How Final Fantasy XIV’s Toughest Battles Are Designed: A Deep Dive On Savage Raids

Final Fantasy XIV is a game of many facets. I’ll always remember it for its emotional storytelling, lovable characters, and how its dungeons and bosses blend gameplay and narrative beautifully–all elevated by an evocative soundtrack. The lighthearted side content like the Manderville quests or the Gold Saucer add levity and embrace the silly side of the MMORPG, and optional boss fights offer great side stories alongside hearty combat trials. But then there are Savage raids, the toughest and most grueling content in the game, fine-tuned for the sickos who love a ridiculous challenge.

Savage raids take the normal raid bosses and turn up the difficulty and complexity of their gameplay mechanics to unforgiving levels. It’s a sort of dance routine with very little room for error for all eight party members–a combat puzzle where precise execution of specific strategies and quick reaction while maintaining your attack rotation are required. And failure means starting over from the beginning. Having been blessed with a smart and patient raid leader, I’ve been able to avoid the frustrations that often come with raiding at this high of a level, and that’s helped me appreciate the nuances of Savage raiding.

With a much deeper understanding of FFXIV from a technical perspective, I was able to speak with FFXIV’s lead battle content designer Masaki Nakagawa (better known as Mr. Ozma in the FFXIV community) about how raids get made. We talked about everything from the development timeline for squeezing the content into a strict release schedule to staying creative and reasonably challenging over a decade’s worth of raid tiers. We also covered the intricacies of working within the confines of the game’s limitations, how the team made the trickiest raid boss yet in Patch 6.4, and the influence of player feedback.

The insights are fascinating for those who’ve touched raids in FFXIV, but even if you haven’t, this deep dive into FFXIV’s design paints a much more complex picture for one of the many pillars that make the critically acclaimed MMORPG a special experience.

GameSpot: What does the development timeline look like when you have to design raid bosses? For example, when do you start creating them for upcoming patches, when do testers get their hands on them, and how much time do you get for adjustments between normal and Savage versions?

Masaki Nakagawa: The early stages of developing a raid involve tasks such as deciding the design concept, as well as creating artwork for the boss and scenery. The timeline for these tasks vary depending on the situation, but it typically starts between about six months to a year before release. Since eight-player raids are usually released during even-numbered patches [6.0, 6.2, and 6.4 for example], it’s safe to say that when a raid is released, we’re most likely already working on the next raid tier.

As for playtesting, we adjust both the normal and Savage versions in tandem, rather than wait until the normal version is complete. This is because changing something in Savage might affect the normal version, and vice versa. After the content is built, we have about four to eight weeks to repeatedly test and tweak to make further improvements.

What are some of the biggest lessons you’ve learned in high-level raid design for expansions and patches through the years in terms of balancing difficulty, player enjoyment, and creative ideas?

I’ve been a part of FFXIV’s battle content development for over 10 years now, and I’ve honestly learned a lot. There’s also a lot I’ve gleaned from our players through their feedback and livestreams, which I’m truly grateful for. When we look back at our first high-end raids, the Coils of Bahamut [compared] with the latest raids, there are many aspects where we’ve improved. That includes making adjustments to difficulty, taking mitigation and damage buff timing into consideration, balancing melee attack range, game design aspects like time limits and hints for mechanics, creating lore or scenario-based depictions in an encounter, our specifications for phase changes and loops, and so much more. There’s so much to cover that I could spend an entire day talking about them!

A specific example would be how “simple mechanics with heavy emphasis on reaction time” are used very differently from “mechanics requiring several decisions at once with lenient time to react.” Mechanics with heavy emphasis on reaction time may make some players feel like they’re not making progress no matter how much they practice, whereas other players might pick it up right away, which bears resemblance to action games. But with a mechanic that requires multiple decisions at once, if players are given plenty of time to think about them, with practice, they’ll feel like they’re getting further, and the majority will eventually learn how to solve the mechanic perfectly. Currently, high-end raids in FFXIV are built around mechanics which fall under the latter category. That said, that doesn’t mean we completely avoid mechanics that emphasize reaction time; we incorporate them in moderation to strike a balance.

There’s a continual need for new MMORPG battle content that surprises our players and keeps them from getting bored. However, creating something new always involves risk and lots of development resources. Conversely, battle content with your typical mechanics have the advantage of providing a stable level of entertainment with little risk and cost. When I lead my team, I provide direction based on maintaining the overall balance between fresh and typical content. For example, in Pandaemonium: Anabaseios, the Ninth and Eleventh Circles were built around familiar mechanics, while we challenged ourselves to introduce new ideas in the Tenth and Twelfth Circles. I believe we gained this sense of balance from all that we’ve learned during our many years of operations, as well as the various feedback we receive from our players.

What are some raid mechanics you wanted to do but couldn’t because of design limitations (either from a technical game engine perspective or having to adhere to things like arena shape, two-minute burst windows/downtime, etc.)?

The development team always creates battle content under numerous limitations, and the number of ideas we’ve waived might rival the number of stars in the sky! To give a specific example, dynamically moving the arena is an idea that was raised multiple times by different battle content designers. These were ideas that you might see in action games, such as swinging the floor back and forth like a pendulum, fighting on the back of a flying dragon, or battling atop two moving ships. These sort of ideas involving dynamic movements of collision would normally be unfeasible under FFXIV’s systems, but we cleverly worked with our given limitations to create arenas like Leviathan’s rocking boat, Diamond Weapon’s dual flying platforms, and Byregot’s hammering the floor into shape [from the Aglaia alliance raid].

The 120-second burst windows and the limitations of player uptime on bosses are aspects where we received lots of player feedback, and have undergone lots of trial and error. We tend to run into issues with encounters that temporarily prevent players from targeting the boss, so we’ve significantly reduced the frequency of bosses becoming untargetable. Furthermore, we’ve reduced the frequency of what our team calls “sports meet phases,” where the boss becomes untargetable while players run around the arena to resolve mechanics. We received lots of feedback regarding them, primarily from our players outside of Japan. I have the impression that more players enjoyed resolving mechanics while attacking the boss, and I feel the same way when I play the raids myself, so a lot of bosses are designed for maintaining uptime nowadays. That said, certain mechanics can only be expressed through these “sports meet phases,” so I’d like to keep an eye on striking a good balance to incorporate both designs.

Some consider P10S [Anabaseios: The Tenth Circle (Savage)] the toughest instance in the Pandaemonium series, and it’s quite different from the rest of the raids. What was your approach to getting creative with it?

For the Tenth Circle (Savage), our goal was to create something that would vividly remain in the minds of all players. A big reason for that was because the namesake of the raid series, Pandaemonium, was going to appear as a boss. To achieve our goal, we took a two-pronged approach. First, we had an idea to incorporate spider web-like mechanics. I felt that by doing so, we already had grounds to create unconventional mechanics from our initial design stage. The second concept was to make the encounter more challenging than previous second-turn raids to give our players a strong sense of accomplishment. I wanted to design a powerful encounter befitting of the name “Pandaemonium.”

After its release, I saw many of our players on social media and livestreams comment that it was their favorite second-turn boss. I even saw some people praising it as a “masterpiece” in their videos and blog articles, which made me truly happy. On the flip side, others hated how it was harder than any other second-turn raid. While I believe that, in many ways, we achieved the goals we set when we started working on the tier, the resulting difficulty was a bit harder than we anticipated. We certainly set out to make this second-turn more challenging than previous ones, but considering the players attempting the raid with strangers in Party Finder, the damage of Harrowing Hell [one of the boss’s late-stage attacks] was a bit too harsh. The development team’s estimations for Harrowing Hell did not use Tank LB3 [Limit Break 3], but the damage was too high and it became the norm for parties to use Tank LB3 there, which prevented them from using healer LB3 to recover from any later mistakes. I believe this was one of the biggest reasons why the Tenth Circle (Savage) felt too difficult.

The encounter had both great points as well as room for improvement, but of all the raids I’ve worked on, I feel we succeeded in creating something quite interesting. I will be reflecting on the areas that I can improve to leverage them in the next opportunity and create even more interesting bosses in the future.

What’s your favorite thing about watching the community take on new Savage raids when they come out, and have you been surprised by strategies that you didn’t think of?

After the raids are released, I like watching livestreams of players enjoying themselves and figuring out their way to a clear. Our players are amazingly skilled, and I’m always surprised at how quickly they figure out mechanics and get through encounters. It’s especially interesting to see videos of players who restrict themselves even further, despite how difficult high-end content can already be. It’s always a surprise to see players find clever ways to clear with different party compositions like having only one healer, all tanks, or only seven players.

Having designed so many high-level bosses, what can we expect in the future to keep challenging us in new ways?

I’d like to thank everyone who enjoys FFXIV’s high-end raids. In Endwalker, we released the three raid tiers of Pandaemonium, and two Ultimate fights–Dragonsong’s Reprise and The Omega Protocol–and we received so many comments from our players via social media. I’ve also had the pleasure of watching many of you stream your raid progress in real-time. There’s been times where I succeeded, and other times where I fell short, but all of you have taught me so much with each experience. I’ll continue giving my all to deliver the feelings of excitement and achievement that can only be experienced through high-end raids, so I hope you’ll be looking forward to more raids to come!

FFXIV is currently available on PC, PlayStation 4, and PlayStation 5, and is coming to Xbox Series X|S in Spring 2024. The upcoming expansion, Dawntrail, starts a new era for FFXIV’s story following the conclusion of Endwalker and the Patch 6.x series, and is set to launch sometime in Summer 2024. FFXIV has a free trial that includes all of the base game A Realm Reborn and the award-winning expansion Heavensward, and lets you play up to level 60 with no restrictions on playtime.