to cheat.
What fun you had creating this! Here's something you might like to add.... a "Pack Sorted by suit" with choices of how you want the pack sorted... Ace-King (or reverse?) but so you can pick Spades/Diamonds/Clubs/Hearts or any combination of the suits?
Yeah a few people on reddit suggested that. This is actually a feature that commercial casino shufflers have. They share a similar card recognition system, though it is locked to be exclusively used for the deck re-ordering feature to prevent sketchy casinos from abusing it like I can do with mine :)
Well shot, well edited, well narrated, straight to the point with enough information it's interesting but not too much that it gets boring or confusing. Oh, also a great idea that was executed well Amazing work, instantly subscribed at the end of the video!
This absolutely is what shufflemaster patent shows. Great video and great work showing us. This is definetly what the casinos do. I miss the good old hand shufflers now they say most states banned hand shufflers. No its banned bc they want more $$
I want this. Actually what really got me looking was less poker, as its easy to shuffle for poker with some practice, but there are so many board and card games with even larger decks and cards that I want to sleeve, that decks become difficult to shuffle, and it would be great to have something with a bigger hopper that could shuffle up large decks while the rest of a game is being setup.
Nice recreation of what we did in the mid 70's (OCR and handler not cards but the concept is the same) Personally I believe the rotary storage is a better method, but that is my opinion When you see the shuffle master sort the cards the first time this should tell you they have full control
5:40 Making a "code explanation video for NERDS" video or something similar for projects may allow for moar youtube treadmill/algorhythm feeding, as well as helping others learn how to do the "behind the scenes" (granted may be other sources to learn obviously, but moar videos/content = moar win?) (also would require NLVE-ing time which is ...eh (at least for me) so may not be worth it in your opinion, just throwing this idea out there)
This is a very impressive machine that goes well beyond the scope of anything I desire from a shuffler. However I hope you can push me in the right direction, I'm playing a board game called Talisman which has several decks containing hundreds of cards, with the main "Adventure Deck" being over 600 cards with all the expansions and is a tedious pain to set up. I'm looking for a shuffler that can shuffle US Mini sized cards and won't destroy them either. I've seen some of the demos of the cheap mass produced shufflers and they look like they'd wreck these cards. Manual hand cranked versions seem more viable but they would still take too long and don't seem to provide a good shuffle judging from some videos I've seen. Apologies if this was impertinent to ask but I'm at my wits end trying to find something good that won't break the bank and your video came up in the results.
Cheap shufflers will definitely destroy your cards, and I highly doubt any quality off the shelf solution exists for mini cards like that. You could definitely do something similar to my design, but keep in mind two things. 1. This design is pretty dependent on how much force is applied to the card being fed. So if you have too much force, you will feed 2 cards. And not enough you may not feed any. And somewhat related number 2. this feeding mechanism will only work with low friction cards, like bicycle playing cards. If you want to build your own solution to this, you would almost definitely want to flip the feed mechanism, so you can much more reliably control the force applied via a spring loaded mechanism, and this allows for some better feeding geometries which can be a little more resistant to minor damage at the edge of cards. Be warned, if you do want to build a project like this yourself, it is going to be a huge time sink, especially if you want to make it reliable. Good luck!
that's a really cool shuffler. i have one of those cheap noisy ones. lol i barley used it. i was tying to make a card distributor or electric dealer 12 years ago and instead made a device that throws an entire deck around the room in a matter of a few seconds. it was too fun to fix.
Just FYI....most companies that create shufflers don't sell their products not because of hacking but rather their model of business is based on leasing their products to casinos. They would much rather keep getting a monthly paycheck from the casino in using their machines than a one-off payment (not enough casinos open up each week for them to profit off of). The lease has stipulations that the casino has to adhere to or lose possible millions in lawsuits, this allows the shuffle companies to control all aspects of their units. You don't get access because you are not a casino....and contracts are usually set in years and numbers (such as 5 machines at $25,000 per month for 1 year, as an example). There is no "hacking" because unloading data requires a password authentication and a hardware key, otherwise you are physically hacking which can easily be caught by casino managers and surveillance. The software has several modes based on the number of games the casino provides, allowing a machine to be swapped from table to table as needed.....and it also has a sorting and recall function. Despite what conspiracies are out there, no, the machines do not have a "set hand" function as not only would that violate commission rules, but also leaves a vulnerability for casino losses by the shuffle company (which leaves them liable on those losses because the code is theirs by lease). Yes, the machine can sort and recall, but that's just reading, not setting. Just a little FYI
How about if the dealer is required to cut the deck? It makes it 1000x more complicated.. The only way to assure player x wins the pot out of a 9 player Texas Hold'em table, (without the dealer who cuts the cards knowing) is to fix the deck so that 52 combinations all having the same player winning. No matter what point in the deck u start from. Since there are a millions of possible combos, u can easily write a script of code that will make the player u wish to win, win every hand or every other hand to not make it obvious. You just need to tell the computer exactly how many players will be dealt in, and it could be done seamlessly.
Lesson Learnt...Never ever play to the casino in which machine shuffle has been done.....always play on the table on which Manual shuffling has been done by Hand.....Am I right Sir ???...
Haha I wouldn't worry that much, at least not in legal casinos in developed countries. First off, most decks are split by hand after being machine shuffled. But there's also plenty of laws to prevent unfair rigging of games. Besides, why would a legit casino risk getting caught and shut down when they already have the house advantage in all games? Casinos make tons of money for a reason! But no need to worry about being cheated, all of your losses are just statistics :D
great job ! you are quite a inventor! I am thinking of making a simpler version - the one that will just shuffle cards randomly into 4 stacks; no need to recognize cards etc.
Amazing invention ...First time to your channel...You mastered 3D printing, engineering, and coding to create this machine...Bravo.... Imagine what the big casinos are investing in, to gain advantage over players... Do the Shufflemaster one or two models have software/cameras to detect suits in pokers games??? I've been playing for years and been witnessing some extremely suspect hands by dealers..
@@3DprintedLife Biggest difference between what a printer does and yours is the orientation. Printers don't try to take the bottom sheet of paper, they take the top one. The paper sits on top of a spring-loaded tensioner that holds the stack of paper at a constant pressure against the rollers. Also, their rollers are always a flat profile, not the round one you get from the o-rings. I suspect it gives move grip for less tension, so there's a lot more friction between roller and paper than the paper and the next sheet downwards.
Super Cool! Most of the issues you were having with the cards, the biggest one being able to get a single card, I would have to say were the cards. Bicycle cards are great cards, but if you were to use them at a NLTH tournament, they would last about 10 hands max. If you had some plastic poker cards, the Copag Poker Jumbo are great around $10/pack, the cards would become a constant, not a variable. They would always be, and only be, as slick as when you begin, unlike the Bicycle which would start one way and continuously change with use. I really want one, but dont think Im at this level yet.
OCD can be emulated by comparing all master images with the scanned one and take the one with the highest % score. It's not how OCD works, but it's a lot faster.
i had a friend who had a cheapo shuffler. you split the deck in 2 parts and it would shove the cards randomly from each pile in the center, do it a few times and it would be shuffled
Due to strong interest in this project (and wanting to recreate it) I did some research calculations. I found that with 8 bins the chance of the bottom card ending on top is about 790x larger than the top card ending on top. This quickly decreases to 2.3 times after 2 shuffles and to 1.1 (neglectable) after 3 shuffles. I thought these results might be interesting ; ) I hope it's OK to use this concept for inspiration for a future design project of mine.
Bro , this person built this.. now all the casinos have “automatic shuffling shoes” casinos probably spent millions to develop AI technology to use against us
Every engineering channel on RUclips: I found a reasonable version online but “didn’t want to spend the money”, so I spent 10-100x (Stuff made here) the cost using my own time to create a super specialized, cool version that has little to no practical application. Or to put it another way: it’s fun to make shit. I suffer from this too. Getting ready to launch a Kickstarter myself.
the dispenser would colloquially be called a " shoe " in theory , you could have a realistic 1 , 2 , 3 deck-shoe game of blackjack - which was invented to stop card counting , the more decks in the shoe, the harder they are to count
It looks like the image processing is taking way longer than your mechanics can handle. (@11:34 you can see the shuffler just waiting.) It's worth looking into some form of faster computing.
Hey I am a senior in highschool and I’m making my own version of this for my engineering project and was wondering if you have a wiring diagram for it and what voltage you used as your power source for the micro motor on the card shooter. Great project and I’m having a lot of fun building it!
Lol so your first idea is basically what I wanted to do to shuffle MTG commander decks Almost identical (except I was planning on using another wheel to push them off to the sides) Edit: though i think the pusher is probably better and faster, might be harder to split the deck evenly in 2 though
Great design and explanation. I'm wondering if you could estimate build cost (assuming access to a 3d printer, say), and say anything about "maintenance" of the machine given regular use (eg do you need to recalibrate the camera, launcher etc regularly)? Sorry if these questions were answered in the video and I missed them.
Total cost is somewhere around $200-300 and there's no regular maintenance, but the card dispensing mechanism is unfortunately pretty unreliable so most of the work to keep this running involved adjusting the gaps on that part. It also would require more uniform lighting and opaque side panels to get the most reliable card detection due to the way the computer vision code was written, though I now know of some ways to improve this significantly if I ever complete V2 (which I started but is a big project!)
I know you said you had no interest in making these, but let me try and convince you otherwise. Machines like this, with the same capabilities run between $12-15k new, $4-8k refurbished. Let’s say at scale you could keep your cost to $200-300 . I would be willing to pay $2.5-3k for a shuffler like this. All you need is a little refinement, a case and a UX .... could be just buttons or a whole separate small LCD like the Shufflemaster Deckmate II. I’m telling you there is a huge market for something like this for weekly poker nights.
In some sketchier places this may be true, but most of the time they wouldn't bother. It isn't worth the risk for them to get caught doing this when all the games are already designed to favor the house. The shufflers do have this feature but it is strictly used for verifying the deck and re-stacking in the unshuffled order.
@@3DprintedLife I’ve gambled over 500k. I can assure you that all the casinos are cheating with these machines. The games don’t play anything like they used to. Ask any old timers how much easy it used to be 10 years ago when most tables were hand shuffle and you’ll see that the games are not what they used to be. The games are much harder to win with machine shufflers. Go watch a casino shuffler at a poker table and then go watch a hand shuffle game and you’ll see they are very different. This is why high roller tables are mostly hand shuffle because most people with a lot of money aren’t stupid. 10 years ago if you were playing baccarat on a hand shuffle you would find a 10-25 run bank or player. Now it’s so rare to find a run with the shufflers it goes more back and forth. The casino loses big money on long runs and the machine took those runs away. I can go to any casino with a shuffler and prove to anyone how they are cheating in real time. Especially Indian Casinos like Foxwoods and Mohegan Sun.
Poker hands would be harder to rig, because the dealer must cut the cards after retrieving it from the machine. So the dealer would have to be in on it and either do a false or precise cut to get the desired order of cards. Or the programmer rigging the hand would have to guesstimate where they think the dealer will cut cards.
Yes you’re very right and that’s exactly why cutting the deck is a thing. There are some theoretical hands where you can guarantee or almost guarantee a certain player wins no matter the cut but there are few combinations that can do this. And they of course only work with a specific number of players
@@3DprintedLife If it's just two players it is possible to stack the deck. So no matter what cut is made a certain player will win the hand, but no idea if you can do that with 8-9 players.
@@EliteCannagivers For casinos I could agree as they seem to take a fixed number of cards to the side. But if you cut the deck randomly, I don't see how you could keep it fixed.
@@kilfoofan it doesn’t fix an individual player. The deck only gets cut in half. It doesn’t matter how many times a person cuts a deck. The cards are still in order of the machine. That order controls when the bad beat jackpot gets hit. That machine sorts cards to make a lot more people get hands which creates a crazy table. More people go on tilt at a crazy table. Then those players start playing other table games because of emotions. Like I said cutting the deck doesn’t change anything. All it changes is who gets dealt the fixed hands at the table. The tables are all fixed.
@@EliteCannagivers I still have to disagree. In order it to be truly a fixed deck the hands would have to be predetermined and known, i.e. you'll know the outcome of the table even before you deal a single card. However, if you cut the deck randomly, it can't be known as the first card dealt is not yet known. You could know what everyone has, and what will be the community cards (if hold'em or omaha), based on your hand (if you can remember the exact order of the 52 cards for the individual shuffle, that is), but you could not predetermine the dealt cards as shown on the video, e.g. player 1 gets full house, player 2 gets 4 of a kind, player 3 gets straight flush. For this you can't have any randomness to the shuffle, including cutting the deck. It's like if you would take this post, cut it randomly, and then try to read it with the original word lengths from the beginning, but starting from the cut position. The outcome would be gibberish.
@@kilfoofan all you have to do is read the patent on Shuffle Master. It says it all in the patent. It also states in the patent that it can all be done wirelessly. People want to talk so much but no one wants to put in any research to see the real truth. Shuffle Masters don’t shuffle cards they sort cards. It’s just a fact stated in the patent as well. Maybe instead of replying you’ll use that time to read the patent.
would be cool to cut these components from acrylic using a laser since most parts are very boxy and at right angles. awesome projects and r&d! integrating the different components (mechanics, steppers, opencv, codebase) is very challening starting from nothing. how does it handle different sized cards? If I wanted to build this id want to shuffle any deck of cards (not necessarily poker cards).
Yeah the rear plate is laser cut and definitely could convert more components to use this, but I wanted to keep the design as open-sourced as possible so people without access to a laser cutter could still build it. I haven't tried any size other than standard poker cards but it probably wouldn't work great with others. And honestly, this was a fun project but not nearly quick or reliable enough to justify using over traditional hand-shuffling. I do hope to get a V2 made sometime soon, but I keep coming up with other more interesting projects which will get worked on first!
This can also be used to put the cards back into the box in the correct order when you are done playing!
Why would you want to do that though?
@@Fadexpl why would you not?
@@player-8740 because there's literally no reason to do it. It would take more time to get rid of the pattern while shuffling the next time you want to play, and you need to shuffle them anyway.
@@Fadexpl I like my cards at least suited at the beginning for a quick check if all are there.
@@therealpanse you can just count them
Your RUclips channel is very underrated. Glad I stumbled upon it today!
Thank you!
Agreed!
i was gonna write the same lol
Same
Same feeling
Great video. Now really convinced me that casinos is stealing my money right my very eyes with the "shuffles".
Not sure if all but indeed there are models that have an optic sensor.
Exactly! Rigged games...
@@mattiasw.5846 When do you cut the deck? ruclips.net/video/oZn7ZML4nmY/видео.html
It might raise some suspicions when it takes 5 minutes to "shuffle" the deck. Really cool project though.
Haha yes, probably. But that's ok with me, the "cheating" part was mostly just for fun to see if I could, I never plan on actually using it. In fact, I was going to add a red LED that illuminated whenever it was running a cheat routine to make sure it's very obvious, but that didn't make it in the final prototype.
What ya do is have 2 decks worth, a play deck, and a shuffling deck. While one is getting played, the other is being shuffled, and then at the end of a hand, they get swapped.
@@MinorLG I would get too distracted during the game though.
@@billmight5662 It doesn't have to be on the table to distract or maybe noise could be reduced. Also maybe with a smooth shell casing on the outside and noise reduction it would become viable? Definitely a cool product though :)
Back in 6th form when ever we played cards we had 2 decks and what would happen is while the game is going on the dealer would shuffle the cards for the next go. We also had it where the dealer must make everyone a stack of cards in front of them and people would grab a stack to prevent the dealer from dodgy deals and the dealer would have to wait until the other stacks where collected before collecting theirs.
The amount of system integration work that you're able to make work reliably is amazing. Amazing work.
Thanks! I'm actually a system integration engineer professionally, so I really would have no excuse if I failed haha!
it's so nice when the youtube algorithm coughs up something I'm interested in watching. Apparently herring bone gears can be quieter.
Silicone lubricant man, should help a fair bit with the noise in the gears.
I would say white lithium grease. They aren’t sealed so oil would quickly run off.
I recommend vaseline,
I use it on my wife too, stops her from screechin
one of the most important things done with real gears is to use helical gears instead of spur gears. Spur gears are more efficient, but all auto transmissions and tons of machinery use helical gears, just to help reduce noise.
@@Timestamp_Guy those aren't smooth it might cause internal damage
@@OooohReally Not for anything directly touching the cards (you wouldn't want to grease anything that touches the cards either). Just for the drive gears (not the toothed pulleys for the belts either)
Add a "sort" mode as well, for people that wanna get an organized deck on demand. :)
And of course, there is in fact a method to prevent just this kind of deck ordering shenanigans:
The Cut.
True, but there actually are some deck configurations in which, even after a random cut, you will still always win. The stack depends on number of people playing and you can't decide exactly what you get, but still an interesting concept!
@@3DprintedLife Is there a specific example of this?
@@opfromthestart3645 go to jail for a few months... 3 max
thats all it took me to learn every fun card "mechanic" "*MAGIC*" trick for only entertainment purposes
“I’ll see yah in a few more years” 😂😂 gosh I hope not. Amazing video Andrew! This channel is one of my favorites and I can’t wait to see what you do next!
Maybe only 6-12 months...we'll see! But first I need to come up with my next project idea...
Dude you need to keep it up. This was amazing. I just watched the whole thing without blinking.
Let me know if you need some eyedrops. And thanks, I have been keeping it up! My newer videos have way better production quality, though I will admit some of them were better ideas in my head than in practice XD
Welcome back! You should update the software to inform you if any cards are missing.
Yeah great idea! I really need to speed up the card recognition first though, it's too slow to run during a regular shuffle as is
Or duplicates
Nice that it can start from an unordered deck with card recognition; quite a project!
In some competitions (bridge, for instance) all tables need to play the exact same hands. They use devices to create a series of identical decks.
But building your own is of couse way more satisfying, and a nice project indeed!
Evil! Subbed.
Two ideas. First for the feeder, have a look at how printers run their rollers. They almost invariably have three: pick, feed, and separation. On cheaper machines the sep roller is just a pad. Anyways ideally the pick roller (the one that sits on the stack of paper) rotates along with the feed roller, and the sep roller is either stationary or even turning slightly backwards. That's what prevents double feeds. The pick and feed rollers are usually on sprag clutches so they can freewheel. After those three, you'll have two driven pinch rollers that actually deliver the paper.
Second...there should be an even more evil and harder to detect algo in the evil shuffler: make player x really lucky. Don't specify any one hand, just make sure that x always has a better hand than everyone else. Maybe they'd even occasionally lose a hand just so nobody gets suspicious.
If you enjoyed this project, please check out some of my other videos and consider subscribing. I'm releasing new project videos every month! ruclips.net/user/3DprintedLife
Bro, did you consider tagging your video under the section of 'horror' ..This is the stuff that fuels nightmares. These casino machines will flat out put people in the streets.
Love the video! But there is a cut by the dealer once it's done. There are even ways to set up the deck where cutting it anywhere will still result in a rigged result.
This could be modified for "tele-IRL card playing" i bet (pun?).
Kind of like what @Stuff Made Here did for the automatic pool stick?
Haha yeah that's a neat idea!
This is a great project and video! Glad I found this channel. You could maybe add a physical button on the device itself to do the regular random shuffle so that mode can be used without having to open up a browser (if that's the most common use I mean haha)
Plan for V2 is to add a touchscreen UI!
Sidenote, that paper feeder will be of use with some OS Plotter/inkjet printer development !
You spend two years working on this to finally beat your friends at poker. Then the dealer takes the cards from the machine and cuts them.
_sulks away slowly_ guess I need to make some improvements for V2...
Hi, great vid. Just had an idea for the dispenser to make it more reliable. Where you have the pair of rollers spitting the cards out, the top one could briefly rotate in the opposite direction, so if there is a second card on top it gets put back in the deck while the bottom one continues but if there is just one card it will get sucked slightly in then spat back out again. Alternate idea, put the roller on a moving arm that goes on top of the deck so it only pushes the topmost card out at a time, if you springload the arm so the downwards force (and move the tray up a bit for each card dispensed) is minimal the top card wont have enough friction with the one under it to drag it along. Have the arm mechanized to move out of the way (idk with a servo or something) so when the shuffled cards are being inserted it can get out of the way.
Great suggestions! I actually considered something similar to that first concept for V2, but got sidetracked by other projects. But I may end up doing just this once I get back to it!
Make an algorithm that makes the interface more simple. Like: which player will win and then it just makes a random situation in which that player will win. Also, you could add a little slide thingy to the end which can rotate. This then automatically distributes the cards to the different players / to the middle.
You could add so many cool features in the code.
Really neat ideas, I may try those in V2! I've also been interested in calculating certain ways of stacking the deck such that no matter where it is cut, you can still always (or with high probability) select who should win the hand. I know some combos exist but I would love to write my own algorithm to identify all the options for different numbers of players and who's dealing, etc!
Just discovered this channel now through the video. Impressive. Liked and subscribed.
Thanks!
I'm late to this party, but I think the mechanics can be simplified quite a bit. The movement of the dispenser could replace the movement of the slots module - The slots module doesn't have to move at all and the ejection bar for the slots module can unload all of the slots at once if the dispenser is at the very bottom. This would make the sorting mechanism quicker and would allow you to fit more slots in.
If the dispenser was double sided you could have slots on both sides as well, allowing even more slots within a compact space.
Nah the party never ends. I eventually want to build a V2 but I have a few other large projects to finish up first. I appreciate all the feedback though, definitely some things to consider!
Given the ability to set the deck order, and read the deck order, you could actually shuffle in software, and then just build the deck to that shuffle. No need to run through excess shuffles. I'm guessing only 2 would be necessary.
You are right, this should be able to sort into any generic order of 52 cards in two cycles with the following algorithm:
1) Give every card in the arbitrary order an index of 1 to 52
2) Look at the first card to sort. Assuming 8 bins in the sorter, if the card index is 1,9,18, 25, 33 40, 49, 57 (ie start at 1 and count up in 8s) then put the card in bin 1. If the card index is in the set 2,10,19… it goes in bin 2, If the card index is in the set 3,11,20… it goes into bin 3.
3) Reassemble the deck, ensuring bin 1 is at the bottom, bin 2 is next etc
4) Check if the card is in index range 1-8, if so place in the first bin. If in range 9-18 place in second bin and so on up to 52. Do this with all 52 cards.
5) Reassemble starting with bin 1.
This should give you a perfect arbitrary sort - because in the first phase you will have done a pre-sort which puts the cards in an order where the second sort will put it into a perfect order. Because bin 1 in the pre sort contains cards 1,9,10 etc, these will be the first cards sorted in the second sorter pass. In the second sorter pass you effectively have pre-sorted chunks of 8 cards, and the first 8 cards must contain index 1 for bin 1, and index 1 will go to the bottom of bin 1 pile etc.
Hope I’ve explained that enough!
@@AshHartwell I watched this at like, 2AM and didn't want to actually figure out the algo. Thanks for pointing out that Its literally just Radix Sort, base {number of buckets}, in this case 8. And radix sort takes Log (n) sorts, so for 52 cards, you only need 2.
Great project. And the reason that most state gaming regulations required all cards to hand cut is to disable any chance that an ASM can be hacked and used to cheat.
What fun you had creating this!
Here's something you might like to add....
a "Pack Sorted by suit" with choices of how you want the pack sorted... Ace-King (or reverse?) but so you can pick Spades/Diamonds/Clubs/Hearts or any combination of the suits?
Yeah a few people on reddit suggested that. This is actually a feature that commercial casino shufflers have. They share a similar card recognition system, though it is locked to be exclusively used for the deck re-ordering feature to prevent sketchy casinos from abusing it like I can do with mine :)
Me: I wonder if I could make a card shuffler that stacks the deck...
RUclips: Here you go
Well shot, well edited, well narrated, straight to the point with enough information it's interesting but not too much that it gets boring or confusing. Oh, also a great idea that was executed well
Amazing work, instantly subscribed at the end of the video!
Thank you!
This absolutely is what shufflemaster patent shows. Great video and great work showing us. This is definetly what the casinos do. I miss the good old hand shufflers now they say most states banned hand shufflers. No its banned bc they want more $$
I want this. Actually what really got me looking was less poker, as its easy to shuffle for poker with some practice, but there are so many board and card games with even larger decks and cards that I want to sleeve, that decks become difficult to shuffle, and it would be great to have something with a bigger hopper that could shuffle up large decks while the rest of a game is being setup.
I've stopped playing cards....
This is pretty fantastic. It would be fun for, say, having two identically shuffled decks and see who does better at Klondike.
Nice recreation of what we did in the mid 70's (OCR and handler not cards but the concept is the same)
Personally I believe the rotary storage is a better method, but that is my opinion
When you see the shuffle master sort the cards the first time this should tell you they have full control
Nice to see you back. Hope you are doing well.
5:40 Making a "code explanation video for NERDS" video or something similar for projects may allow for moar youtube treadmill/algorhythm feeding, as well as helping others learn how to do the "behind the scenes"
(granted may be other sources to learn obviously, but moar videos/content = moar win?)
(also would require NLVE-ing time which is ...eh (at least for me) so may not be worth it in your opinion, just throwing this idea out there)
How about the shuffle machine Texas Holdem when playing against other players?
This is a very impressive machine that goes well beyond the scope of anything I desire from a shuffler.
However I hope you can push me in the right direction, I'm playing a board game called Talisman which has several decks containing hundreds of cards, with the main "Adventure Deck" being over 600 cards with all the expansions and is a tedious pain to set up.
I'm looking for a shuffler that can shuffle US Mini sized cards and won't destroy them either. I've seen some of the demos of the cheap mass produced shufflers and they look like they'd wreck these cards. Manual hand cranked versions seem more viable but they would still take too long and don't seem to provide a good shuffle judging from some videos I've seen.
Apologies if this was impertinent to ask but I'm at my wits end trying to find something good that won't break the bank and your video came up in the results.
Cheap shufflers will definitely destroy your cards, and I highly doubt any quality off the shelf solution exists for mini cards like that. You could definitely do something similar to my design, but keep in mind two things. 1. This design is pretty dependent on how much force is applied to the card being fed. So if you have too much force, you will feed 2 cards. And not enough you may not feed any. And somewhat related number 2. this feeding mechanism will only work with low friction cards, like bicycle playing cards.
If you want to build your own solution to this, you would almost definitely want to flip the feed mechanism, so you can much more reliably control the force applied via a spring loaded mechanism, and this allows for some better feeding geometries which can be a little more resistant to minor damage at the edge of cards.
Be warned, if you do want to build a project like this yourself, it is going to be a huge time sink, especially if you want to make it reliable. Good luck!
You know, when I was a kid, I made a card shuffler (which worked quite well) out of legos. the wheels are perfect for dispensing cards
Love the content Andrew! I vote you make a video showing off your process for building super simple 1-pager GUIs for python with HTML and PMP.
Why thank you David!
that's a really cool shuffler.
i have one of those cheap noisy ones. lol i barley used it.
i was tying to make a card distributor or electric dealer 12 years ago and instead made a device that throws an entire deck around the room in a matter of a few seconds.
it was too fun to fix.
Just FYI....most companies that create shufflers don't sell their products not because of hacking but rather their model of business is based on leasing their products to casinos. They would much rather keep getting a monthly paycheck from the casino in using their machines than a one-off payment (not enough casinos open up each week for them to profit off of). The lease has stipulations that the casino has to adhere to or lose possible millions in lawsuits, this allows the shuffle companies to control all aspects of their units.
You don't get access because you are not a casino....and contracts are usually set in years and numbers (such as 5 machines at $25,000 per month for 1 year, as an example). There is no "hacking" because unloading data requires a password authentication and a hardware key, otherwise you are physically hacking which can easily be caught by casino managers and surveillance. The software has several modes based on the number of games the casino provides, allowing a machine to be swapped from table to table as needed.....and it also has a sorting and recall function. Despite what conspiracies are out there, no, the machines do not have a "set hand" function as not only would that violate commission rules, but also leaves a vulnerability for casino losses by the shuffle company (which leaves them liable on those losses because the code is theirs by lease). Yes, the machine can sort and recall, but that's just reading, not setting.
Just a little FYI
How about if the dealer is required to cut the deck? It makes it 1000x more complicated..
The only way to assure player x wins the pot out of a 9 player Texas Hold'em table, (without the dealer who cuts the cards knowing) is to fix the deck so that 52 combinations all having the same player winning. No matter what point in the deck u start from.
Since there are a millions of possible combos, u can easily write a script of code that will make the player u wish to win, win every hand or every other hand to not make it obvious.
You just need to tell the computer exactly how many players will be dealt in, and it could be done seamlessly.
underrated channel !
Appreciate that! :D
Good to see you back. This is honestly a pretty cool and involved project.
Thanks!
Lesson Learnt...Never ever play to the casino in which machine shuffle has been done.....always play on the table on which Manual shuffling has been done by Hand.....Am I right Sir ???...
Haha I wouldn't worry that much, at least not in legal casinos in developed countries. First off, most decks are split by hand after being machine shuffled. But there's also plenty of laws to prevent unfair rigging of games. Besides, why would a legit casino risk getting caught and shut down when they already have the house advantage in all games? Casinos make tons of money for a reason! But no need to worry about being cheated, all of your losses are just statistics :D
Brilliant work! Congrats on nailing it!
great job ! you are quite a inventor! I am thinking of making a simpler version - the one that will just shuffle cards randomly into 4 stacks; no need to recognize cards etc.
Amazing invention ...First time to your channel...You mastered 3D printing, engineering, and coding to create this machine...Bravo.... Imagine what the big casinos are investing in, to gain advantage over players... Do the Shufflemaster one or two models have software/cameras to detect suits in pokers games??? I've been playing for years and been witnessing some extremely suspect hands by dealers..
This: is in my recommended
Me: is not disappointed in the algorithm
Amazing! take a look at the paper pickup mechanism on consumer printers, it might give you some inspiration on how to improve the card dispenser
Great tip, thanks!
@@3DprintedLife Biggest difference between what a printer does and yours is the orientation. Printers don't try to take the bottom sheet of paper, they take the top one. The paper sits on top of a spring-loaded tensioner that holds the stack of paper at a constant pressure against the rollers.
Also, their rollers are always a flat profile, not the round one you get from the o-rings. I suspect it gives move grip for less tension, so there's a lot more friction between roller and paper than the paper and the next sheet downwards.
THIS WOULD BE A FUN PRANK WITH YOUR FRIENDS!!!
guaranteed Vegas has been done had this technology
you could greatly decrease deck restoration time by pushing the bottom 4 then the top 4
I feel like the image recognition part can be done a lot faster with a tiny AI model that can run on raspberrypis.
Super Cool! Most of the issues you were having with the cards, the biggest one being able to get a single card, I would have to say were the cards. Bicycle cards are great cards, but if you were to use them at a NLTH tournament, they would last about 10 hands max. If you had some plastic poker cards, the Copag Poker Jumbo are great around $10/pack, the cards would become a constant, not a variable. They would always be, and only be, as slick as when you begin, unlike the Bicycle which would start one way and continuously change with use. I really want one, but dont think Im at this level yet.
OCD can be emulated by comparing all master images with the scanned one and take the one with the highest % score. It's not how OCD works, but it's a lot faster.
maybe add a button so you can just press it to start the machine with previous config
Hello stranger! Glad to see you’re still with us!
Haha yes I am! Will likely stick to fewer but higher quality videos from here out, so it really may be 6 months until my next one :D
I’m sure casino use these type of programs to increase house favors. There are many CC trying to reject this idea it’s absurd
Can’t wait to see the Uno version
Must cover up wires and
How
Its
Built
(Nvm UNO!)
Thank you, a lot, for showing and explaining this. Very useful. And amazing great work you did, well done!
i had a friend who had a cheapo shuffler. you split the deck in 2 parts and it would shove the cards randomly from each pile in the center, do it a few times and it would be shuffled
You should try using dielectric grease for the noise. It should help out a bit
5:39 When the sreen turned black I thought my PC crashed
Hoping to see v2 of this soon!
How does this channel only have 24k subscribers???
OK now I can get into agriculture! :D Amazing work, bud!
Damn, I didn't even realize I was subbed to you, bought a 3up kit wayyyy back when. Cool project!
Awesome, man it's been a long time since the 3uP days. Well thanks for the continued support!
Due to strong interest in this project (and wanting to recreate it) I did some research calculations. I found that with 8 bins the chance of the bottom card ending on top is about 790x larger than the top card ending on top. This quickly decreases to 2.3 times after 2 shuffles and to 1.1 (neglectable) after 3 shuffles. I thought these results might be interesting ; )
I hope it's OK to use this concept for inspiration for a future design project of mine.
Really interesting, thanks for sharing. And definitely it's ok to base a future design off of this. That's what open source is all about!
This is super cool project! Nicely done
Thank you! Cheers!
Bro , this person built this.. now all the casinos have “automatic shuffling shoes” casinos probably spent millions to develop AI technology to use against us
After doing the normal legit shuffle, the first seven visible cards are all hearts. That seems highly improbable
Every engineering channel on RUclips: I found a reasonable version online but “didn’t want to spend the money”, so I spent 10-100x (Stuff made here) the cost using my own time to create a super specialized, cool version that has little to no practical application.
Or to put it another way: it’s fun to make shit.
I suffer from this too. Getting ready to launch a Kickstarter myself.
Is there any other way?!
Bro, have I have an ad blocker and you are the only channel that I turn it off on
The ultimate compliment, thank you! :D
I appreciate your oxford comma.
2:14 Wrong payout. Cat gets 3:2 on blackjack at a good casino. 😺😺
the dispenser would colloquially be called a " shoe "
in theory , you could have a realistic 1 , 2 , 3 deck-shoe game of blackjack - which was invented to stop card counting , the more decks in the shoe, the harder they are to count
It looks like the image processing is taking way longer than your mechanics can handle. (@11:34 you can see the shuffler just waiting.) It's worth looking into some form of faster computing.
You should add a unshuffle option to put it back into the order of a new deck!
Nice project!
I can’t imagine the time and effort. Wow!
Hey I am a senior in highschool and I’m making my own version of this for my engineering project and was wondering if you have a wiring diagram for it and what voltage you used as your power source for the micro motor on the card shooter. Great project and I’m having a lot of fun building it!
I love these kinds of projects and videos. Keep it up, subbed!
Thanks for the sub! I've gotten such good feedback on this video that I've already began 3 new projects, so more content will be coming soon!
Lol so your first idea is basically what I wanted to do to shuffle MTG commander decks
Almost identical (except I was planning on using another wheel to push them off to the sides)
Edit: though i think the pusher is probably better and faster, might be harder to split the deck evenly in 2 though
Great design and explanation. I'm wondering if you could estimate build cost (assuming access to a 3d printer, say), and say anything about "maintenance" of the machine given regular use (eg do you need to recalibrate the camera, launcher etc regularly)? Sorry if these questions were answered in the video and I missed them.
Total cost is somewhere around $200-300 and there's no regular maintenance, but the card dispensing mechanism is unfortunately pretty unreliable so most of the work to keep this running involved adjusting the gaps on that part. It also would require more uniform lighting and opaque side panels to get the most reliable card detection due to the way the computer vision code was written, though I now know of some ways to improve this significantly if I ever complete V2 (which I started but is a big project!)
You're a genius...
6:20 Ahhhh please don't leave the sticky like that
feels good to be a new memver
Everyone says lube for the gears. Sounds crazy but anti-seize works amazing for 3D prints it helps to wear in the gears with the sand in it.
Very interesting!
I know you said you had no interest in making these, but let me try and convince you otherwise. Machines like this, with the same capabilities run between $12-15k new, $4-8k refurbished. Let’s say at scale you could keep your cost to $200-300 . I would be willing to pay $2.5-3k for a shuffler like this. All you need is a little refinement, a case and a UX .... could be just buttons or a whole separate small LCD like the Shufflemaster Deckmate II. I’m telling you there is a huge market for something like this for weekly poker nights.
Looks like you read the patent on Shuffle Master. The casinos cheat the players using this same style machine.
In some sketchier places this may be true, but most of the time they wouldn't bother. It isn't worth the risk for them to get caught doing this when all the games are already designed to favor the house. The shufflers do have this feature but it is strictly used for verifying the deck and re-stacking in the unshuffled order.
@@3DprintedLife I’ve gambled over 500k. I can assure you that all the casinos are cheating with these machines. The games don’t play anything like they used to. Ask any old timers how much easy it used to be 10 years ago when most tables were hand shuffle and you’ll see that the games are not what they used to be. The games are much harder to win with machine shufflers. Go watch a casino shuffler at a poker table and then go watch a hand shuffle game and you’ll see they are very different. This is why high roller tables are mostly hand shuffle because most people with a lot of money aren’t stupid. 10 years ago if you were playing baccarat on a hand shuffle you would find a 10-25 run bank or player. Now it’s so rare to find a run with the shufflers it goes more back and forth. The casino loses big money on long runs and the machine took those runs away. I can go to any casino with a shuffler and prove to anyone how they are cheating in real time. Especially Indian Casinos like Foxwoods and Mohegan Sun.
11100% would buy this if it were available. Just for the proper shuffling capabilities for our weekly poker game
This is seriously so awesome
Poker hands would be harder to rig, because the dealer must cut the cards after retrieving it from the machine. So the dealer would have to be in on it and either do a false or precise cut to get the desired order of cards. Or the programmer rigging the hand would have to guesstimate where they think the dealer will cut cards.
Yes you’re very right and that’s exactly why cutting the deck is a thing. There are some theoretical hands where you can guarantee or almost guarantee a certain player wins no matter the cut but there are few combinations that can do this. And they of course only work with a specific number of players
@@3DprintedLife If it's just two players it is possible to stack the deck. So no matter what cut is made a certain player will win the hand, but no idea if you can do that with 8-9 players.
Yeah I think I’ve only seen options that work with
This is a StuffMadeHere level project, why doesn't it have more views?
No clue but I appreciate it! I'll catch up to Shane one day 😁
1:11 Hmmm... Call me a cynic but they might not want people knowing about the hacking they've already done.
This is why you cut the deck after the shuffle, and/or remove some number of cards from the top. Nice project though.
Even when you cut the deck it’s still a fixed deck. It’s how the casinos cheat.
@@EliteCannagivers For casinos I could agree as they seem to take a fixed number of cards to the side. But if you cut the deck randomly, I don't see how you could keep it fixed.
@@kilfoofan it doesn’t fix an individual player. The deck only gets cut in half. It doesn’t matter how many times a person cuts a deck. The cards are still in order of the machine. That order controls when the bad beat jackpot gets hit. That machine sorts cards to make a lot more people get hands which creates a crazy table. More people go on tilt at a crazy table. Then those players start playing other table games because of emotions. Like I said cutting the deck doesn’t change anything. All it changes is who gets dealt the fixed hands at the table. The tables are all fixed.
@@EliteCannagivers I still have to disagree. In order it to be truly a fixed deck the hands would have to be predetermined and known, i.e. you'll know the outcome of the table even before you deal a single card. However, if you cut the deck randomly, it can't be known as the first card dealt is not yet known. You could know what everyone has, and what will be the community cards (if hold'em or omaha), based on your hand (if you can remember the exact order of the 52 cards for the individual shuffle, that is), but you could not predetermine the dealt cards as shown on the video, e.g. player 1 gets full house, player 2 gets 4 of a kind, player 3 gets straight flush. For this you can't have any randomness to the shuffle, including cutting the deck. It's like if you would take this post, cut it randomly, and then try to read it with the original word lengths from the beginning, but starting from the cut position. The outcome would be gibberish.
@@kilfoofan all you have to do is read the patent on Shuffle Master. It says it all in the patent. It also states in the patent that it can all be done wirelessly. People want to talk so much but no one wants to put in any research to see the real truth. Shuffle Masters don’t shuffle cards they sort cards. It’s just a fact stated in the patent as well. Maybe instead of replying you’ll use that time to read the patent.
would be cool to cut these components from acrylic using a laser since most parts are very boxy and at right angles. awesome projects and r&d! integrating the different components (mechanics, steppers, opencv, codebase) is very challening starting from nothing. how does it handle different sized cards? If I wanted to build this id want to shuffle any deck of cards (not necessarily poker cards).
Yeah the rear plate is laser cut and definitely could convert more components to use this, but I wanted to keep the design as open-sourced as possible so people without access to a laser cutter could still build it. I haven't tried any size other than standard poker cards but it probably wouldn't work great with others. And honestly, this was a fun project but not nearly quick or reliable enough to justify using over traditional hand-shuffling. I do hope to get a V2 made sometime soon, but I keep coming up with other more interesting projects which will get worked on first!
@@3DprintedLife I would like to help with V2. See github for some questions and a pull request.