Counting eliminations, catches, and deaths is better than relying on memory or reputation. But the same outcome can reflect very different opportunities, situations, and levels of difficulty. If the goal is to understand what an action was worth - or what it says about the player - we need to move beyond the box score.
A recorded result is produced by player ability and the circumstances surrounding it: playing time, position, role, balls received, teammates, opponents, game state, and tactical decisions. Better evaluation progressively separates those influences.
Four levels of evaluation
The progression can be summarized in four cumulative levels. Each solves a problem left unresolved by the level before it.
| Level | Main question | What it adds | What remains unresolved |
|---|---|---|---|
| 1. Record outcomes | What happened? | Creates an objective event record instead of relying on impressions. | Playing time, opportunity, role, context, and opposition. |
| 2. Normalize opportunity | How much was produced relative to the chances to act? | Accounts for court time, throws, attacks faced, or balls used. | The difficulty and importance of those opportunities. |
| 3. Value the context | How much did the action matter in that situation? | Recognizes that identical outcomes can change the expected result differently. | Who acted, who was affected, and whether the choice was good. |
| 4. Attribute contribution | What does the action reveal about the player? | Considers opponent value, difficulty, lineups, roles, and decision ownership. | Requires detailed data, larger samples, and stronger modeling. |
Record and normalize
Level 1 is essential. At its simplest, this is the level reached by most existing dodgeball-statistics software: recording throws, hits, and dodges, then calculating hit percentage or survival percentage, sometimes alongside basic totals such as eliminations and catches. That is more informative than saying that a player "looked good," but it still leaves major differences in exposure unresolved. A player who appears in every set has more chances to accumulate events than a substitute.
Level 2 introduces fairer denominators: points per minute, hits per throw, eliminations per targeted attack, or eliminations per ball committed. These separate total production from rate and resource efficiency. They also help account for role, because primary attackers and wingers may throw more often and may be forced into less favorable situations.
Opportunity adjustment remains incomplete. Two players can have the same number of throws while facing different defenders, attack types, and game states.
Value the situation
Level 3 asks what the event did to the game. An elimination at six-versus-one may change little because the set is almost decided; the same elimination at two-versus-two may sharply improve the team's expected result. A catch can remove a thrower, return a teammate, and alter possession, so its value also depends on the state.
This level evaluates the action before it evaluates the identity of the player. A technically exceptional throw in a hopeless match state may reveal skill while adding little outcome value. Event value and player skill are related, but they are not the same question.
Attribute the individual contribution
Level 4 asks what the event tells us about the people involved. Eliminating an elite defender is harder than eliminating a novice. Removing a skilled caller or playmaker may also be more valuable because it weakens the rest of the opposing lineup.
But the most important opponent is not automatically the best target. A valuable player may be so difficult or dangerous to attack that repeatedly spending several balls on them creates less expected benefit than removing easier opponents first.
Decision quality is not execution quality
One player may choose the best target and narrowly miss. Another may choose a poor target and succeed through exceptional execution. Raw outcomes reward the second action, even though the first decision may have been better.
This distinction also separates callers from throwers. In a planned attack, the caller may own the target selection and call decision while the throwers own the execution. In an improvised attack, one player may own both.
Each level needs the previous one
The levels are cumulative. Context cannot be valued without the underlying events; individual contribution cannot be separated without identities, roles, opponents, and decisions. The higher questions can only be asked when the necessary information was preserved.
The goal is not to stop counting eliminations, catches, and deaths. Those are the foundation. The goal is to understand how much value an action created, how difficult it was, and what it reveals about the players involved. Do not only count who eliminated whom; ask how much the action was worth and why.