P3: Birds of a Feather – Team Komodo Dragons

Artist’s Statement:

Two hundred years from now, the world is flooded and humans are an endangered species. Rooftops stick out of the sea like islands. Storms tear flocks apart faster than they can regroup, and every year there are fewer places left to land and eat. Birds of a Feather (BOAF) is a deckbuilding roguelike about the strays still flying. You gather them into one flock and lead them to the far shore. 

Someone has to fly out front into the wind while the others tuck in behind and rest. Then a storm rolls in, and you get one turn to decide who gets sheltered, who gets moved and who takes the hit.

We want you to get attached. Each bird has a name and a personality, and the longer one flies with you the stronger it gets. When one doesn’t make it, it’s gone for good. 

Our target audience is people who enjoy roguelike games, like Slay the Spire. We also playtested with people outside it to see if someone who doesn’t play roguelikes could still pick the game up. As you play, you’ll unravel the story of this flooded world through your sticker book. It’s an embedded narrative: you piece it together from pages of a banding station’s log, kept for fifty years by the person who counted and ringed the birds. The pages turn up at big moments in a run, and each one comes with a few keepsakes you earn along the way, like a bird band or a door key. Some pages only appear on the harder routes, and the last one waits on the far shore of the hardest.

Staying close is what keeps a flock alive, and it’s also why losing one bird hurts so much. Each bird rides on the bird in front, so when you lose one, the birds behind it tire out and you lose another. 

 

Overview:

Birds of a Feather is a roguelike deckbuilding game where your deck is a flock. Each card belongs to a bird, and you want every bird to make it to the far shore. Some won’t.

Lightning is two squares from your lead bird and a hawk is eyeing the tired ones in the back. You’ve got three wingbeats. Two of your birds are about to take a hit and you get to pick which. Every bird has a name, its own cards and a stamina bar. Run out and it’s gone. Keep one alive long enough and it gets stronger.

Every flight is a puzzle on a grid. You spend wingbeats to move your birds and play their cards. A bird flying behind another bird pays less to fly, which is what real flocks do, so you want everyone close. Then lightning, hawks and power cables show up and break it up, and there’s usually an updraft off to the side that’s tempting enough to leave for. Lose the bird in front and the ones behind it get tired fast. The lead bird gives the whole flock a bonus, so you end up thinking about who goes up front.

Between flights you pick a path on the map. You can rest, shop or stop at a mystery island, like a fox den, a sunken bell or a kite festival. Some islands help you and some take something. At the end of each act the flock does a crossing over open water, and on the other side you take in a stray. The seeds you spend on resting are seeds you can’t spend on a card.

We want you to try different ways of playing. You can keep the flock in a closed pack and let everyone draft, or you can find other ways to get the most out of your birds. Send one off alone to ride updrafts. Take in every stray you find, or keep a few and look after them.

 

Final Concept Map:

Our concept map for the various game parts coming together as a system.

 

Design Inception and Iteration:

We were excited about modeling the system of bird migration from our very first brainstorming sessions – but wanted to know more about the system as a whole to represent it meaningfully. We started with a deep dive into the intricacies of bird migration, hoping to identify the many parts of this system and note how they interacted. With so many moving parts, we found it difficult to easily model the system – especially in a gamifyable way. We took inspiration from Louis Lafair’s Mapmaker: The Gerrymandering Game, and we were able to condense our thoughts into a more concise set of system components by realizing we only needed to teach the core concepts of bird migration, not each intricate detail thereof. We wanted to represent that during their long migrations, birds flew in a flock to conserve resources and protect themselves. We wanted to represent that birds faced various dangers along the way and had to learn new skills to overcome these obstacles. And as importantly, we wanted managing these resources and strategies to be fun! We wanted the emergence of strategy to be a major source of fun (challenge), and we wanted our narrative to be engaging enough to continue for more (fantasy).

Our initial board ideation.

Our first thoughts were to make a 3D-flight simulator, like Starfox, where the player continuously controlled the path of the flock. While making the first playtestable iteration, however, we realized that a turn-based strategy game with movement confined to a grid would better unite our system elements into showing the system as a whole. Having the player discretely move the birds of their flock, and take actions that each member learned, would be more helpful mechanics in representing our system of choice. Stamina would serve as our fundamental resource to manage (a stand-in for health), and hazards would be dealt with through a combination of grid movement and actions. We took inspiration from Into the Breach and Slay the Spire for the overall integration of these elements into a grid-bound, turn-based strategy game. With our basic game structure worked out, we moved to testing early playtesting with a paper prototype.

 

Phase 1: Paper Prototype and Early Digital Prototype (11/4)

For our first version of Birds of a Feather, we had the player lead a flock of five birds across six flights in an effort to get as many home as possible. Each bird sat on a grid, and hazards to the flock were determined via a dice roll, showing whether the sky was calm, stormy or hiding a predator. With these limited mechanics, we wanted to see what strategies players would develop independently.

Emma and Olu, 11/4

Emma and Olu were our first playtesters, and they played together in class as one player. Both players are Stanford LDT master’s students. Olu plays primarily sports simulation games (FIFA), and Emma plays narrative-rich games – so neither player is an explicit member of our target audience.

Without intervention, they placed the birds with the most stamina at the front to protect the rest of the flock. To make positioning more strategic, they suggested limiting movement to one space per turn. While playing, they also wanted to design their own formations, have more reasons to swap birds around, and struggled to understand why each bird’s stats mattered. At one point Olu asked, “Why wouldn’t you do a V shape?” We didn’t have a good answer yet; it was the dominant strategy once you found it, and there was little incentive to break it. While we liked that there was incentive to be in a formation, we also wanted this formation to be dynamic to the challenges the flock faced.

Our paper prototype.

Lucas, 11/4

Later in the evening, we finished up the digital version of our basic playable prototype, so we continued testing with that. We had Lucas, one of our class TAs, as the first playtester. He has lots of experience playing Slay the Spire and other roguelite deckbuilders, and is thus part of our target audience.

In our first digital prototype, birds could only move by playing cards – which both allowed for movement and also provided additional actions/abilities to help the flock. We found, however, that some cards dominated over others (i.e., Slipstream, which lets a bird skip its flight cost), and the playtester nearly exclusively used this card. Despite the emergence of a dominant strategy, he still liked the prototype and said it already felt like a game! He also thought it was “too noisy.” His biggest suggestion was to make moving a basic action instead of something you need a card for.

 

Phase 2: Building and Balancing Our Systems (11/5 – 11/10)

Following our playtesters’ feedback, we gave every bird the ability to move and placed limitations on how many times a card can be used (i.e., Slipstream could now be used only once per flight). To limit the number of actions (either movement or cards) a player can make on a single turn, we gave players four energy per turn. We also moved weather forecasts onto the grid, so that players can develop strategy in advance. Besides balance, our biggest goal was to make players care about their individual birds, so we added species synergies, skill growth paths, and lessons that teach birds new cards.  limited recruitment to the end of each act because we wanted players to spend time with the same birds instead of constantly replacing them.

We noted, however, that while some mechanics synergistically worked to improve our game, others worked together to break our game. Drafting and limited recruiting together caused a negative feedback loop that often ended in a loss. Losing one bird made it much easier to lose the next. Once a bird was gone, the birds behind it used their stamina more quickly and went down too. With no way to replace these birds until the end of the act, most runs were irrecoverable when a player lost their main bird. The fewer birds you had after this loss, the harder it was to keep the rest alive. To soften this effect, we made flying cheaper for a number of turns after a loss, added regional weather patterns, and added seeds as a currency for recovering stamina. Losing a bird still hurt, but it didn’t end the run as often.

Paprika, 11/9

Paprika, another fellow CS377G student, playtested our next iteration. She enjoys Zelda, simulation games, and choice-rich games. She also has some experience with roguelike mechanics, making her somewhat part of our target audience.

Paprika liked that the game had real consequences, notably reacting with “What happened! *gasp!*.” However, she struggled to tell apart the initial setting up of formation from the later taking of a turn. Weather effects and predators were confusing, and she suggested bringing more attention to the “Undo” feature so players could feel more comfortable experimenting.

After her playtest, we played through the game ourselves and realized something more concerning: our game wasn’t that fun to play. Drafting greatly reduced stamina cost and the forecast showed every hit coming – so most flights could be completed by simply keeping the birds in formation. Emma and Olu had pointed this out during our first playtest, but we had spent so much time adding mechanics that we never truly balanced this. The only significant thing players could do during a flight was avoid a hazard, so it often felt like there was little choice.

Paprika playtesting our early iteration.

 

Phase 3: Onboarding, Readability, and Replayability (11/11 – 11/13)

To give players a reason to break their basic formations, we expanded our grid, and added more hazards and rewards to it. The board was expanded from 5×4 to 7×5 and now introduced  both thermals that rewarded players for leaving safe formations, and terrain obstacles that forced movement. Staying in formation was still useful, but it now wasn’t always the best decision.

Sabrina, Jeffery, Aiden, and Yuina, 11/11

We were able to get several playtests in class. Sabrina and Jeffery are both part of our target audience since they have lots of experience playing deckbuilders and other roguelites, such as Slay the Spire. Aiden and Yuina are outside of our target audience, but they are interested in games providing strategy and some narrative, like ours. Aiden enjoys games involving history and science (especially birds!), while Yuina enjoys Minecraft Dungeons, Overcooked, and Animal Crossing. Sabrina is a junior, Jeffery is a master’s student, and Yuina is a CS master’s student with a lot of coding experience.

Through play, the players picked up the game’s core ideas and mechanics. They figured out that keeping birds in formation and conserving resources mattered. Onboarding was still a problem, however, as this process proved confusing for the playtesters at first. Even experienced deckbuilding players struggled. Sabrina noted, “I don’t know how I am getting this action,” for example, and wished she could see which cards belonged to which birds. Jeffery wanted to drag birds instead of clicking to form his formations, and couldn’t discern whether a skill belonged to an individual bird or the entire flock, wondering “if it is a bird skill or flock skill.” Despite lots of this information being a part of the tutorial, few of the playtesters read the large information boxes provided, and Yuina couldn’t understand what happened during a mystery event because the text was too small. Aiden asked, “Why did I die? This game is unfair”, when he eventually lost his flock.

Our group playtest.

What worried us most was that even people in our target audience struggled to learn the game. Their roguelike experience helped somewhat, but didn’t make up for the rough onboarding experience. We were trying to teach too much at once, and left them to fend for themselves too early. Players were losing birds before they understood how to keep them alive.

 

Phase 4: Making the Game More Engaging and Accessible (11/14 – 11/16)

Based on our player feedback, we added some features that seemed to improve information design and ease of interaction: birds could now be dragged around to set up formations, and the cards were redesigned to better showcase which bird they belonged to. We designed a new tutorial on paper, focusing on teaching the players mechanics through doing rather than reading. Players now start their play without cards and first learn about positioning and stamina, before they then face growing hazards (weather), where we introduce the cards to help solve. We stopped trying to teach mechanics like skills and growth paths before players needed them, and aimed to have those be introduced later in the game, when players already have a good understanding of the core mechanics.

The flight screen was completely redesigned to improve information design (i.e., more concise information and better placement) and to feature a more charming aesthetic to help with the narrative. To similarly create the engaging narrative we wanted from the beginning, we developed a “sticker book” to look through during the game, that would be populated as you found new entries during gameplay. We introduced a collectible logbook written by someone at a bird banding station, telling the story of how the world gradually flooded over fifty years. To scale with different audience’s abilities (based on seeing how different playtesters had widely differing skills), we added four difficulty levels: Normal, Hard, Extreme, and Impossible, so players can pick a challenge that best matches their skill. Each difficulty tier has its own map, weather and logbook pages, with Impossible being designed to beat even experienced players most of the time. For P3, though, we only focused on balancing Normal, and we plan to test the other difficulties more thoroughly later on.

Lucas, 11/14

Lucas, the TA, once more playtested Birds of a Feather for about an hour. Of the many additions we made to improve the game’s feel, surprisingly it was being able to drag the birds around that made a major difference. Once he found this new feature, he started making formations just for fun, which was our favorite moment. After completing his first run, he said, “Everyone made it. This is cool.” He also liked the newly redesigned, Paper Mario-esque flight screen and the storm chasing the flock, noting, “I like that it makes sense. And it’s unique.” He did note that “the battles get a little repetitive,” but the mystery events were what broke them up for him.

The cards were the bigger problem, however. Most of the cards did small and similar things, so playing them felt indistinguishable: “Now I’m having the problem where every card is like the same.” He also described a card that prevents birds from being pushed as “super useless because I’ve never encountered anything that pushes my birds.”, despite there being some hazards this mitigates. Additionally, we still had some recurring older problems. He noted, “It says it costs stamina, but I don’t know what stamina is.” He suggested explaining keywords through hover interactions, like in Slay the Spire. Though we already had some of these, he didn’t realize they existed – so we could do a better job showing these exist in our onboarding. He also said, “Oh wow, that’s a lot of UI,” and wanted anything unrelated to his current decision to fade out. Information design is still a major factor we need to improve on.

Krystal, 11/16

Krystal is a master’s student and one of our class TAs. She playtested our 11/16 build and pointed out that our pixel font made the game difficult to read. Although it fit our art style, it made longer card descriptions harder to follow. About the font, Krystal said, “It was giving me a stroke.” As consistent with prior feedback, information design can be improved to help playability.

 

Phase 5: Improving the First-Time Player Experience (11/16 – 11/19)

As per Lucas’s feedback, we needed cards that worked in synergy and created strategy. We started designing certain advantageous combinations – such as a card that restores stamina whenever a bird moves, paired with a card that can move the entire flock as a whole. We wanted players to get excited about finding new cards that worked synergistically with what they already have in their deck. We wanted players to develop emergent strategy and plan for the future.

We also needed to clean up the UI and improve information design, especially when the player has to make decisions. After much deliberation, we decided to cut the storm that chased the flock, due to confusion and problems it consistently brought to our playtesters. We originally had the mechanic of the slowest bird setting the pace for the whole flock, which only mattered for staying ahead of the storm, but few playtesters remembered or planned for this rule. Despite it seeming like a big change to cut it at first, we saw how little it cost us to rebalance, and how much it smoothened the flow of the game. Instead of having a potential detriment, we decided to make the leading bird now carry a bonus that affects the whole flock, bringing additional strategy. Additionally, after Krystal’s playtest, we swapped out the font to improve legibility. We switched our body text to Atkinson Hyperlegible Next and kept the pixel font for titles.

Sabrina, Xingyi, and Olu, 11/16

For our in-class playtest, we wanted to see how our changes improved playability with players both inside and outside our target audience. Sabrina has extensive experience with deckbuilders and roguelites, while Xingyi doesn’t play video games, and Olu, one of our first playtesters, mostly plays sports simulation games like FIFA.

Sabrina was noticeably excited about our roguelike elements, reacting with “ooo relics goated!” and “elite elite elite!”, both similarities to a game she enjoys: Slay the Spire. She also described the game as “perfect for mobile,” which inspired us to later look into this direction! She initially kept trying to drag cards directly onto birds – potentially another avenue for streamlining player experience. Heavy birds, and their additional movement cost, caught her off guard as well. “F*** he is heavy,” she said, which told us even experienced players needed a better explanation of certain concepts during onboarding. She described the game as strategy and stamina maxing, but said it was hard to maneuver around, and she wished she could stack Brace.

While experienced players could dive in comfortably, first-time players had a much harder time. Xingyi struggled throughout the first crossing and eventually lost every bird. When the fog went away and the run ended, she said, “oh, what!” She barely used cards, and similarly to Sabrina she didn’t understand why heavy birds cost more energy to move – nor did she feel she had enough energy to experiment. She also thought she had to choose between seeds and cards as a reward instead of collecting both, and thus sometimes chose not to take them. Olu skipped the tutorial halfway through because it felt repetitive.

Our newer playtest with better drag-and-drop mechanics.

Following this playtest, we implemented drag-and-drop mechanics for both cards and birds, replaced hovering with tapping, and increased text sizes. We wanted these to improve legibility and streamline the play experience – especially on smaller devices. To help with newer players and onboarding the game’s many concepts, we made the early game less punishing by capping incoming damage, extending flights so difficulty builds gradually, and adding more ways to recover stamina. We made energy yellow to help tell it apart from stamina, and later renamed it wingbeats because people were still mixing the two up.

Our tutorial originally started with eleven short flights – but since Olu skipped it and Xingyi found the onboarding overwhelming, we cut it to eight, more essential lessons. Additional concepts, such as low stamina, predators, and heavy birds are taught with a tip the first time they come up in a real run, so players don’t have to remember everything from the tutorial.

Players still struggled with managing their stamina, however, and did not know how to regain their lost stamina for the run. Over the earlier rounds we had originally added small catch-up rules to keep the early game from being too harsh, such as birds getting stamina back after every flight and flying getting cheaper after a loss. While these worked in the moment, the players could not see these aids and did not learn to play without them. We cut these hidden aids, and added a more direct, play-controlled healing mechanic. We made it such that seeds could also be used to heal birds post-flight, rather than just being used as a currency. While cutting these hidden healing mechanics made the game much harder, we rebalanced with seeds, and better scaling of elite flights and weather hazards.

In our own playthroughs, we agreed that there was too much information and too much happening at once during crucial decision-making moments, which detracted from creating strategy and noticing synergistic combinations. We ultimately cut down on the number and depth of combinations that could be made, as the myriad of different choices and stats proved too difficult to keep track of.

Nicole and Minh, 11/18

Nicole has played Minecraft and Battle Cats, and Minh has only played Minecraft, so they’re both outside our target audience.

In this version, both of them played to get as many birds home as possible, and they spent a lot of their time learning what the different bird types do. It was evident that they got attached to their birds – and when some birds died, Nicole said, “omg noo! The birds!” Minh’s biggest complaint was that the UI didn’t pop out enough. That was a little confusing for us, because earlier players said our UI was too busy. So we still need to find a balance where the UI tells you enough without crowding the screen.

A playtest of our version on mobile!

Bruce, 11/18

Bruce is a UC Berkeley sophomore with lots of experience playing MMOs and FPS games. Though he hasn’t played roguelikes, he’s playthroughs of Slay the Spire. We consider him outside our target audience, but someone we’d like our game to be enjoyable to.

In his initial playthrough, he spent much time in a tight formation and swapped his birds in and out. He noted, “it’s pretty hard to move the whole flock around unless you have the cards that can auto-form a formation.” As the difficulty ramped, he began to lose birds from his flock. At one point he said “they’re all dead,” and he told us “the buildings knocked out like everyone!” (note, his whole flock did not all die). He took in four strays along the way during this setback and was able to keep going. By the end of his playthrough, all three of his starting birds were gone and only one bird was left: Flit the barn swallow, one of the strays. With only one bird, Bruce said he “abused the heck outta the free move 3 squares and -2 leading penalty to hop around to the updrafts to keep alive.” In a way, that’s what we wanted thermals to do: give a bird a reason to leave the formation – so it was nice to see an alternative and more dynamic play strategy emerge.

In the end, Flit made it to the far shore – and because a barn swallow got home, the run unlocked the barn swallow for next time. Though Bruce doesn’t play roguelikes, he still finished the run in its entirety and said “it was fun tho.”

His run is a good example of the unexpected consequence in our game. Players get surprised when they lose a bird, but eventually find out that losing birds is nearly unavoidable. We designed the game to be so, and teach these realisms of migration. A few phases ago, losing even one bird usually ended a run, but Bruce lost six of his seven birds and still got home. That’s close to what real migration is like: plenty of birds don’t make it, and the flock keeps going.

Runs like his, with one bird left, however, had a problem we needed to fix: a lone survivor used to get only one card to pick from after a flight. Rewards now always offer three choices.

He also ran into a small bug: after he lost Honk and recruited another goose, the tip popup still said Honk – so we made sure to fix that for consistency!

Final Playtest – Rusdy, 11/19

Rusdy is a childhood friend of Noe’s and has played lots of roguelites, so he is inside our target audience. For our final playtest, he played a full runthrough, taking about two hours of play to get from the first flight to the far shore. At the end he noted “That was fun!” and didn’t think there were any particular moments in the game that felt bad. It was hard, but he noted that he was fine with that.

In his smaller feedback, however, he said “I honestly probably don’t want that many birds. It’s a lot easier to manage with just a few,” which is the opposite of how Bruce played. Halfway through his playthrough, he was sitting on 37 seeds halfway – and despite how he failed every island gamble, he felt seeds didn’t feel scarce to him.

This playtest also turned up a few bugs. Twice, Rusdy lost his progress. Early on he said his run restarted, and about an hour and a half in he said “I lost everything,” which is a pretty terrible bug. We investigated and found two causes in the code. Firstly, opening the game wiped the saved run, so reloading the web page did exactly this, and starting the tutorial again resets the save. He also found a card pick where “I can’t pick any of them.” – and we found it was because the flock was full. He then went looking in Settings for a save and found none, so we added a Save and Quit button to the pause menu.

He additionally gave lots of helpful feedback on further information design and points where he had confusion. He asked “what shape does it use?” about the W-formation card, and he wanted to see what a card does before picking it. He asked who the “healer” was, and why a bird’s stamina trait stopped working during Act III (it only works while drafting, but nothing showed that). We added V and W to the card keywords to help, we grey out bonuses whose conditions aren’t met and say why in a few words, and the “healer” line now names the bird.

 

Final Video

Important Timestamps:

4:39 – Here the player already notes the info boxes to hover over, and is doing so to get more information.

7:02 – The player checks Pole’s flock bonus unprompted, after learning how to check Elder’s flock bonus. It was nice to see the interest and how this might develop strategy!

13:31 – Checks Honk’s bonus, and puts him in front instead since it’s more helpful.

18:20 – The player hovers over the “arrow” definition to see what an arrow is, to know what the anti-arrow card does.

21:15 – Fun playing with seeds! 😀

21:47 – The player looks at flock for more info on stamina before choosing what best option is (rest, recruit, train).

29:03 – More engaged with dialogue and narrative. The player begins to speak more for the whole run after this, which was very rewarding to see!

29:20 – Evaluates new bird’s abilities to see if worthwhile.

35:00 – Finally setting up strategic turns, “Things are starting to make sense!”

40:30 – “It would have been better to place him at the top” –  player notes deficits in current strategy and how to improve for next time.

41:40 – Player checks deck to see probability of getting next cards, to plan next moves. More advanced strategy!

52:21 – Player notes “end of the world” theme after reading log.

53:13 – “I should try and conserve as much stamina as possible”, planning for future turns.

54:13 – “What does the lightning do?” – doesn’t know, and checks the lightning icon immediately. Successful onboarding!

58:35 –  “These seeds I’m not going to be using anytime soon. There’s a market way up there”, notes the purpose of seeds, and gauges when they will be useful by looking at map, and uses icon to discern where the market is.

1:02:21 – “What if I use it here?”, uses card synergistically with flock formation, i.e. in centerpoint of flock.

1:13:23 – Player realizes one bird is a healer, and finds out by checking bird stats.

1:26:12 – “Seasoned bird” rotisserie chicken joke :3

1:28:02 – Actually splitting up flock, instead of staying in formation, because more advantageous to avoid hazards. Nice to see the formation finally being broken up!

1:30:12 – Sees what V and W formation cards can do, and notes that it will be very useful to recover from situations in the future.

1:33:39 – “Perfect time to use the V!”

1:52:21 – “Probably don’t want that many birds. Easier to manage with just a few.” Developing strategy with existing birds, and realizes it’s not always advantageous to get more

2:19:36 – Notes didn’t have to spend stamina, but doesn’t have the card (just bought) that carries stamina over to next turn. Already planning strategy with new card bought, even though haven’t used it yet.

2:31:51 – Victory 😀

2:32:35 – “That was fun!”

2:34:20 – Player notes getting cards that focused on movement and stamina. Forming strategy.

Deliverable:

And finally, here is the game!

😀

AI Disclosure:

We used Claude to help with coding and for much of our asset generation, though we used references to help generate these images, and did much hand-polishing afterwards.

About the author

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.