It's been roughly a month now since we first talked about One For All, our first featured game mode, and in that time our team has been hard at work getting the last few issues resolved. Starting this weekend we'll be testing the game mode on PBE. Given that, I wanted to take some time and give a bit of insight into some of the challenges we've run into getting the feature off the ground. As we promised in the initial discussion, we really want to strive for an open dialog about each of these game modes, and discussions like this will be the first of many steps along that vein.
One For All, at the outset, seemed like a pretty straightforward game mode to develop, but we ran into some unique challenges along the way. As some of you might have seen when One For All was temporarily "available" a few weeks ago, when multiple copies of the same champion are on the same team, their abilities tend to clash in unique ways (like only two Heimerdinger turrets per team, regardless of the number of Heimerdingers on the team). This was due to the way these abilities were coded, so our first road bump was to tidy them up - no big deal. The real challenges we faced came from more nuanced interactions.
A fundamental design principle that we value highly is that of "Intuitiveness" - effectively, when you look at a skill or ability can you easily extract what it will do when used. Big ice bolt that explodes? Probably does some damage and slows. Spell called "Transfusion" that pulls blood from the target? Probably a damaging and healing spell. Making champion abilities intuitive, even when multiple copies get stacked on top of each other, was an important project we wanted to tackle. An example here would be Akali's Twilight Shroud: should it stealth every Akali who enters it? Sure! What about giving them the magic resistance and armor bonus per additional shroud? Maybe not.
Here we decided to break apart each skill to identify their core fantasies. Like Akali's shroud, where the stealth was important while the defensive stats were secondary, we went through every ability in the game to create a roughly balanced experience. In two cases with Teemo and Karthus, however, we ultimately couldn't find a balanceable solution for them in blind pick and had to take more drastic measures. For this particular release, we opted to actually remove them from being selectable champions in the matchmaking queue, although they will be selectable in custom games for those who want to try. Fighting five Teemos was like reenacting World War 1 with landmines everywhere, and fighting five Karthuses (Karthi?) was just getting aced every 160 seconds. So... don't say we didn't warn you!
Another open question was determining how people would select their champion. In a matchmaking queue we needed a solution to help a group of random people come to a seemingly fair resolution. We knew we wanted a banning phase and, since the captain would be responsible there, we first started with having the captain choose the champion for their team. After a few playtests though we saw the captain just choose whatever champion they wanted, regardless of their team's input.
We next explored solutions with voting, hoping it would help drive toward team consensus. This didn't work out much better. In boardgaming there's a style of play called "Kingmaking," where a player stops playing to win and instead works to tip the scales in favor of another player - we saw the same approach adopted here as well. Frequently we'd see four players disagree and each vote for a unique champion, leaving the fifth player with a choice. Either they could vote for their own champion and maybe get it via random choice in a tie-breaker, or they could vote for one of the other champions that, while not their first choice, was still acceptable. The value of voting last quickly caught on, and soon everyone was waiting to last second, hoping to get that kingmaking opportunity. Ultimately the final solution we came to ended up being fairly simple, even though it took a while to reach: a champion would only be selected if it ended with a strict majority of votes instead of simply being "the most."
The final question we faced has more to do with game modes in general than One For All in specific: availability. After a lot of discussion, we came to the realization that if we wanted to deliver fun, unique game modes with a lot of ongoing player engagement, launching them as permanent fixtures whenever we felt confident enough to ship wasn't the best approach. Having a game mode permanently available has a number of costs, the first of which is increased stability and support costs across the entire game. While adding a single game mode has a known impact, the multiplicative effect of multiple game modes living side by side means we'd soon be bogged down by simply maintaining these game modes, unable to work on the next unique thing.
A second consideration is that having a game mode permanently available actually is a big restriction to the type of experiences we can create. If we want the mode to have a healthy player base a year or two years from now, each game mode needs to be designed to remain engaging after thousands of hours of play. Due to the difficulty of that task it means every game mode would take an exceptionally long time to develop, and we wanted to give you guys stuff much faster than that. By intentionally going into this with the expectation that a game mode is temporary, it freed us up drastically, and it really allows us be much more experimental with the game modes we try.
After we cycle out a game mode, we'll go back to the workshop to tinker with it based on player feedback and, in the meantime, we'll deploy a few other game modes we've thought up over the past while. This lets us rotate any number of game modes in and out as we refine them, and it gives us the freedom to even take community votes for a "player's choice" month!
For the release of One For All, the matchmaking queue will be for Summoner's Rift, but you'll be able to play the mode via custom games on any map. Please let us know what you think of this mode and what changes you'd like to see to it when or if it comes back. Have fun with the mode when it is released, and I look forward to talking with you again about the response and what is coming in December!