· 6 min read
Beat battle judging: the criteria, and a scoring sheet worth using
What quality, originality and mix actually mean when you are scoring an entry, why category scoring flattens results, and the sheet to use instead.
Every battle publishes roughly the same criteria — quality, creativity, originality, mix, crowd reaction — and almost none of them say what the words mean. Two judges using the same list can score the same entry three points apart and both be applying the rules correctly. This is what those words are usually being used for, and what to score instead.
The criteria, and what they are really asking
| Criterion | The question underneath it | Where it goes wrong |
|---|---|---|
| Quality | Does this sound like somebody meant it? | Gets read as production budget. An entry made with four sounds can be higher quality than one made with forty. |
| Creativity | Did they do something you had not already heard tonight? | Rewards weirdness for its own sake if nobody checks that it still works. |
| Originality | Is this theirs? | Confused with novelty. A familiar pattern played well is original; a strange one copied is not. |
| Arrangement | Does it go somewhere in the time it has? | Unscoreable on a short entry. Eight bars cannot have an arc, so do not ask it to. |
| Mix | Can you hear everything that is meant to be there? | The most over-weighted category in any timed battle — see below. |
Do not score the mix in a timed battle
In a four-minute round nobody is mixing. What you are actually scoring is whose default sounds happened to sit together, which is a fact about their kit rather than about them. In an async battle with a week on the clock the mix is fair game, because everybody had the same week.
Why category scoring flattens the result
Splitting a score into four categories sounds fairer and is not. An entry that is competent in every category beats an entry that is remarkable in one and thin in the rest — so the format systematically rewards the safest thing in the room. Interesting is what everybody came for, and a scoring system that punishes it has quietly changed the game.
One number for the whole entry keeps the judgement where it belongs: on what you just heard, as a whole, the way anybody actually listens to music.
The sheet
- One score per entry, out of five, in half-steps. Nothing else.
- Score immediately after the entry finishes, before the next one starts.
- No going back to adjust once the next entry has played. Your memory of entry two after hearing entry six is not evidence.
- Everybody scores every entry. If self-scoring is allowed, say so before the first round and keep it the same all night.
- Scores stay private until every entry has been played.
Half-stars matter more than they look
A five-point scale sounds like enough and is not. In a room of near-equals the interesting judgement is almost always a bit better than that one, and whole stars cannot express it — so everything lands on three and four and the result is decided by whoever happened to round up. Half-steps give you ten usable values and cost nothing.
Judge the entry, not the producer
The single biggest scoring error is knowing whose entry it is. It is not dishonesty; people cannot help it. Play everything back anonymously, in a random order, and reveal the names once the scores are locked. The format piece covers the other three ways a battle goes wrong.
Four minutes is shorter than it sounds.
Play a round