Who got kicked off Dancing With the Stars is determined by a weighted mix of judge scores and viewer votes, with the lowest combined total eliminated — and in the show's newer seasons, judges can also save a dancer in the bottom two. The names change every season and every week, so no static list stays accurate. What stays stable is the machinery: how scores convert to percentages, how vote totals are ranked, and why a strong dancer can still go home. That machinery is what this explainer covers.
The scoring-to-ranking pipeline
Each couple performs, and the judges award paddle scores. On the main ABC run those paddles were typically scored out of 10 per judge, and with four judges a perfect night was 40 points (the show used three judges for stretches of its earlier seasons, making 30 the ceiling). Those raw judge points are not used directly as vote weight. The show converts each couple's judge total into a percentage of the night's overall judge points, then converts viewer votes into a percentage of total votes cast, then combines the two percentages.
The reason for percentage conversion is arithmetic fairness. Raw vote counts can run into the millions while judge points top out at 40. Adding those numbers directly would make the vote meaningless or the judges irrelevant. Converting both to shares of their own totals puts them on one scale, and the two shares are then blended in a fixed ratio. The couple with the lowest blended share leaves.
A worked example makes the mechanism concrete. Suppose three couples remain and the judges award 40, 36, and 30 points, a night total of 106. Their judge shares are about 37.7%, 34.0%, and 28.3%. If the 40-point couple draws only 10% of viewer votes while the 30-point couple draws 55%, the blend can flip the ranking entirely. This is why the leaderboard after the judges perform is not the elimination order. It is a provisional ranking that the vote can overturn.
Quantified context: where the vote weight actually bites
The published split has varied by season and platform, but the structural point holds across eras: the viewer share is large enough that a judge leader can be eliminated and small enough that a judge trailer rarely survives on votes alone. When the vote share is described as roughly half the decision, a couple at the bottom of the judges' board needs an enormous vote margin to climb out. Moving from last to safe against a mid-table couple typically requires the vote share to swing by double digits in percentage points.
Turnout scale matters here. A season finale vote can run into the millions of combined votes across phone, text, and online channels, and those totals are large enough that a few thousand votes from one fan campaign do not move the share much. Early-season votes are smaller in total, which makes organized fan blocs more potent in week two than in the finale. That is a season-progression effect, not a rule change.
Judge score compression is the other quantified force. When every couple scores 8 or higher, the spread between first and last place might be 6 points out of a night total near 150 — under 5% of the judge pool. In that state, the judge component barely separates couples, and the vote share dominates. Early weeks, when scores range wider, give the judges more leverage. Same formula, very different effective weight.
The bottom-two twist and judge saves
Newer seasons added a structural layer: the two couples with the lowest combined totals become the bottom two, and the judges then vote on who stays. This changes the mechanism from a pure calculation to a hybrid. The math still decides who lands in the bottom two. The judges then override the combined result for one of them.
The save is not unlimited. Judges typically must choose by majority, and when they cannot agree or when the rules specify it, the couple with the lower combined total goes home. This is the counterintuitive detail worth internalizing: a couple can lose the combined vote and still survive, because the final step is a discretionary vote rather than a continuation of the math. Conversely, the couple that "won" the bottom two on points can be sent home by the panel. The elimination is therefore partly algorithmic and partly editorial, and viewers who track only numbers will mispredict save nights.
What the formula does not capture
Three edge cases break naive predictions. First, ties in the combined percentage are resolved by rules most viewers never see, and tie handling has shifted across seasons. Second, withdrawals and injuries remove couples without an elimination vote, so the count of "kicked off" dancers in a season is not the count of elimination ceremonies. Third, the judges' save can produce a bottom two in which neither couple had the lowest raw vote — the vote share only determines the pairing, not the final exit.
A judge leaderboard-topper can be eliminated in the same week — judge share is a component, not a guarantee. A couple can survive the bottom two while having the lower combined total — the save is discretionary. Heavy early-season voting by one fan base can distort week-two results far more than finale results. Withdrawal and injury exits inflate "kicked off" counts and are not elimination votes. Score compression in late weeks shrinks the judge component to near noise.
Professional caveat: the exact percentages and tiebreak rules have been revised between seasons and platforms, and the show does not publish a full vote audit. Any specific ratio quoted as universal is an approximation of one era, not a permanent law. Treat precise percentages as season-specific unless a current official source confirms them.
Comparison table: what each input actually controls
| Factor | What it sets | Where it stops mattering | Typical magnitude |
|---|---|---|---|
| Judge paddle scores | Raw points per couple per night | Converted to a share, so raw points alone cannot decide | 0-10 per judge; 30-40 night ceilings |
| Judge percentage share | Judge half of the blended total | Capped by night total; compresses when scores cluster | Roughly a third to half of blend |
| Viewer vote count | Raw popularity signal | Converted to a share of total votes | Millions in finales |
| Viewer percentage share | Vote half of the blended total | Diluted as total turnout grows | Roughly half of blend |
| Bottom-two selection | Names the two at risk | Does not itself eliminate either | Two couples |
| Judge save vote | Picks the survivor of the bottom two | Cannot rescue a couple outside the bottom two | Majority of panel |
The table's lesson is that no single row is decisive. The decision emerges from the interaction of the score share, the vote share, and the save step, in that order.
Timeline table: a typical elimination week
| Stage | What happens | What it tells you |
|---|---|---|
| Rehearsal week | Couples prepare dances; running order set | Nothing about elimination yet |
| Performance show | Judges score each couple live | Provisional judge leaderboard forms |
| Vote window opens | Viewers vote by phone, text, or online | Turnout starts accumulating |
| Vote window closes | Voting cut off at a fixed time | Final vote totals locked |
| Results night | Combined totals computed; bottom two named | The math has spoken; save pending |
| Judge save | Panel votes on the bottom two | One couple survives, one exits |
| Elimination announced | Departing couple confirmed | Season count of exits increments |
The timing detail that trips people up is the vote window. Votes cast after the cutoff do not count, and the cutoff is announced on air, not in advance on a fixed clock. A late surge during the results broadcast cannot rescue anyone.
Practical actions: predicting and reading eliminations
Track the judge share, not the paddle total. Divide each couple's score by the night's total judge points to get the share that actually enters the blend. A 40 out of 40 night in a compressed field is worth less than it looks.
Watch turnout as a multiplier. In early weeks, when total votes are lower, a coordinated fan push moves shares more. In late weeks, the same push is diluted. This is the single most useful adjustment for reading a surprising result.
Before results night, rank couples by judge share and flag the bottom third as at-risk. Estimate vote strength from social engagement volume, then discount it in later weeks as turnout grows. On bottom-two nights, treat the judges' save as a separate prediction from the math. Exclude injury and withdrawal exits from any "kicked off" count you compile. Treat quoted vote-weight percentages as era-specific, not permanent. Distinguish the provisional judge leaderboard from the final combined ranking when discussing who is in danger.
On reconciliation of thresholds: some commentary describes the vote as roughly half the decision, some describes judge scores as the tiebreaker of record, and some describes the save as decisive. All three are true at different steps. The vote and judge shares set the blended order; ties and the save operate after that order exists. A prediction that applies one threshold to all three steps will fail. Apply each threshold only to the step it governs.
Because season-specific cast names, current-week results, and changing vote ratios are not stable facts, this article deliberately stops at mechanism. For the current season's actual elimination list, consult the show's official results and the network's published recaps for that week; any explainer that freezes names into a permanent answer will be wrong by the next episode.
Limitations
No public audit of raw vote totals exists, so nobody outside the production can verify a specific elimination's vote margin. The scoring ceilings, vote weights, tiebreak rules, and save mechanics have all changed across seasons and platforms, which means historical comparisons of "unfair" eliminations often compare different systems. Fan-engagement signals are a proxy for vote share, not the vote itself, and they overstate organized blocs relative to casual viewers. Finally, a count of dancers "kicked off" is not a count of eliminations once withdrawals and injuries are included, and conflating the two produces inflated season totals.
This article is general educational information about how a television competition's elimination format works. It is not a prediction service, not a statement about any current season's cast or results, and not a substitute for the show's official results for the week in question. For the current elimination list, check the network's official recaps.