Hello, I wanted to take a moment to share a bit of my craft in QA on TFT (and hopefully educate a little along the way). This is genuinely one of the tougher parts of game dev, and it’s something I’ve wanted to talk about for a really long time. It just never felt like the right moment. But now that our move to Unreal Engine is finally public, with Set 18 live, I think it’s about time the community got a little insight from me as a QA Lead on TFT. I’m going to start with an example from TFT K.O. Coliseum (Set 15). I actually wrote a version of this a long time ago and felt it was best to wait. I was worried people might not understand, or not care, or that it’d be met with cruelty. The timing never felt right and a big reason was that our Unreal work simply wasn’t public yet, so much of the story I wanted to tell was tied to it. But I feel really passionate about TFT and I’ve wanted to share a side of this work that doesn’t get talked about often. Why start with Set 15? Because that’s when we saw a noticeable shift in player perception of quality. ___ Every TFT set brings its own challenges depending on its scope, mechanic complexity, and where we are in development across multiple future sets at once. From a player perspective Set 15 felt buggier than past sets, but the data tells a more nuanced story. Internally, Set 15 launched with a larger than usual backlog of non-critical issues (roughly double what we’d consider a healthy starting point). These weren’t game-breaking or exploit-level bugs. They were a high number of smaller, low to medium priority issues that compounded over time and contributed to a “buggier feel” overall. Think of it as death by a thousand papercuts. Each small issue adds friction, and collectively it shapes perception. At the same time, our critical and exploitable bug rates were actually among the lowest we’ve had in years. The majority of major issues were caught pre-launch, and our escape rate (the bugs players find that we didn’t catch before release) was significantly lower than previous sets. So from a raw quality standpoint the game was safer and more stable than it might have felt. Where we fell short was live support capacity. Our live team was much smaller than normal while we spun up other high priority work across the team (the TFT Unreal Engine migration). Fewer engineers, designers, and QA dedicated to live meant we couldn’t burn down the backlog as quickly after launch, and the fix window stretched much longer than any of us wanted. Another factor was the complexity of Set 15’s mechanic, Power Ups. That feature alone accounted for over a third of our total bugs. Not because it was poorly made, but because its scripting touched nearly every system in the game (even minor adjustments could have ripple effects). And since Power Ups were the mechanic, players naturally noticed any oddities with them more than they might with smaller content pieces. To put it simply, we had more total bugs at launch, fewer resources for live fixes, and a highly complex mechanic amplifying the visibility of issues. Player perception dipped even though the underlying severity and escape data actually improved. So while the set’s public perception reflected real player pain, it wasn’t a sign of declining quality. It was a reminder that quality is as much about how quickly we can respond and recover as it is about how clean the build is on day one. ___ 1/2 🧵