Looking Back on Sentinels of Light

As we reflect on 2021, we want to talk about how learnings from Sentinels will help future events.

Hey all. As the year winds down and we reflect back on 2021, we wanted to talk about what we learned from the Sentinels of Light event this past summer. We want to own up to the fact that Sentinels missed the mark, and share how we'll use our learnings to help future big events.

One EXTREMELY important note to keep in mind: We're focusing on LEAGUE PC's involvement, not Riot overall.

This is a long one, so here's a summary of our takeaways:

  • Event metagames need to be accessible to regular players, not just super engagers.
    • Pace big events so regular engagers feel like they'll be able to get through the main portion, with big grinds saved as optional challenges for hardcore folks
    • Streamline metagame design so players understand how to access everything
    • Ensure canon lore in particular is accessible without over-the-top gameplay grinding, and provide ways to access the lore outside of the metagame
  • Narrative and character tone need to match player expectations.
    • Keep simultaneous forms of storytelling consistent
    • Only incorporate faceless player stand-in characters when it makes sense, rather than by default
    • Don't bend champion characterization
  • In-client visual novels have scope limitations we need to respect
    • Use them to add color and detail to our biggest lore beats or provide specific windows into those stories, but not to deliver massive stories in their entirety

Ok, let's get into it.

The idea

Sentinels of Light was the culmination of League's first multi-month narrative event, beginning with Ruination at Season Start. This overall arc was our biggest expression of core Runeterra lore in League in ages. We tied together multiple new champion launches (Viego, Gwen, Akshan; originally Vex before her delay) and released skins depicting champions as they appeared during the events of the story.

We also wanted to push further on the success of Spirit Blossom's in-client visual novel and character interactions, using them as a way to let players interact with the Sentinels during the fight against Viego and add immersion to the standard event progression experience.

Learnings

Event progression rate

With Spirit Blossom in 2020, super high engagers completed all of Spirit Bonds in the first two or three days and then spent the rest of the event grinding Spirit Petals. For Rise of the Sentinels, we wanted to narrow that story progression gap between hardcore engagers and normal engagers. We designed Rise of the Sentinels so progress increased every week, meaning hardcore engagers would still complete the early content faster, but normal engagers would be mostly caught up by the time we reached the finale. We thought this'd also be a cool thematic representation of the Sentinels' momentum rising as they bolstered their ranks.

Ultimately, we ended up patching the event partway through to greatly increase progression per game played, which removed the impact of deciding which regions to visit first, as well as basing progression around the Sentinels to begin with. On the one hand, this was because we flat-out missed on event tuning: Not even the hardcore engagers were progressing at the rates we designed for. On the other hand, backloaded progression was a risky choice from the start. Even when tuned well from an event-long perspective, slow progress in the first few weeks can feel demoralizing. Normal engagers in particular can end up feeling like they have no shot at completing the story.

Our takeaway is that Spirit Bonds' backloaded grind was a better event experience, even though players completed the main portion of the event at radically different times. Faster upfront progress means that if you're playing at a regular pace, you're still immersed in the story and world from the start. You can quickly get a sense for whether the event interests you, and if it does, feel confident that you'll complete the primary stuff by the end.

For both regular engagers and the hardcore folks, we also need to be more diligent in our tuning to be absolutely certain players will progress at the intended rate.

Gating lore behind gameplay

While related to the last topic, this one warrants its own callout. We gated the Rise of the Sentinels story behind event progression so you would feel like a part of the fight against Viego. In part because of the harsh event tuning, but also as a general outcome of a progression-locked design, players had to play more games than usual to fully access the story. Many folks without extra time to play more League never experienced the conclusion themselves, and had to rely on outside sources to fill them in. Moreover, it positioned gameplay as an undesirable obstacle lore fans had to overcome to get to what they really wanted, worsening their relationship to League rather than improving it.

Our takeaway is that lore should be accessible, especially canon Runeterra lore. We'll avoid putting major canonical lore beats behind significant grinds in the future, and metagames like Rise of the Sentinels won't be the only way to access the lore.

Multiple versions of the story

This section will be a little tricky since it's about the overall Riot landscape, but we're only directly speaking to League PC's portion of how the story was told (Rise of the Sentinels in the client and the cinematics on the web).

For a long time, League PC was the only game shipping story content for the League Universe. This made it simpler to keep our narrative and storytelling consistent, though even then, we had challenges. As League has grown, more teams are telling stories in the League universe. We didn’t put the correct systems in place during that growth to maintain narrative cohesion among all these teams.

During Sentinels, we released four versions of the story through (1) Rise of the Sentinels, (2) the Ruination, Before Dawn, and Absolution cinematics, (3) Wild Rift's in-client Sentinels HQ, and (4) their Steadfast Heart comics. Each of these teams made changes to the story to fit their specific storytelling vehicle, though we tried to keep the same core story the same. And for the sake of final clarity, here's the core story:

  • Thresh resurrects Viego
  • Viego attempts to take the piece of Isolde's soul inside Senna, but Senna and Lucian escape
  • Lucian and Senna travel across Runeterra, attempting to stop Viego from collecting the other pieces of Isolde's soul. They encounter Ruined champions and recruit new Sentinels along the journey.
  • Akshan is one of those Sentinels. He had been training under his mentor Shadya, the Lost Sentinel. Shadya was recently murdered by a Shuriman warlord, and when we met Akshan, he was attempting to kill Shadya's murderer with the Absolver to resurrect her.
  • As Viego's quest nears completion, the Sentinels take the fight to him in the Shadow Isles
  • Senna willingly surrenders to Viego, as the piece of Isolde inside her insists it's the only way to save the world
  • Senna dies as Viego completes Isolde's soul and resurrects her
  • Isolde secretly tells Akshan to kill her, resurrecting Senna since the cause of her death was Isolde's soul leaving her body
  • The Sentinels defeat Viego
  • Thresh siphons the souls from the Ruination, gaining enough power to restore his human form and leave the Shadow Isles

An example of a non-core, but still impactful, detail that was changed to fit the medium was Rookie. Rookie was never given a canonical appearance (or gender, age, etc) in Rise of the Sentinels, so the cinematics team adjusted the story to make sense without Rookie, whom they had no way of showing.

Even for those of you who only experienced the League PC content, these adjustments made it unclear which version of the story was the 'real' one, and what the other versions were supposed to be. This led to a poor experience especially for hardcore lore fans, who were the main players we were building Sentinels for.

Our takeaway is that simultaneous forms of storytelling that support the same lore moments need to be consistent, rather than optimized for individual products. It's alright if one storytelling form can only tell a part of the story due to scope limitations, but that part must be consistent with the whole that's told elsewhere.

Champion inclusion & visual novel scope

We wanted the Sentinels to feel like a rag-tag group of unlikely heroes from all over Runeterra that Lucian and Senna recruited by happenstance, rather than an ideal team carefully assembled over the course of years. While this premise works on paper, we weren't able to develop all the champions enough to succeed in practice. Even with 'only' about a dozen main characters in Rise of the Sentinels, many of them felt like they were kind of 'just there'. Players never really got the rag-tag sense we intended, so the 'unlikely' inclusions felt like bad fits, and champions who better fit Sentinel ideals (ex. Braum) felt like conspicuous absences.

In retrospect, the scope of a global Ruination was far beyond what we could deliver in a 5-10 hour visual novel. Most of the experience came in the form of dialogue between a handful of characters, which didn’t lend itself well to a world-encompassing battle. For example, many of you wanted to know what key leaders from the major regions (ex. Swain or Jarvan), were doing, even if they wouldn't realistically join up with the Sentinels. But if the Sentinels didn't directly encounter that leader, the visual novel didn't have satisfying ways to include them.

The global scope put immense pressure on the story and roster as well. Visiting every region and recruiting a Sentinel from each meant we couldn't spend too much time in any one region or include too many of its champions. The Shadow Isles suffered from this the hardest: Lore fans were left wondering about Yorick, Maokai, Kalista, and Hecarim, who all have ties to the Ruination storyline, but were minor characters or absent entirely.

Our takeaway is that metagames and visual novels can only go so far in paying off a big lore event. A feature like this can add color and detail to lore beats (like how the Council Archives supported Arcane), but it can’t carry the entirety of a massive story by itself. If we had done a story smaller in scope, maybe only including a couple regions with a tighter and better-developed cast of characters, we could have told a more satisfying slice of the overall whole. Dedicated storytelling mediums are better suited for broad-reaching narratives.

Including Rookie

The player stand-in character was a tool we used in Spirit Bonds, and we used it again in Sentinels to give players a front seat to the action. Spirit Bonds made sense as an experience centered around your personal interactions because the Ionian spirits weren't pre-established, so you were literally meeting them for the first time during the event.

Unlike Spirit Bonds, though, Rise of the Sentinels was more about bringing the conclusion of the Ruination story to life—a story which for years had been told exclusively through champions you all have already gotten to know over the years. Routing the central perspective for the story through a new, faceless Rookie felt like we were stealing the limelight from the roster and distracting from the connections between champions you expected. This was exacerbated by dialogue choices often feeling formulaic and like they didn't lead to meaningful impact.

Our takeaway is that faceless player stand-in characters as a tool aren't something we should assume by default when developing visual novels, and that relationship-building doesn't have to always be the focus of them. Some of you suggested, for example, that we could've told Rise of the Sentinels through Lucian's or Senna's eyes, which still would've given that front seat experience without feeling like a deviation from previous Ruination narratives.

Tone

Narratively, the Ruination was a bleak story. We wanted to add some lighthearted moments and jokes to ease some of the heaviness and tension since Sentinels was a longer event, but it’s clear that we leaned too heavily into comic relief. This created a couple issues:

First, players felt like there weren't enough serious or neutral responses to choose from with Rookie’s dialogue choices. For these players, that prevented Rookie from accomplishing their role of feeling like a stand-in for the player.

Second, the tone was inconsistent from one scene to the next. Sometimes the story was dead serious with very high stakes, but other times it was comedy hour. This made it hard to get a feel for what kind of experience Rise of the Sentinels was supposed to be.

Third, it caused Rise of the Sentinels to underdeliver on the stakes and gravity of a global Ruination, the first world-level threat we've introduced in current Runeterra.

Our takeaway is that tone needs to be more consistent across an event. When the storyline calls for it, we should be comfortable with things being a lot darker and more serious, with only small moments of comedy rather than swathes of it.

Mischaracterized champions

A specific sub-point of the above is that some champions felt like they were being portrayed inaccurately. Specifically, some dialogue options elicited responses like "My champion wouldn't say that." Several champions were called out as being too cartoonish, with Rengar as the biggest example and Pyke coming up a lot as well. The other frequently highlighted case of problematic characterization was Lucian in the early chapters. We wanted him to feel like his back was up against the wall because of the danger Senna was in, but he came off as cold, angry, and even cruel instead.

Our takeaway is that even when we're layering event-specific context onto champions, they still need to feel true to the core of their previous characterizations; they'll feel unnatural if pushed too far. The previous takeaway regarding tone will also help here, as there'll be less pressure for us to use champions as vehicles to add levity where it wouldn't organically appear in the first place.

Overly complex metagame systems

Rise of the Sentinels featured ten Sentinels, each with one of two progression mechanics depending on the order you recruited them in. The in-client hub sent players from the central chamber to character bio pages to the world map and to region maps. For many players, this complexity created too high of a barrier to really get into the event.

As one particularly frustrating callout, some of you who played a ton of games in one sitting returned to the Sentinels hub to find that much of your progress was lost, because you unknowingly completed a region partway through the session and didn't know you had to choose another region to keep earning points.

Our takeaway is to streamline the design of future metagames; both mechanically and in terms of interface and experience. This doesn't mean less metagame content overall, just that metagame content needs to be less confusing to experience.

Looking forward

Sentinels of Light was the capstone to our first multi-month story arc, our first big, immersive lore event on PC in years, and our first lore event across multiple products at Riot. While there were a lot of things that could've been better, we do think there were things that went well. We're still happy to have launched an event at the same tier of scale as Spirit Blossom, including Ultimate Spellbook, our first new event game mode in years, and our first coordinated champion release across four Riot games at once.

Though it's a little cliche to say, the sheer amount of feedback we heard from you, and the depth and care reflected in much of that feedback, proved to us that there is an appetite for us to do more with League's universe and champions. Lore events in League of Legends are not dead.

We hope this post encourages those of you who were let down by Sentinels, and that our takeaways line up with what you'd like to see us do better next time. See you all in 2022!