It's the first of April, and we're back with another blog post!
One of our toughest jobs as devs is dealing with the "bad apples" of the community. We're talking players who AUTO n chill, players who pick Fiesta, or players whose opinions we simply don't like.
For years, some of you have speculated that we manipulate things like your matchmaking queue times, card drop luck, unit targeting, and how often you run into smurfs. That’s ridiculous. We don’t do that. AI does it for us.
We're proud to showcase our proprietary tech: Data-Oriented Game Enforcer, abbreviated as DOGE.
Principles
1. Get that juicy D (Data)
Every time you play, DOGE collects and analyzes everything. Your build order, your positioning, your roll luck, how quickly you hit “play again” after a win versus a loss—it’s all in the system. Think we don't know you started to type "fk the devs" but then erased it before hitting send? Think again.
2. Snoop around. Privacy is a fallacy.
As powerful as DOGE is, it can't peek into your Steam messages or eavesdrop on your Discord rants. For those, we rely on our eyes and ears in the community. Our dedicated staff sniff out bad actors, many before they even type a single message. Sometimes, just the seeing that "is typing..." flicker on and off before a wall of text is enough to know it's not gonna be good. Give them a mute, knock off a few likeability points, and move on.
3. Calculate LikeabilityScore™ and adjust game accordingly
Our proprietary GPT-4.5-based model runs 24/7, calculating a univariate weighted sum (LikeabilityScore™) that combines all your behavior data points into a number between 0 and 100. Roughly (the model is always evolving), this maps to something like:
- 100 = You rock. We will name a Campaign boss after you. The odds will always feel in your favor, until you decide to throw.
- 75-99 = You're a good person. Your game will almost never crash, reconnecting actually fully works properly, and cards will come with regularly.
- 50-74 = Most players land here. You probably roll Pota+Mask more than you should, but otherwise pretty vanilla
- 25-49 = You're on our watchlist. 60% of the time, your units will lose focus every time. You will tend to match against smurfs, and good luck finding a card. Your game is also more likely to crash/lag.
- 0-24 = We won’t even read your bug reports. All above de-optimizations and many more, like a tendency to match with teammates from a country you don't like.
Additionally, if your LikeabilityScore hits 0, you may receive a message prompting you to provide five bullet points on why you should still be allowed to play the game. Failure to respond could result in your game uninstalling itself.
Takeaways
Found yourself in an unfair match? It really has nothing to do with the amount of players in your rating bracket currently queuing. Rather, consider that your LikeabilityScore may have recently nosedived. Reflect upon your recent behavior, and consider being a better person instead of blaming a perfect system.
Got 700+ Pity Points? Wondering why your units just focused the full HP Siege Ram instead of the 10% HP wave? You know the answer now.
Closing Thoughts
DOGE is still learning, adapting, and evolving. By ~summer, we estimate it'll be able to dynamically tweak your connection quality, split aggro consistency, and even how often you match with teammates who will listen to your pings. Hopefully with this post, you now know why the game and its community are just so perfect.
Fellow game devs, we hope to license this tech to external parties starting in Q3. Reach out to learn more.
Congratulations on making it this far. Happy April Fools everyone!
- 2024 April Fools - World's first-ever game to support next-gen NVIDYA Visage™
- 2023 April Fools - We're rebranding as Aye-Aye-Attack Games!
- 2022 April Fools - 1v1 Event & Balance Changes
P.S. If you didn't zoom into the screenshot to find the URL, here's a direct link: Legion TD 2: DOGE Admin Panel. Check this Screenshot for a hint on what to do next.