Every week Big Brother airs, the same search spikes: who got eliminated from Big Brother? Recap sites answer it after the episode, and spoiler threads answer it earlier but leave you deciding which anonymous poster to believe. This page tracks a third kind of answer, a federally regulated exchange where people trade contracts on which houseguest leaves next, updated around the clock. Last Thursday it answered Kamu Kirk. This week it is looking at something stranger. Three of the 11 people left are on the block, and two of them are in the Head of Household's own seven-person alliance. The exchange has picked one of those two to go. We asked eight AI models to price the same week without letting them see a single quote, and they came back with a different name on top.
The Quick Answer
The most recent houseguest eliminated from Big Brother was Kamu Kirk, voted out 8-1 in last Thursday's live eviction after Rick Devens used a Diamond Power of Veto to pull Dee Valladares off the block and put Kamu up in her place. This week's eviction, Thursday, August 27, has Drew Campbell, Taylor Brown and Mallory Aurichio on the block after Head of Household Dee Valladares nominated three and LaTrice "Lala" Verrett won the veto and saved herself. Kalshi's contract prices, read Wednesday morning, make Mallory the most likely to go. Our eight-model panel, which never saw those prices, ranks Taylor first by a hair and calls it a coin flip between the two. The full board, the panel's reasoning, and the one season-long number that explains why nobody should be confident here are below.
Stay ahead of the markets.
Daily insights and expert picks on Kalshi, Polymarket, and what's moving markets.
Free forever. Unsubscribe anytime.
These are market prices and model estimates, not predictions of fact and not financial advice. New to this? Start with how prediction markets work.
How Three Of The Same Alliance Ended Up On The Block
The week began the way most Big Brother weeks begin, with the Head of Household picking targets. Dee Valladares won the competition, becoming the season's first houseguest to hold the job twice after opening the summer as the Week 1 Head of Household, and nominated Drew Campbell, LaTrice Verrett and Taylor Brown. Three nominations, not two, because of the BB Block Buster: the Head of Household names three, all three compete on eviction night, and the winner is pulled off the block before anyone votes. Only the two who remain face the house.
Then the veto rewrote the week. Lala Verrett won it, which was its own story, coming two weeks after she became at 58 the oldest houseguest ever to win the Wall competition. She used it on herself, and Dee needed a replacement.
The replacement is what turned this from an ordinary week into the one worth watching. Dee named Mallory Aurichio, a member of her own alliance. The seven-person group, called the Mosh Pit, is Dee, Rick Devens, Angela Murray, Drew Campbell, Barrett Pfeiffer, Melody Morris and Mallory herself, per TheThings' account of the ceremony. Two of the three people now on the block are inside it. The third, Taylor Brown, is not.
By Tuesday the feeds had a plan and a problem with it. EntertainmentNow reported that "Mallory has emerged as the likely eviction target," and in the same breath that "concerns about a possible last-minute vote flip remain, meaning Dee and her allies cannot necessarily consider the eviction settled." Feed watchers had Dee and Devens describing the move to each other as a blindside they would apologise for later, with only Drew and Barrett aware that Mallory was the real target. On the other side, Melody Morris was reassuring Mallory she had the votes to stay, counting herself, Barrett, Yash Patel and Haley Thogmartin, plus Lala if it came to that.
Two whip counts, both live, two days before a vote. That is the week the exchange has to price.
The Board: The Market's Favourite Is Not The Panel's
Here is the board as it stood Wednesday morning, roughly 36 hours before the episode. Prices are Kalshi's midpoint between the best bid and the best ask, in cents, where a contract pays $1 if that houseguest is eliminated. All quotes as of August 26, 2026, 9:25 a.m. ET.
| Houseguest | Status | Market Price | Yesterday | Panel Blend |
|---|---|---|---|---|
| Mallory Aurichio | On the block (replacement nominee) | 58.5¢ | 53¢ | 40.5% |
| Taylor Brown | On the block | 37.5¢ | 43¢ | 40.9% |
| Drew Campbell | On the block | 13.5¢ | 20¢ | 15.3% |
| Barrett Pfeiffer | Not nominated | 2.5¢ | 2¢ | 0.5% |
| Rick Devens | Not nominated | 2.5¢ | 1¢ | 0.5% |
| Yash Patel | Not nominated | 2.5¢ | 2¢ | 0.5% |
| Haley Thogmartin | Not nominated | 1.5¢ | 2¢ | 0.5% |
| LaTrice Verrett | Veto winner, safe | 1.5¢ | 1¢ | 0.5% |
| Melody Morris | Not nominated | 1.5¢ | 2¢ | 0.5% |
| Angela Murray | Not nominated | 1.5¢ | 2¢ | 0.5% |
| Dee Valladares | Head of Household | 1.5¢ | 1¢ | 0.5% |
One housekeeping note on that table, because it changes how you read it. The 11 prices add up to 124.5¢ rather than 100, and that is not a bookmaker's overround. These are 11 separate contracts rather than one market carved into slices, and the exchange lists them as such: each pays $1 on its own name, and the rules settle on elimination rather than on the vote, counting "voluntary withdrawal, disqualification, or producer removal" and stating plainly that "multiple simultaneous eliminations count." Nothing forces the column to sum to a dollar, so 58.5¢ on Mallory is already the market's read on Mallory, no arithmetic required, and the eight names quoted at the exchange's one-to-two-cent floor are the price of those non-vote branches rather than rounding error. Nobody off the block is worth exactly zero. If you want the board on the panel's basis instead, where 100 points are split across the whole house and exactly one person leaves, scale all 11 prices back to 100: the market reads Mallory 47%, Taylor 30% and Drew 11%, with the last 12% sitting on the eight names off the block.
Now the panel. We put the same week in front of eight AI models and showed them none of the prices above. What they got instead was a card assembled at run time: every contract's exact settlement language, every settled week of this season and who the exchange paid out on, the reconstructed board shape going into each of those evictions, the full Week 7 sequence from the feeds, this season's vote margins, and the Block Buster's format and past winners. Then each model distributed 100 points across the 11 names, read the other seven models' anonymised reasoning, and got one chance to revise.
| Seat | Taylor Brown | Mallory Aurichio | Drew Campbell |
|---|---|---|---|
| Fable | 41.0% | 39.5% | 15.5% |
| Opus | 39.8% | 47.6% | 10.2% |
| Sonnet | 37.1% | 43.7% | 15.7% |
| ChatGPT | 40.8% | 41.2% | 14.0% |
| Gemini | 42.7% | 38.2% | 15.1% |
| GLM | 42.0% | 38.0% | 16.8% |
| Kimi | 40.6% | 39.7% | 16.5% |
| DeepSeek | 44.0% | 42.5% | 9.5% |
| Panel Median | 40.9% | 40.5% | 15.3% |
Every seat on this panel is graded against real market settlements — records to date: Gemini 86% on 4,952 graded calls · GLM 81% on 2,544 graded calls · Kimi 83% on 2,534 graded calls · DeepSeek 80% on 2,583 graded calls. Recomputed daily; the full scoreboard is public.
These are model estimates, not predictions of fact and not financial advice. Every number in this piece gets graded in public once the market settles, on the full graded scoreboard. The medians are taken name by name, so they come to 100.5 rather than 100; the eight unnominated names sit between 0.45% and 0.5% and are shown rounded, which is why the column as printed reads 100.7.
The headline is the gap on Mallory. The market pays 58.5¢ for her to go; the panel's middle answer is 40.5%. That is 18 points of daylight on the board's most heavily traded name, and the gap survives the stricter comparison: on the scale-to-100 basis the panel itself used, the market still reads 47% against the panel's 40.5%. What that arithmetic does not capture is the more interesting disagreement, which is about the shape of the week rather than the size of one number. The exchange has Mallory clear of Taylor by 17 points. The panel has them tied. On the size of the gap, every seat that explained itself gave the same reason. Six of the eight arrived at the same two-stage arithmetic before they arrived at a number: the Block Buster is played first, so a nominee is only evicted if she loses a physical competition and then loses the vote. Mallory has already won one of these, in Week 3. Drew won Week 4. Taylor has never won one. Strip roughly a third off whoever the house intends to send home and, however clear the plan looks on the feeds, the panel did not put any nominee as high as 58%.
The season's own record points the same way. Every settled week of this season is on the exchange, and the board going into each eviction can be reconstructed from the hourly candles. The favourite has gone home four times in six real eviction weeks, so 4 for 6. But sort those six weeks by how confident the board was and the split is not close:
| Eviction-Eve Favourite | Weeks | Favourite evicted |
|---|---|---|
| Priced At 60¢ Or More | 4 | 4 of 4 |
| Priced Below 60¢ | 2 | 0 of 2 |
Ashley Trail at 71¢, Rome Seymour at 61¢, Jason De Puy at 60¢ and Lyric Medeiros at 61¢ all walked out the door. Then came the two coin-flip weeks. Haley Thogmartin went into Week 5's eviction as the 49¢ favourite and won the Block Buster, taking herself off the block before a vote was cast; Chuk Anyanwu went instead at 48¢. Haley went in again the next week as the 51¢ favourite, watched Angela Murray win the Block Buster, and survived the vote 8-1 as the house sent Kamu Kirk home from 42¢. The favourite has never lost above 60 and has never won below it. Mallory is at 58.5. Six weeks is a shape and not a law, and one model said so at length below, but it is the exchange's own history and it points exactly where the panel already went.
More live boards from the same panel: Big Brother 28 winner, where Dee Valladares leads the field at 23.5¢ while running this week's block · most Emmy wins, with Widow's Bay out front at 67¢ · Oscars Best Picture 2027, where The Odyssey sits at 57.5¢. Prices fetched August 26, 2026.
Hottest Prediction Markets Right Now
- 2028 Democratic presidential nominee · $194M traded
- 2027 Pro Football Champion · $72M traded
- 2028 U.S. Presidential Election winner? · $62M traded
- Pro Baseball Champion · $60M traded
- 2028 Republican presidential nominee · $60M traded
- Bitcoin price at the end of 2026 · $33M traded
Every market above links to our full AI model verdict; browse them all on the OddsShopper prediction markets hub, and see how every settled call actually scored on the full graded scoreboard.
Where The Panel Changed Its Mind
The revision round is where this gets interesting, because the models moved toward each other from opposite directions and one of them refused.
Fable came in hottest on the house plan, at 54% on Mallory, and cut it to 39.5% after reading the others.
"'Mallory is the likely target' is worth about 40, not 65. She sheds a third of her risk to a competition she has already won once, and the reported whip count leans her way." — Fable
Gemini moved the other way, from 27% up to 38.2%, having started by treating Taylor as the near-automatic casualty and finishing with more respect for the plan actually on the feeds.
"The biggest error in a naive reading is pricing Mallory near 70% simply because she is the reported target. Doing so entirely ignores that she is 35% likely to remove herself from the block before a single vote is cast, and that her counter-whip makes the vote itself genuinely competitive." — Gemini
And Opus went up, not down, ending as the only seat with Mallory clear of the field at 47.6%. Its reason for holding out is the sharpest thing any seat said all week, and it is aimed straight at the base-rate table above.
"Four-for-four above 60c is real. But 0-for-2 below 60c came at 49c and 51c — coin flips, expected hits ~1.0, observing zero is a routine 25% outcome. That is calibration noise, not proof favourites are fragile. Several forecasters used it to lift Taylor above Mallory; I think that over-reads two tosses and discards the card's headline fact, that the HOH has a target and this house stampedes." — Opus
That is the honest disagreement on this board. Two data points can look like a pattern and be nothing. The season's vote margins back Opus up: the six evictions went 14-0, 12-1, 12-0, 8-3, 9-1 and 8-1, so five of the six were decided by seven votes or more. Once this house decides, it does not split.
What Would Change The Panel's Mind
Every seat was asked what would move its number, and the answers converge on four checkable things. All of them are visible on the live feeds before Thursday's episode, which is what makes them worth watching rather than guessing at.
- A Locked WHIP Count Against Mallory. If Wednesday or Thursday feeds show Barrett Pfeiffer, Haley Thogmartin or Yash Patel committing firmly to Dee's plan, Melody's counter-count collapses and Mallory moves toward the mid-fifties, with Taylor under 30.
- The Blindside Leaking. If Melody's group works out that Mallory is the target and consolidates on Taylor instead, the board flips: Taylor toward 50, Mallory under 30. Several seats named this as the single largest swing available.
- The Block Buster Result Itself. The competition is played on the episode, before the vote. Whoever wins it is off the block before the vote, which takes almost all of their eviction risk with it and pushes it onto the other two. This is not a forecast that improves with information; it is a branch that closes.
- Anything The Settlement Rules Cover That Is Not A Vote. A withdrawal, a medical exit or a removal would lift one of the penny names off the floor and shrink all three nominees at once. The panel left roughly four points of the board on those eight names for exactly this reason.
Dark Horses The Panel Won't Dismiss
Eight houseguests are not on the block, and every seat still gave each of them something. It is worth understanding what those pennies actually represent, because they are not a vote count.
The contract's secondary rules resolve YES on "voluntary withdrawal, disqualification, or producer removal" as well as eviction, and they specify that multiple simultaneous eliminations count. That last clause matters more than it sounds this deep into a season, when double evictions start appearing. The exchange has also already proved the opposite branch is real: the board dated July 9 settled with every single contract at NO, a whole week in which nobody left. A model pricing these eight at zero would have been wrong about that week.
Rick Devens is the interesting name at the bottom. His contract is quoted at 2.5¢ and he is not nominated, safe, and holding the strongest position in the house. Yet the exchange's trade tape shows 28,394 contracts changing hands in his name on Tuesday alone, nearly three times the next-busiest name on the board that day, all of it at one or two cents. Before that reads as a signal, do the arithmetic the way a trader would: at a penny or two a contract, that entire pile is about $355 of premium. It is somebody buying cheap lottery tickets in bulk, not the house turning on Devens. Contract counts and dollars are different units, and on a board this cheap the count flatters itself.
Barrett Pfeiffer and Yash Patel sit at the same 2.5¢ for a more legible reason. Barrett is the swing vote both sides claim, named in Melody's count to save Mallory and reported on the feeds as sticking with Dee and Devens. Yash spent Tuesday trying and failing to recruit Barrett into a move against the returning players. Neither is in danger this week. Both are the reason the week is not settled.
The Story Around Taylor Brown That The Vote Will Not Price
The week has a second story running alongside the game, and it centres on the one nominee who is not in the Mosh Pit.
On August 22, Yahoo Entertainment reported that several former Big Brother houseguests had publicly criticised how Taylor Brown and LaTrice Verrett were being treated in the house. Season 27's Morgan Pope wrote that she was "deeply disturbed by the microaggressive behavior we're seeing in the Big Brother house this season." Season 4 winner Jun Song said Taylor and Lala had been "generous and gracious, feeding these [expletive] and keeping the house clean" only to be "painted as aggressive and angry." Jason De Puy, evicted in Week 3, questioned why certain houseguests "are allowed to be emotional, frustrated, loud or upset — while Lala and Taylor displaying those same emotions suddenly seems to be getting characterized very differently." Season 27 winner Ashley Hollis wrote: "Do I think Taylor's housemates are unfairly villainizing her? Yes."
The reporting traces the reaction to comments captured on the live feeds, including CinemaBlend's account of Drew Campbell saying he hoped Taylor and Lala would clean the house if they lost a competition, and telling Rick Devens he wanted Lala to be first in jury so she could "make us food." CBS has not issued a statement. We are not adjudicating any of it, and it is worth being precise about what it does and does not do to this week's number.
It does nothing directly. The people trading these contracts do not vote, and the eight houseguests who do vote cannot see any of it. A backlash outside the house has no mechanical path into Thursday's result.
What it does is describe, from outside, the same position the panel identified from inside the data. Taylor Brown is the only nominee outside the seven-person alliance, the only one with no reported campaign working to save her, and the name several models flagged as the one a lazy read of the week most underrates. Both the criticism and the model have noticed the same structural fact: this house has spent the week deciding what to do about its own, and Taylor is the person it was never really arguing about.
"Taylor Brown is the only nominee outside the Mosh Pit, has no reported whip count, no comp pedigree, and loses the vote almost by default whenever Mallory comes off the block. The name a naive reading most underrates is Taylor herself: anyone pricing 'the reported target' misses that she is exposed in two of three branches and defended by no one." — Kimi
When This Settles
| Contracts Close | Thursday, August 27, 11:59 p.m. ET |
| Expected Settlement | Friday, August 28, 10:00 a.m. ET |
| Settlement Source | The official program broadcast, website, verified social media or press releases, per the contract's own rules |
| Next Catalysts | Wednesday and Thursday live-feed vote counts; the BB Block Buster, played live on the episode; the vote itself, Thursday 8 p.m. ET on CBS |
| Following Week | Week 8 nominations, expected the weekend of August 29-30 |
The contract resolves on elimination as its rules define it, which is broader than the vote, and settlement happens on the exchange once the result is official. Prices on this page refresh through the news wave and carry the timestamp they were read at. When the story moves, the board gets re-scored, and once the vote is official this page carries the result and the panel's grade for it.
The panel's calls are on the record. Our 8-model AI panel puts a graded, public verdict on the season itself, not just this week, at the live model verdict on the Big Brother 28 winner.
More on this: Rams 49ers Receiving Yards: Who Got A Line · Big Brother 28 Winner Odds: Kalshi Sleeps On Kamu Kirk · NFL Coach Of The Year Odds: The Cowboys' Coach Everyone Wrote Off · Who Will Be Charged With A Federal Crime? Kalshi 2026 Odds · Big 12 Championship Odds 2026: The Panel Vs. Kalshi
The Bottom Line
The house has made a decision it has not finished making. Dee Valladares put a member of her own alliance on the block, let Rick Devens, Drew Campbell and Barrett Pfeiffer in on it, and left the other seven houseguests a day and a half to work it out. The exchange's answer to that, 58.5 cents on Mallory Aurichio, is the tidy version. The panel's answer is that a week with two live whip counts and a physical competition standing in front of the vote is not a 58-cent week, and the exchange's own six-week history says the same thing: every favourite above 60 cents this season has gone home, and neither favourite below it has.
If you check back Thursday night, the useful thing to watch is not who the feeds say is going. It is the Block Buster, because whoever wins it takes a third of this board off the table before a single vote is read.
