If you have ever played a competitive game, you are probably familiar with tiers. They seem to be everywhere, and can shape the competitive landscape of a game. But have you ever wondered why tiers even exist in the first place? Today, we find out why!
The following is a reproduction, and has been modified for this site. The original article, and many more, can be found at RemptonGames.com
If you’ve ever played a competitive game, you’ve probably been in a situation where you fall in love with a character, weapon, skill or deck, only to discover that it is not competitively viable. You may do your best to try and discover a new strategy, or a new combo that will make your choice work, but to no avail. You try time and time again, but end up getting beat every time.
If you have been in this situation, odds are that the strategy you have fallen in love with is “low tier”, and simply can’t compete against the more powerful, “top tier” options. This situation comes up all the time in competitive games, whether they be fighting games, MOBAs, or CCGs, and every time it can be frustrating. It can make you wonder why this is the case – why can’t every character be equally powerful?
Today, I am going to take on the topic of tiers in competitive games. In the process we will learn what competitive tiers are, why they exist, and why they may actually be beneficial in some circumstances and unavoidable in others.
If you have been involved with a competitive game before, you may not be familiar with the idea of tier lists. If you have, you can probably skip this section.
A tier is a set of characters, weapons, cards, etc. that are considered to be of a roughly similar power level. Tiers can be labelled with numbers (Tier 1), letters (A Tier), or even have specific names (Overused). Lower numbers / letters usually represent higher tiers, while higher numbers represent lower tiers. The most common exception to this rule is “S Tier”, which is usually placed above “A Tier” in an alphabetical tier list.
The “S” stands for “Super Special Awesome”
Tiers are generally seen as the result of game imbalance – because the game is not perfectly balanced, some strategies end up being more powerful or competitive than others. Some players argue that if the game were to be perfectly balanced, tiers would cease to exist, and that this would create a better, more competitive game.
However, this is not necessarily the case. There are many reasons why perfect balance in a game may not be desirable, and even if it is a desirable trait it is often not achievable. In this article I am going to look at several reasons why tiers exist, and why they may even be a good thing for some games. Not all of these reasons apply to every game, but every game will have to deal with some combination of these factors.
The above statement is not self-evident, but it is true. Consider Chess – this game does not even provide different options to the player – every player gets exactly the same pieces, in the same amount, in the exact same positions. However, even in this game there is a tier – White is strictly better than Black because it has first move advantage. By making the first move, white gets a number of advantages – firstly, it immediately puts Black into a reactionary position. Secondly, it is always either a move ahead of Black or tied, but never a move behind (an average of ½ a move ahead). Suppose that both White and Black would be able to Checkmate the other in 10 moves – because White moves first, they will reach that Checkmate first.
Statistical analyses of thousands of Chess matches have show this to be the case – White wins somewhere between 52% and 56% of the time. While some have argued that there are psychological factors at play that may be affecting these outcomes (and these almost certainly play some role), analysis of games played between two computers have shown similar results.
And people still try to say white privilege doesn’t exist
Sure, you could try to give White some sort of handicap, to make up for their inherent advantage. However, it is unlikely that your handicap will be able to exactly match the value of the first move advantage. If you do not handicap enough White will still have an advantage. However, if you handicap too much it now tips the balance in favor of Black.
This difficulty only becomes more pronounced when you actually begin to add distinguishable choices into the mix. Take a fighting game for instance. Characters have variations in the size of their hit-boxes, attack speeds, movement speeds, health points, attack ranges, attack damage, and many other factors that will tip the balance in one direction or the other. With all of these variables to deal with, it is extremely unlikely that every character is going to end up at exactly the same power level. When you add additional factors such as different players having different objectives (like in Overwatch) this only becomes more complicated.
Although developers are dealing with an overwhelming amount of variables to tweak and factors to adjust, they often do their absolute best to make their games as balanced as possible. Unfortunately, there is no single definition of game balance that will satisfy everybody.
Going back to our generic fighting game example, suppose that you are a developer who is trying to make this game as balanced as possible. Lets suppose you have the data from dozens of matches between playtesters, and you are trying to analyze this data to determine whether the characters are balanced. What do you look at?
Perhaps the most obvious thing to look for would be the win-loss record of each character. If every character has a win/loss ratio of 50% each, then the characters are balanced, right? Perhaps, but not necessarily. Suppose that one character, a speedy ninja who attacks with ranged throwing stars, has an even win/loss record overall, but wins 75% of the time against another character (say, an armored knight). Does this count as balanced?
The sneakiest ninjas wear bright yellow
Even though everything may seem even overall, this highly unbalanced matchup may actually hurt the knight’s viability in a competitive landscape – players may be wary of picking that character knowing that they will have a severe disadvantage based on what their opponent chooses, and may choose to go with a lower variance character. How do you go about balancing this? You have several options:
Make adjustments to one or both of the characters to make this matchup more even
This option seems enticing – If you can adjust these two characters so that their matchup is 50/50 then every character will be viable against everybody else. Unfortunately, changes cannot be made in a vacuum. Any changes that you make for this particular matchup will affect other matchups with this character as well. While you may make this one match more balanced, you will likely end up unbalancing something else in the process.
Try to balance the characters in such a way that their good and bad matchups even out
The idea behind this strategy is that if the ninja has a particularly good matchup against the knight, they will have an equally bad matchup against another character (such as Patrick Swayze from Roadhouse). In theory this could create a rock-paper-scissors type strategy where part of the strategy of picking a character is predicting what your opponent will pick. In my experience, this strategy is used more often in practice. However, it only works if the amount and magnitude of good and bad matchups among characters are relatively even, and no single matchup is too lopsided.
Of course, it is mathematically impossible for Swayze to have a bad matchup
Either of these ideas could be considered “balanced”, but they are very different from one another. In addition, they do not take a number of other factors into account. If your game has different maps these also need to be balanced. If your game allows equipment or items these may need to either be balanced in a vacuum or in context of the different types of players that may use them. If your game allows team-play then different characters may not only need to be balanced individually but as members of the team.
All of these factors can not only make the process of balancing more difficult, but actually change the goalpost of “balance” that you are aiming for. Each of these factors introduce new potential definitions of balance that the developers can strive for, many of which are incompatible with one another.
Suppose that you plan to balance the fighting game discussed above. Lets be modest and say that the game will only contain 10 different characters, and 5 functionally unique maps to fight on. How do you go about testing? Well, if we want to make sure that all of the different matchups between characters balance out we should probably test matchups between each character. Also, to make sure that the maps don’t affect the outcome too much we should probably test them on different maps.
Of course, you can’t just test the characters one time – that doesn’t give you enough data. How many matches is enough to get good data? Let’s be extremely generous and say that the testers can get good results with 20 matches each, 4 on each map.
Each of your 10 characters has to face up against 9 other characters, which leads to a total of 90 different matchups. Assuming that the fight is symmetrical we can cut this number in half to 45 (Knight V. Ninja is the same as Ninja V. Knight). If we had an asymmetrical game like Overwatch, or even something as asymmetrical as League of Legends, you would probably want to test them on both sides.
Unless you think a dragon and a baron are the same thing…
This means that you would have to run, at minimum, 900 different test matches. It also means that any changes made to the maps or characters could potentially invalidate this data, and you would have to start over. It also raises the question of how you will measure these matches – if one character has a win loss record of 85/95 do we call them balanced?
In most real games there are likely to be even more characters, and even more confounding factors such as teams, items, skills, etc., and all of this will need to be tested while the game is still undergoing changes. However, lets say that they are able to achieve this – they play several thousands of rounds with different variables, and according to their data everything seems to be perfectly balanced. Does this mean that the players will feel the same way?
Absolutely not. While the developers may be able to play thousands of rounds, the players will play millions. For a major studio release, it is likely that more rounds of your game will be played in the first hour after release than in the entire time that it was in development. Not only will the players play it much more than any developer could, but they will also be communicating with eachother on a world-wide scale, sharing strategies and tactics.
This means that players will discover things that the developers missed. They will learn how to break a seemingly balanced character, deck or item and make it completely overpowered. The players will tear the game apart, and in the process discover things that the developers didn’t even know were there. They will develop their own tier lists, share them across the internet, and in turn influence other people’s perceptions about the value of different strategies.
However, even after millions or billions of rounds, not even the players will agree on which strategies are actually the best. The tier lists they create will not all agree, and they will not be stable. Characters will move up and down the lists as preferences change, new strategies are discovered, or updates are released. Tiers can be impossible to prevent because at the end of the day they are not real, tangible things, but constructs that the players feel driven to create regardless of how balanced the game actually ends up being.
While tiers are associated with competitive games, most of these games have more than one way to play. Not only are there different forms of competition – team play, individual play, modes with different goals – but most of these games also have single-player campaigns as well. All of these modes will have different needs, and some options may not work as well in every mode.
Take for example Magic: The Gathering. There is no single correct way to play Magic – instead, there are numerous formats, each with their own specific requirements. Standard only allows cards from the last two years of expansions, whereas Legacy allows almost any card ever released. Constructed formats allow players to build their decks ahead of time, while limited formats require players to build their decks as they go.
Some make you put your head on the table
Some formats have 40 card decks, some 60, some 100. Some restrict what colors you are allowed to play with, others what rarities you are allowed to play with. Some let you have 4 of each card, others only allow 1. Some formats are 1 on 1, others are multiplayer. Even among multiplayer formats some are team based, where others are free-for-all.
With so many different ways to play, it should be clear that the same cards and decks are not going to be equally viable (or even legal) in every format. One big reason why something may be low tier in a particular format is simply because it was not made for that format. A card may shine in booster draft but be absolutely unplayable in Standard, and vice-versa.
That is all I have for this week! Ihope you enjoyed this article about tiers in competitive games. There was a lot that I wasn’t able to cover, but I tried to hit most of the main points. If you have some disagreements, additional points that you think I should have covered, or other constructive comments, let me know in the comments or on Twitter and I may do a follow up addressing those points.
If you like this article check out the rest of the blog and subscribe on Facebook, Twitter, or here on WordPress so you will always know when I post a new article. If you didn’t, let me know what I can do better in the comments down below. And join me next week for another game design article!
Read more about:
BlogsYou May Also Like
.png?width=300&auto=webp&quality=80&disable=upscale)
