We're hoping this column becomes your window into roleplaying design and development -- or at least the way we approach these things here at Wizards of the Coast. We'll handle a wide range of topics in weeks to come, from frank discussions about over- or underpowered material, to the design goals of a certain supplement, to what we think are the next big ideas for the Dungeons & Dragons game. All of this comes bundled with a healthy look at the people and events that are roleplaying R&D.
The following design notes, straight from Rob Heinsoo, concentrate on how he created Three-Dragon Ante as a game you can play with both your D&D group or your non-gaming friends. In a couple of weeks, he’ll return with another article touching on games within fantasy worlds -- specific on Three-Dragon Ante as it appears inside the world of D&D.
Today’s article is broken into three sections.
- Section 1 covers how I came up with Three-Dragon Ante, and how it took shape during playtesting at Wizards of the Coast.
- Section 2 explains why Three-Dragon Ante deliberately plays differently than the king of gambling card games, poker.
- Section 3 details some game design lessons I learned (or relearned) while hammering the game into its publishable form. It’s something like an outtakes reel, a series of bad ideas that got better in the final cut.
1. Origin Story: A One Year Turnaround
Three-Dragon Ante started miles and miles away from most Wizards of the Coast games -- I was snorkeling above a coral reef while on vacation in Hawaii. My thoughts drifted out of vacation, back to playing D&D on the mainland; I realized that I wanted the PCs in my Wednesday night campaign to play a card game in a tavern they were about to visit, on a mission of mercy for a dying friend.
So I wondered what card game they’d play in the tavern. Not poker. Playing poker didn’t make any sense to me. My background is in cultural anthropology, and I couldn’t stomach the sloppiness of pretending that D&D’s fantasy world could (or would want to) recreate the specifics of our own culture’s 52-card deck, or even its disguised cousin, the tarot deck.
I circled around a chunk of green coral and thought about how I might focus a gambling game played in the world of D&D. Something involving dragons seemed right. That way I could use the ten main colors of dragon as the suits, with each type of dragon having its own unique power. There’d be big piles of gold on the table, and each player would play dragon cards to try and win the stakes.
By the time I swam back to the beach, I’d worked out the game’s basic mechanics, in which high cards help you win the stakes at the end of each hand, and low cards win micro-rewards by being more likely to trigger their powers.
I thought about the game for the rest of the vacation, tinkering with possible card powers and card-drawing mechanics while hiking or snorkeling. I had the first deck ready by the time I got back to the mainland (this was mid-January 2005). I knew it didn’t quite work, but I eventually got a game together with three other Wizards designers. Two or three big rules changed in the middle of the first game (see section 3 below!), but the concept looked like it was going to fly.
The First Playtests
A draft or two and a couple weeks later I played a game with Andrew Finch, Wizards of the Coast’s Director of New Business. By coincidence, Loren Greenwood, WotC’s CEO, came into Andrew’s office just as I’d started demonstrating the game, and stuck around to play a couple hands. He won the hands resoundingly, playing an all-Gold Dragon flight and collecting gold from every player as well as picking up the stakes. Loren left the office saying that he really liked the game; and no, we hadn’t tried to hand our boss the win. He genuinely kicked our butts.
In the next couple weeks, I played the game with my bosses in roleplaying R&D and the D&D Brand folks. The consensus was that making a card game played by Dungeons & Dragons adventurers was a good idea and that the flavor and mechanics of Three-Dragon Ante handled the job. So with remarkably few complications, the game went onto the publishing schedule, aiming to be out by the end of 2005 -- which made it an extremely quick turnaround.
The Look and Feel
From the start, I’d said that I wanted the game to look like a deck that might be used in the D&D world. That meant we weren’t going to use the modern-illustration style from our Monster Manuals and D&D book covers. We had to find an artist with a style that could look something like woodcuts done in a fantasy rather than a straight-medieval world.
The sample piece that sold the art directors on using Craig Phillips as the artist later transformed into the game’s beautiful card back. I’m aware that some fans are surprised by the art style, expecting something more like standard D&D illustrations; nevertheless, I’m extremely happy with the cleaner style Craig used. If the game does well, there might even be room for another style on a deck produced by a different socio-magico strata in the D&D world… because there are certainly multiple Three-Dragon Ante deck designs in each D&D world, just as there have always been multiple playing card designs in our world.
2. UnPoker: Because Poker’s Been Done
As a gambling game, Three-Dragon Ante lives in poker’s world. Poker is a splendid game, and I haven’t been certain that it needs more variants. So my goal with Three-Dragon Ante was to create a different gambling-game experience. I wanted to create UnPoker, a game that would share poker’s mercenary attitude but avoid the things I like least about it.
In part, poker is about knowing when to fold. If you’re playing right, you’re dodging out of many unfavorable hands. However, I wanted Three-Dragon Ante to keep everyone at the table involved at all times – no folding in the middle of hands. In my first thoughts about the game, I decided that low cards would be useful in that they would have powers that could earn you little micro-rewards, reasons to feel at least a little good about needing to stay in. So when you’re not playing to win a gambit, you can still play to steal gold from the stakes (using Black and White Dragons), demand gold or cards from other players (Green and Brass Dragons), or draw new cards to set yourself up for the next gambit (Silver and Gold Dragons). Of course, if somebody plays the Druid and its power triggers, you might just end up in the running for the weakest flight and still sneak away with all the gold at the end of the gambit.
Plus, if you’re lucky enough to have three dragons of the same color or the same strength, you might be able to collect more gold (or at least more cards) than the eventual winner of the gambit. Even starting with two dragons of the same color gives you a shot at finding a third, especially if they’re good dragons or you know how to finesse your way into running out of cards so that you have to buy more before the end of the gambit.…
Table Stakes, Not Unlimited Stakes
Poker has another feature I wanted to avoid: infinitely escalating stakes and bullying. Poker is a full-force psychological brawl, and players with more money can often beat lesser players into submission with their bankrolls. I don’t have a problem with that from a game design sense, my issue is personal. I wanted to design a game anyone could play with their family! This meant that I needed to keep the game simple, preferably about as simple as Hearts. And I also needed to keep it from being all about psychological warfare, because my family won’t play poker with me.
This isn’t to say that Three-Dragon Ante is a game for people dedicated to being nice to each other; try playing three Green Dragons in one hand and see what the player to your left thinks of you. The big moments in a given gambit usually involve screwing at least one other player. But by saying that every Three-Dragon Ante game starts with each player supplying 50 gold pieces they’re willing to lose, the potential losses for a given game are limited to roughly 50 gold apiece. When the player who’s in the lead antes big to try to force out the weakest player and win the game, the weakest player is fighting to keep on playing and win some gold back, rather than fighting to hold onto all their worldly possessions.
3. The Nitty Gritty: Specific Design Issues
The Zeroth Draft
Game Design Issue: Simultaneous card play.
The very first draft of Three-Dragon Ante asked all players to lay their cards down simultaneously. I’m trying to reconstruct my reasons for thinking why that could have been a good idea. I must have thought it would be quicker, which wasn’t necessarily going to be true given that power-triggering still depended on looking at the card played by the neighbor to your right.
Maybe I thought it would be more dramatic to have all cards played simultaneously. Certainly I was staking out ways in which Three-Dragon Ante could be different than other card games, and I liked the visual drama of players placing their card in front of their forehead and then slamming them down on the table.
But simultaneous play increases a game’s randomness while removing most elements of strategy. As we started our second hand/gambit, Henry Stern and Mike Donais wanted to know what simultaneous play added to the game. I had no good answer, so we started taking turns playing cards, clockwise around the table….
Killing the High-Low Flip Flop, Tinkering the Ante
Game Design Issue: Randomness reduces strategy.
Before writing this article, I’d forgotten that there was another dead-end complication that failed to survive the first five minutes of actual game-play. I’d come to the table thinking that we’d flip a coin at the start of each hand. If the coin came up heads, we’d be playing a Fangs hand, and higher cards would be the cards that triggered their power during their turn. Tails? The lower cards’ powers would trigger.
How embarrassing. As soon as playtest cards hit the table, the Fangs/Tails conceit revealed itself as an ugly monkey-glitch tossed into the game’s potentially elegant high-low triggering system. By the third hand, we stuck to the idea that high cards were good because you could win the stakes and low cards were good because they were more likely to trigger.
That original zeroth draft version of the game had also turned over a random card to see how much gold each gambit was being played for. Yes, completely random stakes, another overly-random simplification I’d included to make the game play quickly. It turned out that the game wanted a bit more strategy and less luck. By the time we were finishing the first playtest, I’d realized that players should be anteing cards out of their hand to determine the stakes. Now players have to judge their chances of winning a gambit to determine which card they want to ante, and if you pay attention to the card an opponent antes, you’ll have a better understanding of the cards that might be in their hand.
Getting rid of the coin flip cleared the way for the ante-card as the means of determining the stakes, which in turn opened up good powers for cards like the Bronze Dragon and the Archmage. The game never looked so bad again, and I nearly forgot it had ever been that dumb.
Bad Point Spread
Game Design Issue: Wide differences in card power create hopeless situations.
From the start, I’d rated the dragons’ strength ratings to simulate the power of D&D’s different types of dragons. The strongest White Dragon was only a match for a middling Blue or Red Dragon, the Red Dragon had the highest average strength among the evil dragons, and the Silver and Gold Dragons ruled the good dragons’ ratings. But my initial numbers were too widely scattered, ranging from 1 to 23, making the difference between high cards and low cards too extreme. Mike Donais, Rob Watkins and Mike Turian advised narrowing the scale, and when I started messing around with that I found the 13-point scale from our earthly 52-card decks staring me in the face.
Three-Dragon Ante decks still don’t try to make all the suits equal: some dragons are more equal than others. But the 13-point scale gave me a nice impressive number for the biggest dragons (Bahamut, Tiamat, and the biggest Gold) and worked well with the ante mechanism, while still giving me enough range to differentiate between the dragon types. The second try at the point-spread survived through publication.
Good & Evil
Design Issue: Capturing the flavor of good and evil dragons with game mechanics.
The original design said that good dragons had powers that let you draw cards and that most evil dragons had powers that let you steal cards from other players. The idea was that the good dragons could create something new. The evil dragons just ‘redistributed’ resources, forcing you to buy new cards eventually if you only played evil dragons. Cute. But it turned out that this meant too much stealing. No one could form a coherent strategy when they were always losing cards they had intended to play later. The problem got highlighted when my boss’ boss, Bill Slavicsek, spent his entire first game buying new cards to replace the last cards that had been stolen before he had a chance to play.
I tackle design problems like this in my background-brain, thinking about them while driving or hiking. It took about two weeks to put together a new set of powers, keeping the idea that good dragons can set you up for winning future hands and that evil dragons mainly give you the instant gratification of grabbing things that don’t belong to you.
Correcting the Downward Spiral
Design Issue: Recognizing an opportunity for better gameplay.
The game worked well with its third draft. But there was still one problem: I was still playing that in order to trigger, a card had to be less than or equal to all cards played in the round. That was a death spiral effect, and not much fun. When Richard Garfield and Bill Rose played the game one weekday, they both had the same comment: “Play so that your card triggers if it’s not stronger than the last card played in the round.” This single change got rid of the death spiral, and meant that a person playing a tiny card was only going to screw the player to their left, not everybody. This change made all the difference. Three-Dragon Ante games became fully fun, and there wasn’t another big change made to the mechanics.
That’s it for now. See you in a couple weeks with more notes on Three-Dragon Ante place inside D&D.
About the Author
Rob Heinsoo started playing D&D with the original brown box in 1975. He's now the lead designer of D&D Miniatures.
Thoughts or suggestions for this article? Topics for future Design & Development articles you'd like to see covered? By all means, please feel free to write directly to the authors, at: firstname.lastname@example.org.