SolForge Fusion began more than a decade ago when Richard Garfield and I met at the PAX Dev conference. Richard is the legendary game designer who created Magic: the Gathering, a game that changed my life.
In 1997, I won the Magic US National Championship. My winnings allowed me to pay my way through college and travel around the world playing Magic. The skills and friendships I built eventually helped me launch a career in game design. Since then, I’ve worked on games like the World of Warcraft Miniatures Game, the Marvel Vs. DC trading card game, and my own deckbuilding game, Ascension.
In 2011, I attended one of Richard’s talks on design. At the end of the talk, someone asked Richard what his favorite game at the time was, to which he replied, “Ascension.” I involuntarily leaped up from my seat and shouted “Woohoo!” After the event, I took the opportunity to meet up with him and talk about design.
Our conversation circled around one question: What’s next in the evolution of gaming?
Our answer: digital trading card games. More specifically, SolForge. This was in 2011, before the era of Hearthstone and Magic Arena, so it was a relatively new space for design.
We started by creating SolForge’s unique three-level system, where cards level up as you play them, creating a unique gameplay experience where you get to customize your deck both before and during play.
SolForge’s Kickstarter launched in 2013, nearly doubling our funding goal—a huge success. The game ran successfully for four years before being sunsetted in 2017.
Since then we’ve been looking for a way to bring back SolForge, but neither Richard nor the team at Stone Blade believed that a straight relaunch was the correct plan. Back in 2011, SolForge was “the next evolution in digital card games,” but ten years later we were worried it was “just another digital card game.” If we were going to bring back SolForge in a sustainable way, we had to return to the question that started it all: “What’s Next?”
The Origins of SolForge Fusion
It was one of Richard’s other designs, Keyforge (no relation), that sparked the idea for the next iteration of SolForge. Based on advances in digital printing technology, it is now possible to make custom cards and decks, something that would have been impossible even a few years ago.
In Keyforge, the decks were pre-generated and 100% unique. This meant that no two players would have the same deck, completely upending decades of trading card game tropes, where players would often copy whatever “netdeck” was winning that month.
I started to imagine what else could be done to leverage this new technology. In other words, “What’s next for digital printing technology?”
The result, SolForge Fusion, is the best of both worlds, giving players the option to customize their deck but still ensuring every player’s deck is 100% unique.
SolForge Fusion is the world’s first Hybrid-Deck game, a game where players combine different unique half-decks to customize their play experience.
SolForge Fusion goes beyond just digitally printing unique decks, but also uses a unique algorithm to digitally create the cards themselves. By “Fusing” two card halves together, SolForge Fusion will have over 17,000 possible cards in our first set alone (more than was created in the first twenty+ years of Magic: the Gathering!), with more unique deck possibilities than there are atoms in the universe.
Moreover, all SolForge decks can be scanned into your online collection, where they can be exported and played via Tabletop Simulator.
Turning a Purely Digital Game into a Digitally Printed Game
In order to make a good sequel to a game, you need to know the core tension of your game. A core tension is the primary constraint that players push against as they play the game. For those who want to know more about the core tension as a concept, you can read more about it in my book Think Like a Game Designer, or in this article.
The core tension of SolForge is the leveling of cards. Players can play any card they want each turn, which also levels that card up. The tension comes from the fact that the more powerful cards, in the short term (i.e. good Level 1 cards), tend to come with weaker leveled up versions. A player will be asking themselves, Do you want to invest in the future by playing weaker cards now or do you want power immediately at a risk in the late game? Keeping this tension high and creating a lot of interesting level progressions is the core of what made SolForge so much fun.
From Digital to Physical: “Certainly the thing that surprised me most about the game was that Justin expected me to physically level all these cards—I wanted the computer to do that. That’s why we had it there. But a lot of work was put in and the more we played with what the design space offered, it seemed worth it, and the more we played with the organizational aspects, the more it wasn’t that much of a burden either.” – Richard Garfield
SolForge was originally designed to be a TCG that uniquely takes advantage of the digital space. In it, each card had three levels which evolved as you played the card, allowing us to tell more interesting stories and give players more interesting play decisions than with a single static card.
Leveling up cards is easy when a computer handles it for you because the process is seamless and automatic. In a physical game, where leveling up is manual, we had to work on making the leveling process fast and efficient for the players. We used several tools to achieve this:
- Graphic Design
- Board Layout
- Card Design
We updated the game engine to make the process of leveling as obvious as possible. Graphic design is a key to making gameplay intuitive. Here are some of the tools we employed to make leveling fast and effective during play:
- All three levels of a card share an art piece and a name to make higher levels easy to find.
- Creature cards show a preview of the stats of their higher leveled version for ease of reference.
- We redesigned cards to have a clear progression on powers so it is easy to remember (or guess) what a leveled card will do.
- Each card’s level is clearly indicated by an icon and a colored border.
- The playmat has zones for players to store their level two and level three cards so they can be easily accessed during play.
- Reduction in deck size.
As you play cards each turn, they are removed from your deck and you replace them with a higher level version. By reducing the size of the deck from 30 cards to 20 cards, the number of cards becomes easier to track from round to round. The deck will always be 20 cards when shuffled until the final rounds of the game. This also had a beneficial effect of reducing game variance and giving players more opportunity to see their leveled-up cards each round.
SolForge Fusion uses a shared turn model, where players alternate control of “The Forge.” Whoever has the Forge plays a card first, followed by the player without the Forge. This process is repeated once more before combat begins. Playing cards one at a time allows the person without the Forge time to search for their level-up cards, making the process less burdensome.
This streamlined turn structure also makes the game go faster and creates a dramatically different feeling from turn to turn depending on whether or not you have The Forge.
Definitive End Games and the Forgeborn
The Forgeborn are the most iconic characters in SolForge. They represent the most powerful characters in the game, those who can actually wield the power of the SolForge directly. In the original SolForge, the Forgeborn were the only creatures that had a fourth level. This gave players something fun to strive for, but since few games would last long enough for the fourth level to come into play, the Forgeborn rarely lived up to their potential.
In SolForge Fusion, we placed the Forgeborn front and center.
Each Faction half-deck comes with a unique, algorithmically generated Forgeborn. Players pick which Forgeborn of the two combined decks they want to use. Every three turns, players shuffle their deck and level up their Forgeborn, granting them access to a new unique power that is guaranteed to have an impact on gameplay.
The Forgeborn are a huge part of the SolForge lore and players can now embody those characters in a far more compelling way. In addition, the Forgeborn are a fun way to track the game progression, as the Forgeborn levels up at the same rate that a player gets access to higher level cards. After the Forgeborn’s mighty fourth-level powers are unlocked, the game goes into sudden death mode, ensuring a dramatic conclusion within a reasonable period of time.
These changes were intended to keep the core of SolForge intact while providing an exciting answer for the question, “What’s Next?”
Algorithms I’ve Known and Loved
(AKA How to Balance a Game with More Permutations than Atoms in the Universe)
I’ve spent my entire adult life thinking about playing, competing in, and building trading card games. After winning the US National Championships, I made a living as a professional Magic: the Gathering player and paid my way through college. I was hired to develop the Marvel Vs. DC trading card game to be competitive on the multi-million dollar Pro Circuit. I’ve built and launched multiple multi-million dollar trading card games in both physical and digital forms.
In addition to my own experience, I have a team of incredibly talented game designers and developers who have worked for many years in this space. And I can say without a doubt that SolForge Fusion was one of the most challenging games to develop of anything we’ve ever worked on. Let’s explore what the process looks like:
TCG Development 101
Developing a trading card game is far more complicated than developing a traditional single-box game. You need to account for all the possible permutations of typically several hundred and over time several thousand cards. This requires you to think categorically about not just the cards you are working on today, but also all the cards that might come in the future.
I’ve written extensively about how to develop these kinds of games both in my Think Like a Game Designer book as well as online articles, but let’s just go through a brief overview of the questions that need to be answered for each release of a trading card game
- What is the core tension of the game?
- How does the current mechanic/card/theme play with the core tension?
- Where is the fun coming from in each faction?
- Are the themes well balanced across each of the Factions?
- Is there enough for each Faction to do?
- Is there enough in the set to appeal to different types of players?
- What fundamental needs does each faction address?
- How does each player psychographic approach each faction?
- How do we balance mechanics at various levels of play?
- What if players have only a few cards vs. an entire collection?
- How do limited formats work?
- How are the different factions balanced and are their core themes coming across?
- Are there viable counter-strategies to each theme?
- What is the appropriate range of power levels across each cost in the game?
- Is there a good balance between different costs and cards for early game, mid-game, and late game?
- Is there enough depth for experienced players?
- Is the learning and discovery curve appropriate for new players?
- What does the average board state look like?
- Are players suffering from analysis paralysis?
- Where are we spending our complexity points?
- Are there the right level of simple, intuitive cards at each rarity?
- Are there enough exciting build-around cards at each rarity?
- What is the experience like of opening a random pack?
- What is the experience like of opening your 10th or 100th pack?
- Are there exciting things to look forward to and discover as you open more packs?
- How long are the games taking?
- Are there mechanics slowing the game down?
- What range are we looking at when considering game length, how many turns?
- Rules and Clarity
- Are the rules clear for the game and on specific cards?
- How do player assumptions match up with the correct way to play?
- What assumptions are we making about player experience with these mechanics?
As you can see, there is a lot to consider! After twenty years in the industry, my team and I can go through the above checklist relatively quickly, though it still takes many months to get a new TCG release ready to go.
When it comes to SolForge Fusion, we had a whole other set of fascinating considerations to account for. Let’s dig into a few of these considerations:
Cards that Level as you Play Them
Level progression is one of the most exciting elements of SolForge Fusion. Because every card you play levels up and becomes available in future turns, you get the experience of not only playing your deck, but also building it during the game. In many ways this was the main fusion between Richard’s TCG concept in Magic: the Gathering, and my deckbuilding game Ascension.
As with the original SolForge game, there is no card playing cost in the game. One of the easiest ways to balance TCG cards is by increasing the cost, and in SolForge, instead we balance cards by the level progression. This means (typically) that the better a card is at level 1, the worse it will be at later levels. To make this work correctly, we consider the following:
- Level Progression
- What is the appropriate amount of power to give out at each level
- Are any cards too dominant at one level?
- Are there viable strategies that allow players to focus on winning at each level
- Do card progressions feel natural and make sense to players
- Are cards balanced appropriately by their level progression?
Algorithmically Generated Decks
SolForge Fusion is the world’s first Hybrid Deck Game. That means that each faction deck of cards is 100% unique. You customize your deck by combining any two different faction decks. This means that you can’t just copy some deck you see online, and you don’t need to spend your time chasing after rare cards hoping to keep up with your friends who buy more packs than you.
That’s the good news. The challenge is that in order to deliver on that promise, we need to build an algorithm that makes sure decks are reasonably balanced against each other and that there are enough synergies to make playing a combination of any two decks interesting. One of the key lessons we learned in this process is that cards can’t be judged in isolation, they can only be judged by how they operate within each faction deck.
Moreover, as Richard pointed out to us during one team brainstorm session, traditional concepts of rarity and faction identity don’t apply in a game like this. In a traditional TCG, if you introduce even one good card at a certain theme (e.g. a card that destroys creatures), then that faction immediately becomes about that theme, since players will just all collect and play that one good card. In a hybrid-deck game, however, a card that is good at a theme but rare will still have to be played in the context of all the other cards that come with it, forcing a player to make tradeoffs between gaining access to the desired card vs. other cards that they want in the faction. This forced a huge shift in our way of thinking and opens up the door to a lot more design space, including allowing a small percentage chance that ANY card could “jump factions” and appear in a deck that otherwise would never have access to it!
In addition, we had to find a way to make sure that cards that showed up in a deck had a purpose. There is nothing worse than finding a powerful Zombie Lord in your deck, with no zombie minions to command! So we engineered our algorithm to search for these cards and ensure that there were always synergies for players to find in each half deck. Beyond that, we made sure that each faction had at least some hooks that would play into another faction’s mechanics. So, for example, even though the undead Nekrium faction is home to the vast majority of Zombies, you can still hunt for a few in the other factions to help round out your Zombie deck.
In general, algorithmic deck creation forced us to add the following list of considerations
- Algorithmic Deck Generation
- Does each theme have enough connections to match with each other faction?
- Are there enough representative elements for each theme to ensure that decks don’t become too homogeneous
- Are all of our cards tagged correctly for our algorithm to pair them?
- How are rarities impacted by our building algorithm? Do we need to add more (or remove) elements of a theme to balance this?
- Are Algorithmically generated decks within an acceptable power band
- What themes are we willing to bleed into other factions and which will stay more tightly controlled?
Algorithmically Generated Cards
I’m pretty sure that at this point any sane development team would have stopped innovating. There is limitless content and tons of interesting design space just from the innovations of leveling cards and being the world’s first Hybrid Deck Game. But we did not stop there.
Richard and I are always excited about pushing the boundaries of what is possible, and we realized that digital printing doesn’t just let us create unique decks on demand, but it also lets us create unique cards on demand.
A traditional TCG will typically release cards in sets of a few hundred at a time, which is enough cards to change the game, but not so many that the developers can’t playtest them all. Using the algorithmically generated cards system for SolForge Fusion, our Set 1 alone has over 17,000 possible cards in it! Not only that, but each set will exponentially increase that number very quickly. There is no way our small team of 6 developers can possibly test them all, let alone understand what each card would be like in the limitless number of hybrid deck permutations. This forced us to innovate on how to design and balance these cards.
Understandability and the Learning Curve
For many players, the idea of 17,000+ possible cards is exciting. For others, it’s intimidating! And for good reason. The joy of playing trading card games in large part comes from exploring the space offered by a new batch of cards, gaining mastery and understanding until you feel like you’ve got a good handle on it, then going through the discovery curve again when new content comes to shake things up. But how do we make the concept of 17,000+ cards understandable? By using the power of Fusion.
Each algorithmically designed card is really a fusion of two card “splices” that represent half of each card’s functionality. Each half contributes some abilities to the card and in the case of creatures some of their attack and health. Each splice also contributes to the name of the card, with one splice forming the “noun” (e.g. Spectre) and the other splice forming the “adjective” (e.g. Necrotic) so when combined, they will create an Adjective-Noun name that will be consistent and unique to that card.
So players can talk about cards by name without confusion, and you can gain mastery over even the 17,000+ permutations simply by learning around 200 card splices.
Each deck will also come with at least a few traditionally designed “wholistic” cards to help players quickly find cards they recognize and understand even while they are still absorbing the breadth of splice designs.
Forging the Forgeborn
The most ambitious card type by far are the Forgeborn. Each faction deck comes with one Forgeborn who represents the leader of that Faction. Forgeborn start in play and gain powers every time you shuffle your deck (which happens every 3 turns). The Forgeborn are important characters in our lore, so we wanted them to have a strong identity and be mechanically cohesive. But since they are such an integral part of every game, we also didn’t want them to play out too similarly and felt this was the best place to take advantage of our digital printing technology.
To solve this challenge of wanting both consistency and variety, we decided to design each Forgeborn with 4 distinct abilities, each named and with a clear level progression from levels 2-4. When a Forgeborn is generated, it randomly selects one of those 4 powers at each tier, creating 12 possible permutations for each Forgeborn, but still allowing players to see the same sets of powers consistently to begin to understand what the Forgeborn is about and see them as a holistic character.
One of the more exciting things we are planning in the future is to have the Forgeborn morph over time, gaining and losing different abilities based on the events of the lore, which are in turn influenced by player choices at tournaments and organized play events. Because digital printing is faster than traditional printing, we can react to player actions more quickly than most card games. Increasing player investment in these main characters by making you feel like you are part of the story is a holy grail of game design and something we are hoping our digital printing technology lets us do better than any games in the past.
Balancing Algorithmically Generated Cards
So how do we balance 17,000 cards? More algorithms.
Since it is impossible to review all 17,000+ cards individually, we relied on additional algorithms to surface “problematic” potential combinations. This algorithm would flag cards where certain stats seemed out of line for the appropriate level (e.g. a Level 1 create with 10 attack). Once a card was flagged, our team could review it and decide if the aberration was acceptable or not.
Beyond algorithms, there is still no substitute for the bread and butter of card game development- playtesting, review, and instincts. Our lead developer Jason Zila was meticulous about logging all of our games in a shared spreadsheet, taking notes on things like game length, challenging cards, play patterns, etc. which we would review each week to highlight the appropriate areas of focus. We would rotate play experiences between modes where we had only a few decks to pick from to when we would have dozens to see what the differences in play and power level felt like.
Another strategy that was very helpful to us was to treat the game as though it was a traditional TCG for a few rounds of testing. We could look through the file and build our decks (still following the rules of our algorithm) to be the best possible decks we could imagine. Then we played those decks against each other and against more traditional algorithmically generated decks to ensure the experience was still fun and that there weren’t any degenerate combinations that we were able to exploit.
Finding the Fun
At the end of the day, our job is to create the most fun and engaging experience for our players, and the way we know we are doing a great job is how much fun we are having ourselves. When our weekly reviews turned into mostly moments of reliving our epic games, crazy combinations, and interesting deck combinations, we knew we were in a good place. The next piece of the puzzle is you.
Even for a traditional trading card game, there is no way a small team of developers can compete with thousands of players when it comes to testing a game and finding issues. This is even more true with a hybrid deck game.
The beauty of building a game with the Kickstarter community is that you get to give direct feedback that can still change the game and make it better. We experienced this last year with our Ascension Tactics campaign and we have already seen our community have a huge impact on SolForge Fusion in our discord channels and playing our free mod on Tabletop Simulator.
Please come join us when our Kickstarter launches on September 7th and be a part of the next chapter in SolForge Fusion!
Exclusive Game Design Insights
All my articles are first written in a physical journal and I only post about 10% of them on this site. If these articles are interesting to you, I recommend signing up for the newsletter.