The problem ratings try to solve
If you want to know who is better at a game, the simplest answer is to make them play. But with thousands or millions of players, everyone cannot play everyone else. A rating system solves this by giving each player a number and using the results of the matches that do happen to adjust the numbers, so that the whole population ends up roughly ordered by skill.
A ranked ladder is the public face of that process: a visible list or set of tiers where your position reflects your recent results.
The basic Elo idea
Many systems are based on a method originally devised for chess and named after its creator, Arpad Elo. The core idea is simple and has two parts.
First, the system predicts the result. If two players have the same rating, each is expected to win about half the time. If one has a clearly higher rating, that player is expected to win more often, and the bigger the gap, the more lopsided the expectation.
Second, after the match, the system compares the actual result with the prediction and adjusts. Winning when you were expected to win raises your rating a little. Winning when you were expected to lose raises it more. Losing as the favourite costs more than losing as the underdog.
A simple illustration with made-up numbers: suppose two evenly matched players each have 1500 and the system moves ratings by up to 20 points in a game. The winner would gain 10 and the loser would lose 10, because each was expected to win half the time and the swing is shared. If the favourite wins as expected, they gain only a small amount, and the underdog loses only a little. The exact size of the adjustment is a setting chosen by the designers, and real games use different values.
An important feature is that points move between players, so the total stays roughly the same. This keeps the scale stable over time.
Why new players move quickly
The system starts with very little information about a new player. Early results therefore count for more, and many games use bigger adjustments or special placement matches at the start. As more matches are recorded, the system becomes more confident and moves your rating by smaller steps.
Some systems track this uncertainty explicitly, keeping a second number for how sure they are about your rating. Several well-known approaches do this, and the idea is straightforward: a rating based on ten games should be trusted less than one based on a thousand.
Tiers, divisions and hidden numbers
Players rarely see raw ratings. Many games translate them into named tiers or divisions, and a rank can be divided into smaller steps. This is easier to read and gives players visible milestones.
Some games also keep a hidden rating used for matching players, separate from the visible rank. This can make the displayed rank and the experienced difficulty differ. The connection between ratings and matchmaking is explained in the article on skill-based matchmaking.
Team ladders and how they differ
When a team plays, the system has to decide how to credit each person. Some games average the ratings of players on a team and compare the averages. Others rate players individually and adjust each based on the result. Teams of friends of different skill, or very small groups mixed with strangers, make estimates harder, which is one reason team ratings are often noisier than solo ones.
What ratings can and cannot tell you
A rating is an estimate, and estimates have limits.
- Short-term swings are mostly noise. A winning or losing streak can happen even when your skill has not changed.
- Different ladders are not comparable. A rank in one game, or even one season or region, does not translate to another.
- Performance is more than results. Ratings usually consider only win or loss, not how well you played or what role you had.
- Ladders can drift. If the player base changes, the typical rank can shift, so a tier name means different things at different times.
- Cheating and abuse distort the picture. Shared accounts, boosting and unfair play can corrupt a rating, which is partly why anti-cheat software matters.
Ratings are also affected by the connection quality of the match, since high ping time or other issues can influence results.
Seasons and resets
Many ladders run in seasons. At the end, ratings may be partly or fully reset, so that players have something to climb again. The details differ, and resets can help the ladder stay lively while making a rank less permanent than it may feel.
Where ratings meet organized play
Ladder ratings are sometimes used to decide who can enter qualifiers or who receives an invitation. Otherwise, organized tournaments use brackets, and leagues use league seasons, to decide winners. The ladder is the open, always-running layer beneath those formats, as outlined in the overview of competitive play.
Keeping it in perspective
A rank can be a fun goal, but it is one imperfect measure of how well you play. Frustration over a number is common and understandable, and it can help to remember that the number is a statistical guess rather than a verdict.