The question "who was eliminated on Dancing with the Stars" cannot be answered with one permanent name, because eliminations change every week and every season. What is stable is the mechanism: a contestant leaves when the combined judge-score and viewer-vote result places them at the bottom, and the show converts two very different measurements into one survival decision. Understanding that conversion—percentages, ties, and vote weighting—explains why the "wrong" person sometimes goes home.
How a dance score becomes a survival percentage
The core problem is that judges and viewers measure different things. Judges award points, typically on a scale where each pair receives a score out of 10 per judge. Viewers cast votes in raw counts. Those units cannot be added, so the show converts each into a share of its own total.
Here is the mechanism step by step. First, each couple's judge points are summed across the panel. Second, that sum is divided by the total points awarded to all couples that week, producing a judge percentage. Third, each couple's raw votes are divided by all votes cast, producing a vote percentage. Fourth, the two percentages are combined, usually at equal weight, and the lowest combined total is eliminated.
The key insight is that the conversion normalizes away scale. A couple with 24 out of 30 judge points and a couple with 30 out of 30 both become fractions of the night's total. This means the gap between first and last in judge points shrinks after conversion. On a three-judge panel scoring 10 couples, the judge spread might run from roughly 5 percent to 13 percent of total points. The vote spread can run from near zero to well over 30 percent. Votes dominate the arithmetic even when the weights look equal.
Plain language: the judges set a narrow band, and the voters decide where inside that band you land.
The arithmetic of vote weighting, in numbers
Consider a simplified week with five couples. Judges award points that sum to 100 across the field. Couple A earns 25 judge points, which is 25 percent of judge points. Couple A also draws 10 percent of viewer votes. Combined at 50/50, Couple A scores 17.5.
Couple B earns 15 judge points, or 15 percent, but draws 35 percent of votes. Combined, Couple B scores 25. Couple B survives comfortably despite a weaker dance.
Now push Couple A to the bottom of the vote at 5 percent. Combined score is 15. Under this arithmetic, a judge favorite with a small, unenthusiastic fan base can fall below a judge-weak couple with a mobilized audience. This is not a flaw in the math; it is the math working as designed.
The practical threshold matters. In a five-couple week, the eliminated couple is the minimum combined score. In seasons with a save or judges' decision, the bottom two combined scores enter a separate resolution step. The exact weighting has varied across seasons and is set by producers, so treat any specific percentage split as season-dependent rather than universal.
Why ties and score compression create edge cases
Edge cases cluster around compression. When many couples receive identical or near-identical judge scores, the judge percentage stops discriminating between them. The vote percentage then carries almost the entire decision.
A tie at the top of the judge leaderboard is mathematically interesting. If two couples share the highest point total, they share the highest judge percentage, and their combined scores differ only by votes. A single percentage point of vote share can decide placement.
There is a counterintuitive detail here that even engaged viewers miss. Being the judge leaderboard's top scorer does not protect you if your vote share is genuinely low, because the leaderboard advantage is measured in a few percentage points while vote deficits can run in the tens. Conversely, the couple in last place with judges is not safe either, because a strong vote share can lift them above a middling couple with a weak fan base.
Professional caveat: the show does not publish exact vote totals or the precise combination formula in real time. Any public reconstruction of "who should have gone home" is an inference from judge scores and reported outcomes, not a verified calculation. Treat fan-made elimination models as estimates.
The scoring system compared across formats
Different competition formats convert judge and audience input differently, which changes who survives.
| Format | Judge input | Audience input | Conversion method | Elimination rule |
|---|---|---|---|---|
| DWTS-style combined | Panel points | Raw vote counts | Both converted to percentages, then weighted | Lowest combined score |
| Rank-based scoring | Ordinal ranking | Ordinal ranking | Ranks summed, lowest total wins | Highest rank sum eliminated |
| Pure audience vote | None or advisory | Raw vote counts | Votes only | Lowest vote share |
| Judge-only | Panel points | None | Points only | Lowest point total |
| Save-plus-vote | Panel points | Raw votes | Combined, then judges choose among bottom | Bottom two resolved by panel |
The distinction between percentage-combination and rank-combination is the most consequential. Rank methods compress outliers; percentage methods let a large vote margin swamp a small judge margin. DWTS-style formats lean toward percentage combination, which is why fan mobilization matters so much.
What the elimination timeline looks like across a season
Eliminations follow a predictable cadence rather than a single event. The time course below is a typical structure, not a fixed schedule.
| Stage | Timing | What decides it | Key risk for contestants |
|---|---|---|---|
| Premiere weeks | Weeks 1-2 | First combined scores; no prior fan base | Low name recognition |
| Early middle | Weeks 3-5 | Vote base forms; ties common | Judge compression |
| Midseason | Weeks 6-8 | Fan bases mature; double eliminations possible | Split vote nights |
| Late season | Weeks 9-11 | Small field; every vote share counts | Judge leaderboard gaps widen |
| Finale approach | Final weeks | Combined scores plus format-specific saves | Bottom-two resolution rules |
The early weeks carry the highest variance because fan bases have not stabilized. A contestant with a strong week-one dance but no established audience can still land in the bottom. By midseason, vote patterns harden, and the judge percentage becomes a narrower lever.
Common misconceptions about who goes home
Misconception one: the worst dancer always leaves. False. The eliminated couple is the lowest combined score, and votes can outweigh dance quality.
Misconception two: the judge leaderboard is safe. False, as the arithmetic above shows. A top judge score with a weak vote share can still fall.
Misconception three: voting is pointless if your favorite is mid-table. Also false. Mid-table couples are exactly where a vote surge changes the outcome, because the judge percentage gap is small.
Misconception four: the show reveals exact vote counts. Generally it does not during the season, which limits any outside verification of the formula.
Red-flag warnings for anyone reconstructing results:
Do not assume the judge-to-vote weight is fixed across seasons; it has varied. Do not treat reported percentages as official unless the show published them. Do not compare raw judge points across weeks with different panel sizes or dance counts. Do not model ties as rare; compressed scoring makes them common. Do not assume a saved couple was the higher combined score; save rules can override it.
Practical actions for tracking eliminations accurately
If you want to predict or verify eliminations rather than react to them, use a repeatable method.
Step one: record each couple's per-judge scores and compute the weekly sum before conversion.
Step two: estimate each couple's judge percentage by dividing by the total points awarded that night.
Step three: treat vote share as unknown but bounded, and reason about how large a vote swing would be needed to flip the bottom position.
Step four: flag weeks with identical judge scores, because those weeks are vote-decided.
Step five: check whether the season uses a save or bottom-two rule before calling any result.
A useful threshold: if the judge percentage gap between the bottom couple and the next couple is under about two percentage points, the outcome is effectively a vote contest. Above roughly five points, the vote would need a large swing to matter. Reconcile these two thresholds explicitly: small judge gaps mean votes decide; large judge gaps mean judges set a floor the vote must overcome.
Limits of what can be concluded
The mechanism is stable, but the parameters are not public. Exact weighting, vote totals, and tie-breaking rules are producer-set and season-specific. That means any answer to "who was eliminated" is a record of an outcome, not a derivation you can fully audit.
The deeper limit is measurement. Judge scores are observable; vote shares are not. Every reconstruction fills that gap with assumptions. When assumptions are stated, the model is useful. When they are hidden, the model is guessing.
This article is general educational information about how competition scoring systems combine judge and audience input. It is not a prediction, an official result, or a verified calculation of any season's vote totals. For confirmed elimination results, consult the show's official broadcasts and published recaps.