Using Poker Hands in TTRPGs

Every now and again, the idea of using poker hands as an RNG for a TTRPG comes up, usually in the context of some kind of western. How would that work?

Five Card Poker

The most obvious way to do this is to just deal out a full five-card poker hand to each player, then let them play a hand whenever they make a check. A better hand is more likely to succeed. After playing a hand, they discard it and draw an entirely new hand.

About 50% of all poker hands are a single, and about 43% are a pair, so you would probably want the standard TN for something anyone could accomplish to be Jack-high, and the TN for most things that you would expect to require expertise to be a pair of Jacks. Your odds of getting two pair or better are about 7%, which is pretty close to a critical hit on a d20. It’s unlikely, but not so unlikely that you don’t usually see one or two in every session. A three-of-a-kind, straight, flush, and full house are all much more rare but still common enough that you’ll very probably see them happen over the course of an entire campaign, loosely comparable to rolling triple sixes on a 3d6, which does happen now and then. Once we get into four-of-a-kind and especially a straight flush or royal flush, we’re getting into territory that’s technically possible but unlikely to come up even once across an entire mid-size campaign.

But this raises the question: How do you add in skills? If your gunslinger is real good at shooting, how does that get reflected in this system? The most obvious way to do it is by allowing them to draw more cards to make a hand out of, but working out the probabilities on that is a huge headache and the value of additional cards goes down the more of them you already have. If you want a narrow range of power between the weakest and strongest characters in the game, that can still work, but even a fairly mundane wild west game wants a deadshot to be much, much better at shooting than a shop keep with broken glasses, so this extra cards approach only works if every character in the game must be a gunslinger with more-or-less the same skills or else if you’re willing to have a game where characters draw 20+ cards to assemble a poker hand out of when they’re using their best skill. At that point, you don’t really feel like you’re playing poker at all.

The second most obvious solution is to convert the poker hand into some kind of total number, to which your skill bonus is added. The problem here is that it’s hard to find a conversion that works across all poker hands. For a single or a pair, the answer seems obvious: Just add the cards together. This way a pair is, on average, worth twice as much as a single. Here’s the problem: Two pair (a hand with four cards to add together) is much more common than three-of-a-kind, and four-of-a-kind is less common than every five-card hand except the straight flush and royal flush. To make the convert-to-number method work, you have to give special conversion methods to most hands (although not the most common hands, at least) to make less common hands more valuable. For example, with two-pair sevens and threes, instead of just adding 7+7+3+3 for 20, you might add 7+3 and then multiply by 1.25 and rounding down for 12. Since three-of-a-kind is on average worth 1.5x a pair, having two-pair be worth on average 1.25x one pair is mathematically sound, but also now you have multiply things by 1.25. Then you have to find some similar conversion for the straight, flush, and full house that all give average results between three-of-a-kind and four-of-a-kind, even though four-of-a-kind is only 1.33x the value of three-of-a-kind. Or you can add a multiplier to four-of-a-kind beyond just adding all four cards together, at which point you’ve sacrificed what little consistency this system had.

The best way (of these three that I’ve thought of, anyway) to incorporate skill bonuses into a poker-based RNG is to have skill bonuses just promote the kind of hand you have. If you have a pair of sevens and a +1 skill bonus, your hand is now two-pair of sevens and some other card that is lower than sevens. With a +2 skill bonus, it becomes three sevens. With a +3 skill bonus, it becomes a seven-high straight. With a +4 bonus, it becomes a seven-high flush. With a +5 bonus, it becomes a full house, sevens over something else.

Two things you’ll notice about this system: First, it’s possible to get normally impossible hands through skill bonuses. If you get a pair of twos with a +3 skill bonus, you now have a two-high straight, which is normally impossible. If you have a pair of sixes with a +4 bonus, it becomes a six-high flush, which is normally impossible because the only way for six to be the highest card of a flush hand is if it’s a straight flush. This is weird, but doesn’t really impact anything. Just roll with it.

Second, we only ever care about the number value of one card in any hand. For two-pair, we only care about the value of the higher pair. For full house, we only care about the value of the triplets. For a straight, we only care about the highest card of the straight. This makes things simpler to resolve anyway. Instead of having over a hundred different combinations of a full house each of which represents a minutely different TN, a full house represents the same two-through-ace scale as every other hand. Jack is the default difficulty, but minor circumstantial bonuses or penalties can nudge the difficulty up to a Queen or down to a ten or whatever.

Hand Building

Another way to do this is with hand building. That means that when you make a check, you don’t discard your whole hand. You discard only the cards used to make the check. You can play a single card to get it out of your hand, taking a dive on whatever check you’re making to try and build up a stronger hand. If you’ve got two-of-a-kind, for example, you could hold onto that and junk other cards until you’re able to get three, or two-pair, or a full house, or something. Building a flush would mean taking a dive on a lot of checks, but it’d be much, much easier to build a flush than to draw one straight from the deck. In fact, it’d be easier to build a flush than a straight, since each card you replace when building to a flush has a 1-in-4 odds of building your flush, but only a 2-in-13 (or slightly worse than 1-in-6) odds of building your straight. Actually, the odds aren’t quite 1-in-4 or 2-in-13, since the cards you’ve already drawn from the deck affect the odds in ways that vary depending on exactly which cards are in your hand, but the basic math still works out: Drawing a flush straight from the deck is less likely than drawing a straight, but building a flush when you get to select which cards to replace will usually go faster than building a straight.

And this also encourages players looking to get rid of junk in their hands to find skill checks that don’t matter so they can safely take a dive on them. This leads to players interacting with the mechanics instead of the narrative, attempting tasks which are challenging enough to warrant a check but which have nothing to do with anything just to manage the abstract, game mechanical resource that is the cards in their hand. Proper game design should be making mechanics and narrative harmonious, so we’ve definitely gone off the rails here.

Hand building sounds interesting, but really only works in board games where you can tightly control when checks are called for in order to guarantee that all checks are relevant and you can never safely take a dive. Then hand building is about minimizing the damage of getting rid of junk, rather than wasting time with actions that totally negate the damage. Unfortunately, the open-ended nature of a TTRPG makes it basically impossible to prevent players from finding things to do that call for a skill check while having no real consequences for failure. They can try to play a fiddle and the worst that’ll happen is they aren’t very good, try to chat up the ladies at the saloon and the worst that’ll happen is they’ll get slapped in the face. They can even throw out some junk cards on activities that could hypothetically cause great harm, but won’t do so except for catastrophic failure. They can try to swim a river to get rid of an unneeded seven of clubs, and while they won’t make it to the other bank, they’re probably not going to straight up drown unless they play a two.

Texas Hold ‘Em

The last RNG idea I’m going to look at here is Texas Hold ‘Em. Each player gets a hand of two cards, and there’s a river of between three and five cards shared between them. With a seven-card hand, the odds of a pair or better are 82%, which means you can make pair of Jacks the base TN for challenges that an ordinary person might struggle with but will probably be able to manage. The odds of three-of-a-kind or better are 15%, which is a pretty good spot for something that experts can manage easily but ordinary people will struggle with, which means you have two pair in between if you need more granularity in the “routine for experts, tough for ordinary people” space. A flush has about 5% odds, making it the natural 20 of the system, and even a four-of-a-kind is not so unusual that you wouldn’t expect to see it once or twice in a campaign. Only the straight flush and royal flush are so rare as to be unlikely to come up across an entire campaign.

On the one hand, it’s good that more hands are coming up. There’s more granularity in the scale of TNs, since more of the potential results are actually achievable, compared to the five-card probability spread where singles and pairs totally dominated the space with only 7.5% of the entire scale left over for anything better than a pair. With seven cards, we have TNs that are 80%, 40%, 15%, 10%, and 5% likely to be hit, and then four more TNs higher than that which are mostly the domain of people with skill bonuses, although an untrained person can still hypothetically hit them. This is a pretty good spread. It gives you room for characters with bonuses ranging from -1 to +3 where each step on that scale is a big deal, so you’d expect this to be the kind of game where you probably don’t level up much, which makes sense. You don’t usually get zero-to-hero stories out of the old west.

The problem is that it’s kinda hard to do the Texas Hold ‘Em thing with three different rounds of revealing cards on the river for every single check. Probably you’ll just put five cards on the river all at once, and then replace them all once a check is made. You could replace just one card, with a first-in, first-out system so that the card that’s been in the river the longest gets removed, but then you get into all the problems with hand building and the bizarre meta-gaming behavior it encourages, with people trying to make useless checks to get the river moving when it’s bad, and trying to avoid all checks to keep the river still when it’s good.


I don’t have one, particularly. I just had a string of disconnected thoughts about how to do some kind of poker-based RNG for a TTRPG, and since I have no plans for any game along those lines any time in the future, this blog post was the only outlet.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s