TFT Dev Drop: September 2021
A new set, mechanic, and lab? Oh my! Discover everything coming to TFT in the latest Dev Drop.
At a high level we wanted Preseason 2021 and our work throughout 2020 to be about strengthening and improving the foundations of League for the next 10 years. I wanted to take a quick overview of the state of the game: What we’re feeling good about, and what we want to improve.
We updated the item system to increase strategic choice, satisfaction, and accessibility. We’re generally seeing those goals met when we look at the game widely, but we're still focusing on follow up work to realize the full potential.
Most champions and classes are seeing a wider set of item choices that suit different games, but there are some cases where choices are either not there, or not clear and satisfying enough to feel there. We’re seeing roughly 20% of champs crossing our “Mythic hardbinding” threshold (same Mythic in >75% of games).
We’re going to continue addressing these gaps to get as close to 100% of champions with meaningful build diversity. That said, build diversity is one value among a long list of gameplay goals, so we wouldn't sacrifice too much of other important things like satisfaction, counterplay, or champion uniqueness to achieve it. This means addressing hardbinding will take longer than standard balance concerns since the work required is more disruptive and risky, but getting as close to 0% Mythic hardbinding as possible is one of our ongoing focuses.
In the last few patches we’ve assessed the state of item choices and power level across many classes to tune the greatest outliers. 11.1 saw buffs to underused items like Shurelya’s and Ionian Boots; in 11.2 we adjusted the tank Mythics to help Frostfire Gauntlet and Turbo Chemtank compete with Sunfire Cape; in 11.3 we’re targeting fighter class outliers like Goredrinker and Sterak’s Gage, and shifting systemic healing toward more conventional durability.
Balancing and reworking the new groundwork we laid in preseason will eventually taper off into normal patch work on items. Our hope is that, by the end of spring, every champ has a range of satisfying (even if not totally perfect) choices. As things settle, we’re looking at a few opportunities to add new items to flesh out a few classes.
The item shop overhaul has hit our goal of creating a tailored experience for each type of player from new to master level experience. The upgraded recommendation system using high level player data in particular has been performing quite well: Lots of players use the Recommended tab as their default and it’s been exposing players to more viable choices and teaching them the basics of when each option is worth considering.
I would still estimate that our new shop is only 90% of the way to completely finished though. As we continue to play with it, small but critical usability improvements become clear. Now that we're using modern tech, we hope to continue adding more and more functionality over time to make the shop usability feel like second nature.
With so many item changes this preseason we were anticipating some needed followup on position balance. (We look at both pick rate as well as a position's ability to influence the outcome of a game to assess its power and health.) Bot and Support landed weaker than before the update, but we focused on followup for both during the preseason window and they've recovered to appropriately powerful spots.
One thing on our radar for the future is the Jungle position. It’s one of the most influential/powerful positions, but also low popularity, meaning there's a satisfaction or accessibility problem keeping it from being appealing despite how powerful it is. We've made light-touch changes on this front a few times in recent years, like adding jungling tools to champion kits, putting camp timers on the minimap, and increasing viable path options so more champs can make it through the first few clears. Those light touches expectedly weren't comprehensive solutions, so Jungle continues to be something we're thinking about.
Worth noting: We’ve seen positive reception to the new Jungle item experience of buying a jungle starter and getting to your build right away. This not only streamlined jungling for new entrants to the role, but also freed up build paths for veterans. We're confident there are still more ways like this for us to improve the appeal of the role without undermining the core skill and flexibility that current junglers love.
At preseason launch, many players felt an increase in burst damage (mostly coming from big item actives or passive procs), past the point of where we think the healthiest and richest version of LoL gameplay should be. When damage gets too high, counterplay and in-combat decisionmaking suffers, so we began looking into the issue once preseason data stabilized enough for us to be sure it wasn't just noise right after a big update.
In the last few patches, we've pulled burst from multiple sources and are back to a reasonable state, though still probably the higher end of acceptable. We're continuing to evaluate whether more changes should be made. Generally, we try to stay in the range where burst reliant classes (mages, assassins) can kill an out-of-position carry, but not so bursty where they can do so without hitting a full combo or when missing important skills. We also use mathematical analysis with historical comparisons (how much damage is a Talon combo in S11 vs S10 vs S6 etc). Another method is test cases like LeBlanc's time-to-kill vs a Vayne at typical builds. This accounts for shifts in target defenses as well, which can be missed when just looking at a damage dealer's theoretical numbers.
Something we always watch closely is snowball, which tells us whether the game flows in a way where player decisions are impactful at all stages. One of our preseason goals was to avoid power creeping the item system, and increased snowball would be one of the worse symptoms of messing up because finishing items earlier than your opponent would give you an even bigger advantage. Snowball currently looks similar to the last few seasons, which is good.
Data aside, we know playing from behind in League can be very frustrating even if you still have a realistic chance of making a comeback. Being at a disadvantage will probably never feel good, but ideally comeback opportunities are present and viable enough to keep games fun and exciting to the end.
2020's champions were successful on both the satisfaction and appeal fronts. The themes and gameplay of Sett, Fiddle, Voli, Lillia, Yone, Samira, Seraphine, and Rell aimed for broad appeal among players within their applicable roles and archetypes. Launch balance was also significantly improved compared to 2019, with Samira as the only champ who launched significantly off-target. But we know that there are always ways for us to improve and iterate in 2021.
This year, we’re looking to improve the “good for League overall” side of new champs and VGUs. This includes things like strong and clear counterplay, very unique kits that justify each new addition to our roster, lower complexity on average, and more unique thematics to expand our IP even if less popular. It's worth noting that 2020 was definitely better than 2019 on this axis, but we want to go further in 2021.
We launched a small bundle of Ranked upgrades in both preseason and start of ranked this year. Overall this has led to a smoother climb where players are facing more balanced teams, fewer smurfs, and consistent wins becoming more important than win streaks.
We did discover one issue at the start of season where MMR was adjusting too slowly, causing LP gains to be a bit lower than expected overall, and in some cases caused winstreaking players to experience negative LP gains (eg +12 for a win, -16 for loss). That issue was addressed with patch 11.2 last week. We’re going to continue monitoring all aspects of ranked to be sure that the experience is working as intended. We’re also going to be watching this year to see if these improvements have positive long term effects that make playing ranked a better experience over the course of a season.
Thanks again for playing League with us. Stay safe out there.