/dev: TFT Set 1 Learnings

Teamfight Tactics has been out for a little over three months now, and we’ve nearly reached the end of our first set. Time has flown by so fast, and we can’t wait to show you what’s coming in Set 2: Rise of the Elements. But before the new set hits PBE, we wanted to discuss some of the things we’ve learned in our first three months and how we’re taking those learnings forward into the future.

We’re going to go pretty in-depth today, so here’s a quick summary:

• We want to keep random effects in positive and controlled spaces.
• You can hold us accountable to higher standards of balance so that a wider variety of comps can win in any given patch.
• We’re aiming for less frustrating and fewer overall disables.
• We’re going to have more soft counters than hard counters, especially between traits.
• Rate of change was overall a bit high, so expect B-patches to only be used to nerf strong outliers.
• Bugs are bad, and we’re going to take more time to make sure we deliver a game with less issues.
• We want getting 4th place or higher to always feel like a win, so we’re changing it so you always earn LP for a top-half finish.

Good Random vs Bad Random

Randomness is always going to be a hot topic when discussing TFT. At a high-level, Set 1 reaffirmed to us that TFT is about how you play around the randomness presented to you through the store, item drops, and various effects in game that cause variance. We still strongly believe that utilizing randomness in the correct way creates a wider variety of experiences, which leads to more replayability. In addition, one of the ways players can express skill is through knowing how to adapt to the many possible outcomes that can occur.

That being said, we certainly learned a lot about where inserting variance worked and was received well and where it wasn’t.

For example, two-piece Imperial (most commonly with a Darius and Draven) was a very random effect that boiled down to a coin flip on whether or not your most important carry would do double damage. We found this fit into acceptable random because you opted into the risk, and it was fairly easy to predict the possible outcomes.

The counter proposals here are Phantom and Hextech, which were our most controversial traits. Many players had a very real negative reaction to their most important champion losing all three of their items or being set to 100 HP with nothing they could really do about it. Phantom in particular did achieve our goal of creating exciting moments and chances for you to win fights you might not normally win, but it did so in a way that was likely too costly for the overall experience.

So as Set 1 went on, we made improvements by moving towards more controlled randomness. For example, we revamped the item distribution system so that the outcomes were more controlled than before while also opening up a wider variety of potential outcomes. We saw this lead to a more fair and more exciting experience. We did something similar with Thieves Gloves, where instead of simply giving the champion any random two items, we created an extremely controlled list and limited where the power can show up at various stages in the game. Again, we believe this led to positive results.

With all that said, we’ll be continuing the trend of trying to keep random effects in positive and controlled spaces in Set 2.

Champion Balance & Design

Honestly, this section could be an entire article by itself. Our understanding of what makes a good champion in TFT and what balance qualities they should have has increased exponentially since TFT launched. Let’s try to keep it shorter and discuss some of the most key things we learned though.

Trait tax shouldn’t be as severe: We saw pretty early on with champions like Fiora, Lissandra, and Mordekaiser that having champions who were intended to be bad because they enabled very powerful traits just didn’t pan out. It also limited the types of comps people were willing to run since it meant having to use these less-than-ideal champions. Few people enjoyed having to carry around a bad Mordekaiser just to enable Phantom. It’s okay for them to be less powerful, but not to the degree we launched with.

Even bad champs should have a fantasy: Champions that weren’t the obvious powerhouses were and still are bound to happen—this is fine. But those champions need something to aspire to in the right conditions. Katarina was the best example of this: When given the right items, she could carry the game for you, despite being generally weak. Three-star Veigar was another big one. We aim to create more examples like this in Set 2.

Champions shouldn’t be overloaded: Certain champions simply had too much going on in their kits in Set 1. Pantheon’s spell had a stun, AoE percentage health damage, and a built-in Morello. Akali’s spells only cost 25 mana and were allowed to crit without a Jeweled Gauntlet. Shyvana got a free Dragon’s Claw from her trait, a free Red Buff on her autos, tons of AD, and a burn effect. In all of these cases, the champs had more effects and rules than they needed to. We’re going to keep this more in check for Set 2.

One last note worth bringing up is that balance wasn’t as good as it could’ve been across the patches. We learned that even something as small as 0.05 attack speed or 50 starting HP can make a huge difference. Over the course of Set 1 we tried a variety of tactics—such as nerfing multiple aspects, a single aspect in a big way, or a single aspect in a small way—and we learned what works and what swings things too far. So in Set 2, hold us to an even higher standard for balancing the game so that a wider variety of comps can win in a given patch.

Disabling Effects

We already knew this going into TFT, but abilities that disable your champions are perceived to be less fun. When designing abilities in League, we’ve seen a pretty visceral reaction when they cause a loss of control or inhibit your ability to perform expected actions. In TFT, Glacial initially had some very strong reactions from people who said it was no fun to have your army just frozen in place. However, we knew that this was acceptable because we need disabling abilities in TFT as outlets to damage, otherwise the game simply becomes who wields the biggest damage cannon.

That being said, there are a couple areas we know we can improve on. For starters, we looked back and agree that we ended up with too many disables. Demon, Glacial, Hextech, Phantom, Hush, Cursed Blade, Swordbreaker, Zephyr, Leona, Ashe, Sejuani, Gnar, Kennen, and more… it was too easy to get your army in a state of “I don’t get to play.” Because of this, we will be aiming for less disables overall in Set 2. (Important here, we said less. Not none)

We also learned what types of disables worked and what types didn’t. For example, pre-changed Demon was extremely painful, since it prevented most champions from ever living out their fantasies. Once you got hit once, it was probably over since you lost ALL your mana. New Demon and Hush worked much better—while they did disable you, there was the hope that they’d wear off, or you could still gather enough mana to cast your spell. Predictability also played a big part here. Zephyr ended up in a good spot even though it disabled a character for six seconds (a much stronger effect than Hextech) because it was predictable and could be played around. You can expect to see more disables like these in the future.

Hard Counters vs Soft Counters

Early on in Set 1, we believed that having hard counters was the best design choice for TFT. Dragons had 100% magic immunity as a hard counter to spells, Grievous Wounds blocked 100% of healing as a counter to healing effects, and Wild/Rapidfire Cannon made your attacks undodgeable as a counter to Yordles. (Poor poor Yordles.) In all of these cases, it led to some pretty extreme meta situations where the things being countered were considered to be not viable options by many players. Yordles in particular pretty much always struggled to find a good spot, even with the introduction of Mittens.

In all of the above cases, we saw much more success when we pivoted to a softer counter. When Grievous Wounds became 80% heal reduction (which is still REALLY strong), we saw healing effects being used more, leading to a wider variety of viable strategies and builds, since players could still get some value from Bloodthirsters and Hextech Gunblades.

There was also the case of “general counters,” which we define as something that is just universally good against everything. Noble became a general counter when it swapped from giving armor to giving armor and magic resist. The only way to counter Nobles was the new Void trait, which had the same quality of being just generally strong in all cases. True damage as a whole fits this as well. General counters are usually too good, and we’re going to be moving away from them.

When Set 2 launches, you will notice that the traits are more in the soft counter space. In addition, we’ll be making some changes to the items as well to shift them into a soft counter space in a future patch after Set 2 launches.

Expansion Content in Set 1

With Set 1, we released quite a few pieces of what we call “expansion content,” which is stuff that expands the set beyond its initial launch state. This included:

• Twisted Fate in 9.14
• The four Hextech champions in 9.16
• Pantheon in 9.17
• Item system revamp & Neeko’s Help in 9.18
• Kai’Sa and Sparring Gloves in 9.19

From what we saw, there were some pretty clear takeaways. Having a big piece of expansion in almost every patch contributed to the feeling that the rate of change was too high. In addition, the smaller drops like Twisted Fate and Pantheon were much less successful at generating excitement, where as the larger drops like Hextech and the item system revamp did a much better job.

As for champions, we learned that…

• The 50 champs we launched with was a bit small, especially in the 1-cost area. When we launched Twisted Fate, we swapped Elise to a 1-cost, which led to a better early game state, but TF himself was also a bit of a disappointment at the time.
• Adding a new trait was very exciting (ignoring the quality of Hextech itself), though ramping from 51 to 55 champs was a very large change to the feel of the game. It was immediately noticeable how much more difficult it was to two- and three-star champions.
• Adding the two 5-cost champs at the end was probably not great, since the change in difficulty in chasing legendaries changed so dramatically so quickly. Noble was the biggest victim of this, with Kayle going from a pretty reliable chase to almost luck based—which was a pretty big nerf to the trait. We want legendary champions to be exciting capstones to late game builds that feel great when you find them. If they’re too difficult to find, then you stop trying to chase them.

With all that in mind, for Set 2 we are going to try shipping less total content patches, but each one will be larger. Our current thinking is 3-4 content patches over the course of the set. Our goal is for each of these drops to bring an exciting and meaningful change.

B-Patches & Rate of Change

One of the things we dealt with a lot in Set 1 was, “What is the right amount of change for the game?” Early on we had B-patches and hotfixes out of necessity—we had to address things that needed to change very quickly, including bug fixes and rougher early designs. However, we saw some frustration with this, with some players feeling like it wasn’t worth investing in the game and learning the meta: Why invest deeply in the game if it was just going to change in a few days?

One of the key contributors to this was the size of the B-patches, which sometimes changed as many as 15-20 parts of the game in a single patch. We felt this was just too much change for most players to deal with.

So in the second half of Set 1, we moved away from B-patches. We drew a hard line that unless it was a major game-breaking issue, we weren’t going to hot fix it. (For example, we did not hotfix Void-Assassin in 9.18, but we did hotfix Titanic Hydra in 9.19.) We learned that there were positives to this approach in that we would see the meta shift as the weeks went on, and players began to make counter comps to the popular comps. We want to see more of this. However, this method wasn’t without frustration either, as any time something was even slightly off balance (Void-Assassin in 9.18, Infinity Edge in 9.19), engaged players were stuck with it for two weeks, leading to what felt like a stale meta.

Based on what we gathered from Set 1, we’re going to take a different approach for Set 2. The first patch of Set 2 will likely have a bigger B-patch to resolve any post-ship issues, but after that we will limit B-patches to be extremely small and only for nerfing strong outliers. This allows us to deal with some of the most extreme cases while slowing down the overall rate of change. Hopefully this will lead to a more stable Set 2, but also a less stuck one.

Bugs!

This one is a bit of an obvious learning, but we want to discuss it because frankly it’s very important. Bugs are bad. We shouldn’t have bugs. Due to the rapid pace of building TFT and the content in each patch, we didn’t hit the quality bar we would have liked to, which led to days of ranked being disabled, hidden exploits, and other not great experiences.

We still have a long way to go to get everything up to the bar we think you should expect of us. But we’re committed to getting there. One thing we’re doing in Set 2 is taking a little more time to QA when we add new content. Delaying the new items from patch 9.18 to 9.19 was a step in the right direction, as it allowed us to have far fewer bugs with the new items (though admittedly we still didn’t get them all).

For Set 2, expect us to take our time to make sure we do a better job at delivering a game with less issues.

Other Things We Learned

There’s a lot more we can discuss, but this is a long article, so let’s wrap up with a few final learnings.

With traits, we found that adding the extremely unlikely chase options—like nine Blademasters and nine Sorcerers—created some extremely exciting moments for players to chase after and some memorable moments when you managed to nail it. Watching a Karthus one-shot an entire team with the nine-piece Sorcerer bonus was a personal highlight for me! We’ll continue to add these types of chases in Set 2.

We also learned a lot about player damage and game pacing. When we launched, games were over far too quickly and it was almost impossible to make it to late game. After some adjustments, we ended up in a state where late game was all that mattered because the early game wasn’t doing enough. The relationship between available resources and player damage was also important, as when we added the new item distribution system, we saw a shift back to quicker games. While we learned a lot here, we don’t think we’ve gotten to the ideal state, so we’ll be making more adjustments to the pace of the game as we continue into Set 2.

Finally with our Ranked system, we’ve seen some things that work and some that don’t. We’re going to continue to make adjustments to the system as we move forward, but one change coming in Set 2 is that, no matter the conditions, getting 4th or higher should feel like a win. Losing LP when getting 4th (or in extremely rare cases, 3rd) feels real bad. So we’re changing it so that no matter what, you can’t lose LP if you place 4th or higher.

So there ya go. We learned a lot in our Set 1, and we’re so excited for you to get your hands on Set 2: Rise of the Elements!