Speaking of this, someone devised a system in which you can set a deck such that no matter where it's cut, the button gets the winning hand: https://www.benjoffe.com/holdem/. It's kind of interesting in some ways, but it's unlikely to be effective in practice. Yes, the button always has the winner, but in many cases, the other players are getting weak to nothing hands. A player who happened to turn bottom pair will not likely stack off against someone who makes the second nuts on a four-flush board.


Yeah, good point. Even if the number of players stays constant, how would you know who is at the critical seat when the cold deck is in play? I could maybe see it in a home game, with multiple people in on it. The dealer sees the conditions are correct, the mark is in the proper seat, and hits a button on the shuffler to deal a programmed cold deck. But then how many home games have a $13,000 automated shuffler?

I'd be less worried about someone stacking a cold deck, and more concerned with someone simply having the ability to know hole cards. With the internal camera already in place, this seems like an ongoing possibility. Perhaps your cards reveals what everyone else has. I don't know enough about the device or its software to even speculate how to hack it, but I'm also not as tech-savvy as anyone at IOActive.

Hahaha between stuff like this and RTA, I could envision more rules banning or restricting cell phones. And yeah, it just means another device would be in play. But at that point, if someone cheats me out of a pot via use of a vibrating cock ring, he deserves my money.

Highly unlikely. The seed for the random number generator likely originates from hardware sources, such as the current time.

Current time is not hardware.

Well, that's software, even if that specific function could be interpreted as firmware. Then the time is interpreted by more software which controls the spinning wheel, reads the cards, etc. The point is that the device is hackable.

Then you better not play home games. I bet there are more dealers, mechanics, who can stack the deck than there are who can defeat a DM2 in a live game.

What? Of course it is. It measures physical events and clock is hardware, as opposed to deterministic results from an algorithm. It'd be a poor solution by itself and hopefully Deckmate has a better way but anything is better than relying solely on deterministic software for seeding.

The CPU clock is not software. Go ahead and show me your code for replacing a crystal oscillator that will show the current time if I'm wrong, lol.

Those are words. Say time comes from CMOS, the processor, any chip on the board. How would you access that time programmatically? With a function call -- that's software. There may or may not be a hardware mechanism backing the data. In this scenario, hackers can replace the result of the system time function call to return a banana. That's why it's software.

You don't understand how things work. A crystal oscillator doesn't generate time. It creates an electric pulse at a specified frequency that used for timing. How that's used and interpreted is done by software.

You're wasting your time worrying about the RNG. Hacking the machine will involve the software after the sequence of the deck is generated either by transmitting the sequence as recorded by the software, or by reading the sequence using the camera(s). How the sequence was determined won't be relevant.

Of course there are. And it probably won't be the dealer who hacks the machine. The dealer(s) may very well be oblivious that the machine is compromised.

But there are sharp players who may pick up on a dealer stacking a deck. The best you could hope for with a compromised machine is that the results over time don't add up and a sharp player puts in enough time to notice. That would be a lot harder to detect.

As for whether I'd play in a home game, I would if I didn't have a reason to be suspicious. If I became suspicious I'd watch for things like how the dealer picks up the discards, false shuffles, whether the dealers places the deck on the table before cutting, etc. Not that I'm an expert and would catch the best of them, but I'd probably catch a less talented mechanic if I was looking for it. Ultimately, if I became suspicious I would stop playing even if I couldn't detect anything, as I would be wasting my efforts worrying about whether the game was legit, rather on the game itself.

it creates a VIBRATION at a specified frequency.

This is like listening to a geologist and an astronomer argue about whether the earth is oblate or round.

Which might make for a fun toy for the Mrs. But it wouldn't be of much use in an electronic circuit if that vibration wasn't creating an electric signal.

Hello, I'm one of the three researchers from that article. We wanted to share some information here since 2+2 is such a big platform for poker. Firstly, we've released a 69 page whitepaper that goes into detail on everything we covered in the black hat talk and the article in much greater detail.

You can find that here: https://ioac.tv/3YD9t36.

And then we wanted to answer a few questions/comments you folks had:

The deckmate shufflers have no RFID readers - the deckmate 2 read the cards with a camera as they are being shuffled.

This is interesting! During this research we did quite a bit of initial research online to see if anyone had thoroughly looked at these shufflers (and released the details publicly). We didn't see anything like that on 2+2, but we'd be curious to read these old posts if you can link to them. We did see the Joey Ingram interview that discusses a proposed cheating scenario nearly identical to the one we present in the paper, but didn't actually see it until after the research. Furthermore regarding discussion here and elsewhere that this technique was "known" in the past - we don't disagree per se (after seeing the old threads and videos) but would make a distinction: we were not looking here to speculate about what might be possible, but prove, with thorough documentation, exactly what is possible and how it could be done.

This is mostly correct, though the deckmate 2 is not a wheel shuffler - it has a platform featuring 27 slots (two cards each for a total of 54) that moves up and down to meet the cards as the are being shuffled from the intake platform. Functionally though, it works essentially as you describe. The suit and rank of each card are always recorded for verification purposes, but for a normal shuffle, these values are not used to determine position, and they are instead placed randomly based on their original position in the unshuffled deck. You're correct about sort mode - it does support it and it works by reading the cards with the camera before placing them.

This is more clearly explained in the whitepaper, but the fix you describe wouldn't completely prevent the kind of cheating we demonstrate/describe in the article. The technique we used is capable of accounting for the deck being cut, and in fact the full order of the deck and the contents of all hands at the table can be known as soon as a single one of the cheating player's cards are known. Thus, cutting or riffling the deck after the hands are dealt would limit the cheaters knowledge of the flop, turn and river, but they would still have a huge edge (knowing everyone's hole cards).

This is something we were interested in learning when we started this research - were the shufflers, unmodified and running original software, designed in a way to skew the odds or allow for some kind of cheating? After thorough review of the software of both shufflers, we found no evidence of the shufflers doing anything other than what they were intended to do: shuffle the cards randomly (and verify decks and sort, when relevant).

The purpose of the camera in the deckmate two appears to primarily exists to prevent cheating, actually. It is used to read each card in the deck and report to the dealer if there are any missing or duplicate cards which may have been pocketed or inserted by a cheater. The sort mode seems to be a nice bonus, but not the main reason for the addition of the camera.

The USB is used for two main things - first, the deckmate 2 supports an external display which the dealer can interact with for configuration and for displaying things like the player clock. It also is used for maintenance and software updates.

Regarding the discussion on RNG, we cover entropy and RNG in detail in the whitepaper for both shuffler models, and it's explained better there than I'd do justice to here, so please give those sections a read.

Obviously there is a great deal more discussion in the thread and I won't address it all in one post, but feel free to ask us questions and we're happy to answer.

The 2+2 poker forum is over 15 years old now. It is no longer a baby. And the collective knowledge of the people that post here in terms of years in poker and information is vast. It is a shame that a few bad apples destroy some of the most important threads and topics. And it is the same very small group of *******s causing the problems. I cannot prove or disprove when you saw the Joey Ingram interview. But at least you acknowledged that some of information in your research was available before your research. My issue you is not with you getting the word out, I just wish 2+2 had less *******s bring it down, then the word could have gotten out better and further here.

One small point, it feels like what you wrote to me here does try to dance between the raindrops a little. Your use of the words "proposed" "known" and "speculated" make it sound like you were somehow the first group to ever get your hands on a shuffle machine. Well, I have friends in low places and can tell you first hand that you are not the first group to get your hands on a borrowed or bought machine. The stories told in many a former 2+2 thread including the Joey Ingram stories were not "speculation" about what could be done. It was people telling what had already been done.

On page 12:

"Furthermore, it was pointed out after the 2023 World Series of Poker by a dealer at that event that not only was it possible for players to reach the ports on the shuffler, but that players were regularly and openly doing so – the USB port was being used as a convenient place to charge their phones."

Oh, that's friggin' great.

I think what matters in the end is public information. What your friends in low places did and their results doesn't seem to be something regular players have access to.

You may be misunderstanding our intention with the research. We're not suggesting we're the first group do investigate these shufflers, and in fact if you read the paper you'll see that it's the opposite - we're suggesting that it's likely that the issues we identified, and the cheating techniques they enable, probably have been known to cheaters prior to our work, and could certainly have been exploited in the wild. The distinction, again, is that none of the previous groups which may have worked on these shufflers ever published their findings (for obvious reasons). Our goal with releasing these details is to make poker players, casinos, card rooms etc. aware of the potential cheating, so that players can protect themselves and manufacturers/card rooms/casinos can work to address the vulnerabilities that allow for cheating. Also, it was fun, and the we're hoping others find the details interesting as well.

https://ioac.tv/3YD9t36

From page 34:

"Entropy and Random Number Generator

The DM2 machine controller board does not feature any hardware or cryptographically secure RNG. Rather, it relies on software pseudo-RNG (PRNG), where a relatively unpredictable seed value is used to deterministically generate pseudo-random data. The RNG is initialized as part of system init, in a function dubbed initialize_system_features by IOActive."

Thanks for making that whitepaper available. It was a fascinating read.

Do you plan on releasing any videos, such as on your YouTube channel, demonstrating your hack(s) in action? I'm sure most of us would be interested in that.

He is not going to teach you how to design the software for the usb so that you can hack your home game bro.

Two words:

anal beads

Butt plug.

CMOS clock and windows time are two completely different things.

Don't know why you jumped to THAT conclusion. I meant showing it on a macro scale, in action at the table. Not a primer on how to write the code or how to embed it. I believe they've already demonstrated it publicly, just not widely so we all can see it.