2024.05.02
By Matt Lowe
I’m Matt, the Design Director for THE FINALS. I’ve been working in the games industry for over 17 years now, predominantly on multiplayer and FPS games. To this day, these types of games are still my go-to experience when I want to have fun. The quality of the multiplayer experience really matters to me and that’s a large part of why I’m here today to explain our approach to ranked play in THE FINALS, both now and into the future.
The goal of ranked matches in any multiplayer PvP game is to let players test their skill against other players, show off their abilities, and have fun improving over time.
The ‘fun’ part is critical here. Ranked players don’t really want to get free wins all the time, as that doesn’t test their skills. They also don’t want to get crushed by their opponents all the time because repeatedly losing isn’t much fun either.
As a result, the ideal scenario is creating fair ranked matches, where players on all teams have similar skill levels, which leads to a good challenge but also gives you a reasonable chance to win. For ranked modes to create these ideal matches, we need some way to measure skill.
That’s why most ranked modes use Skill-based Matchmaking (SBMM), which assigns players a skill rating based on past performances and uses that rating to find them similarly skilled opponents.
Measuring skill, however, is really tough. In an FPS game, when measuring skill it’s easy to just look at what you do in the game: eliminate opponents. This often results in players looking at kill/death ratios as the only measure of skill.
The challenge is that most FPS games, especially objective-based FPS games like THE FINALS, aren’t just about eliminating opposing players. There are objectives to capture or steal, revives that can save teams from being eliminated, well-timed Goo Grenades that seal off an objective, or a clutch Dome Shield deployment that saves a key player from death. Is it possible to determine which of these actions has the most impact on a win? Which is the most impactful skill? And is that always the most impactful skill?
These questions make skill ratings in any PvP game hard to define. Some people dedicate their entire careers to determining where skill comes from in particular games and sports. I recommend you watch the film (or read the book) Moneyball to get an idea about just how difficult this can be.
As a result of this complexity, the most common approach to measure skill in games is one that ignores all the individual actions players take in a match that might lead to a win and instead measures the player’s win/loss ratio and the skill level of the opposing teams. This way, you remove the difficult questions about which actions contributed most to the win from the equation. Players who win more often are likelier to have the important skills needed to win, and can therefore be considered more skilled.
This is what rating systems like the commonly known Elo System do. When a new player comes along, the system matches them against a range of players, each of whom has a rating based on how often they win or lose and who they have won or lost against. If the new player wins several matches against Gold Tier players, it moves the new player’s rating towards Gold Tier. When that new player starts losing lots of matches at Platinum Tier, the system adjusts and lowers their rating, as it can see they don’t win as often against that tier of players.
Rating skill for a game like THE FINALS has a lot of additional complexity, but this covers the fundamentals.
THE FINALS | Season 1
There are downsides to Elo-style league systems though, one of the biggest being that they’re quite good at finding a player’s skill level pretty quickly. Let me explain why that can be an issue.
You may have seen placement matches in the ranked modes of other multiplayer PvP games. Quite often, by the time the player finishes their placement matches the SBMM in these games has already mapped out the player’s skill level pretty well. This is good as the game can now give those players close matches to compete in, but it can be bad because once a player’s skill level is established it’s actually very hard to improve it without really hard work and meaningful training or practice. This means for most players, in most ranked seasons, their skill rating will barely move for the rest of the season and this lack of progress and progression for a lot of players can seem quite boring. Some players want to progress as they play, they want to feel like they’re learning, but a skill rating that rarely moves doesn’t often give this feeling.
Ahead of Season 1 of THE FINALS, we felt most of our players would prefer going on a journey through the ranks each season, to get this sense of progression through the leagues, as this was an approach that was becoming more common in PvP games.
In our Season 1 league, the goal was slightly different to Elo-style systems: to take players on a journey through the different leagues before reaching their actual skill rating. This journey feels fun! It’s nice to progress, to have goals to strive for, and to be rewarded for spending time playing the game by moving up through the ranks. For many players, this is more fun than having a skill rating that doesn’t move a lot once reached (like in the Elo-type system).
To reach the close, competitive matches that players want, we still used an underlying skill system, like Elo, that measured the player’s real skill. That rating was converted into a different kind of rating, Fame points, to place the player in the ranking system. Doing this involved a lot of math, but essentially, the system tried to keep the skill rating and the fame ratings somewhat linked.
The system we used for this in Season 1 wasn’t great though. The underlying rating we used to matchmake during the season was okay, but it wasn’t effectively linked to the visible fame or ranks. As a result, any player could theoretically make it to Diamond Tier by the end of Season 1 simply by playing the game enough.
This meant players would often be matched into games where their skill levels were fairly close, but their progress through the leagues was far apart. This made matches seem less balanced than they were, as players would see many different league icons. Understandably, this frustrated our ranked players.
It was clear that Season 2 would need a different plan
THE FINALS | Season 2
In Season 2, we dropped our old fame system in favor of a skill-points system for tracking seasonal progress, a system that was more closely linked to the player’s skill rating than in Season 1. We still tried to preserve some of the seasonal journey by allowing the progress on earning skill points to move slower than the real skill rating.
In Season 2, players were placed below their actual skill rating at the start of the season, in most cases. As they played they would progress through the ranks, but would plateau once they hit their actual skill rating, causing their skill and league ratings to be in sync at that point.
The approach we tried has actually worked more effectively than season 1 in that the league ranking and skill rating remain more in sync and become closer the more the player plays, but it is still confusing, and the lack of information at the beginning of the season didn’t help. Ultimately we still started Season 2 with a similar issue to Season 1, matchmaking based on real skill ratings but displaying league ratings that can be out of sync means players see Silver ranks in matches with Platinum players, even though their skill ratings might be much closer than their current league rating, due to the seasonal journey.
We also found that the time it takes a player to go from the league they were placed in, to the league that matches their skill rating was a little too long, adding to frustration.
The overwhelming feedback from our competitive players is that they would prefer to play a more traditional FPS ranked mode with an Elo-style rating system, where their visible skill/league rating is an accurate measure of their skill level, but where they may not experience much skill growth or rank progress during a season.
So what can we do to fix this?
Skill ratings and league systems are nearly impossible to change in the middle of a season. Drastic changes now would mean sudden and massive changes in match quality, sudden rating changes, and so on. This would end up feeling broken, confusing, and unfair. Instead, each new season allows the ranking systems to be freshly reset without causing ill will.
Our plan now is to rework the ranked league system again at the start of Season 3 to give a fresh experience that should feel better, be more informative, and be more closely linked to actual skill. We plan to include an end-of-match summary that informs you about your performance and rank and also shares the rank and level of your opponents—something we aren’t showing right now.
One change we’re going to make in Update 2.6 is to reduce the duration of the seasonal ranked journey, which will allow players to reach their actual skill tier in the league system sooner than before.
To help the quality of matches, we’re going to reduce the size of ranked tournaments from 48 players to 24. Trying to find 48 players of the same skill level within a short matchmaking time is difficult and has often resulted in tournaments where the range of skill levels is too broad for close, fun matches. By cutting the number of players needed to start a tournament in half, it will be easier to find enough players of the same skill level quickly.
We’re hopeful that these changes will improve the experience for the rest of Season 2 and create a much easier-to-understand ranked system in Season 3.
