Design Methodology in Avatar Warrior’s Journey: Learnings from Balancing Card Packs
In the previous article, I gave an overview of the considerations behind the game balance in Avatar Warrior’s Journey. In this piece, I’d delve further into how to design for blind-pack cards and the considerations behind the pricing.
Cards in AWJ are that of Gear pieces that the character can equip. Each piece of Gear adds a benefit to the player either influencing the rewards earned or makes it easier for the player to complete certain mission goals. Each card has a soft cap that can be increased by acquiring duplicates of the same card and the player will have to run distances with those cards to level up. There are 3 tiers of the same card from a very basic one to an Epic-classed one and these classes affect both their initial bonus and their hard level caps.
Cards are gained by either fulfilling mission objectives or by purchasing blind card packs and card packs come in 3 different flavors — each corresponding to the tiers of cards found in them.
The design for this changed quite a lot through production when we realized we didn’t have the capacity to soft-launch with the number of gear sets we wanted. Initially, the packs would have come with a chance of getting a better-tiered card but we decided against this for the benefit of better rewards clarity.
When designing the benefits of the cards, I classified Gear into several categories — gear that influenced the player’s entire run, gear that influenced only the parts of the run that contained certain features (for instance, only granting a benefit when riding a creature) and gear that influences goal completion.
With the other aspects of gameplay, it’s easier to get estimations of how much of a typical session would contain these features as well as what percentage of players would actually reach these features to get a gauge of how different gear would suit different play styles. There are gear that would gain a very skilled player good rewards during a good run and those that would give any player a reasonable reward increase. Slow and steady vs all or nothing.
Before looking into pricing card packs, Ilooked into the stat gains for each card and set the baseline advantage based on the metrics discussed in the previous article. These advantages are more or less equivalent in gain for an average player playing an average game.
We also wanted some cards to be harder to chase than others and hence put in place a rarity system implemented in design but not on the frontend. A rarity system serves a few purposes. I wanted players to be able to level up at least a card to a decent standard and with a uniformly-weighted card pack, it’s likely that this will take time. With rarity, I was also able to give certain rarer cards a slight advantage boost to the player if we wanted to or to give these cards a bit of a visual upgrade in the future. I wanted to future-proof the game’s design foundations.
With that, I put together a matrix that quantifies these and objectively measures the strength of each pack. Here’s an example pack.
The ‘Approximated Value’ column refers to the worth of the card when measured against the number of games needed to start profiting and the ‘’Difference from adjusted cost’ refers to the difference between the chosen card value and the average value of a card in the pack. I had initially used these fields for a first pass balance against our pack worth to ensure the deviations aren’t out of our expectations. I’ve left them out here for confidentiality reasons.
Once we started on playtests, the system’s flaws became apparent. One of the main takeaways was that with independent probability, there is always the chance that very unfortunate players won’t get the rare cards they want and this translates to their feeling of losing out. Games like Hearthstone reconcile this with a definite chance of at least one rare or higher card in each pack.
Another thing we had to do was to adjust player gains (through mission rewards) because while the system balances out eventually to the play pattern that was desired, early players were finding it hard to afford packs and this was frustrating. What we ended up doing was to map the number of card pack gains to the average player’s expected game time (and on a more micro-level, number of sessions) and adjusted our pack costs thus. I think it ended up being a more forgiving experience that retained players for a little longer.
To summarize, my main takeaways from the project are these:
DO
- Weights work best towards future-proofing content. However, percentages are still better for visualization so having both are useful in tracking your game balance.
- Being organized in data and having appropriate graphs correlating different systems in play saved us heaps of time trying to figure out where to start fixing awry balance.
- It’s almost always a less stressful work flow to adjust your pack costs upwards rather than downwards. People like getting a bargain.
- Be more generous with your starting players
DON’T
- Just think about adjusting for costs. Sometimes, a path of less resistance is to adjust minor aspects of gameplay to fix the issue.
- Plan for the best-case scenario. With our gear system, we had planned for a system where all our assets are available the manner we wanted it to. We ended up with bugs that killed some of the assets we planned to have in game. As a result, we had to compromise on some of the balancing aspects which still didn’t solve the main problems in a satisfactory manner.
That’s it. If you’ve tips on how to work with game balance or would like to share knowledge on your thought process, please feel free to drop a comment. I’d be very interested in a discussion and feedback on what I’ve written is always welcome — positive or negative!
