PROBST Forecast
Last updated: 2026-10-06
PROBST forecast for Survivor 51 Episode 3, made on October 6, 2026 from the data through Episode 2. The game has 19 players left, on 2 tribes. Eric Macksoud, Carter Krull, Ori Jean-Charles and Jenna Doore are the most likely to go home, at 10%, 8%, 8% and 8%. The episode most likely has 1 tribal council (84%). There is a 17% chance of 2 eliminations in the same episode. A tribe swap has a 14% chance this episode. Rob Antonson and Angelica Loblack (Jelly) hold idols, which lower their chances of going home.
Not for betting. PROBST knows only what the aired episodes have shown. The season was filmed months before it airs; the cast, the crew and anyone they told know the result, and leaks reach betting markets long before the edit does. A market price will be ahead of this forecast whenever there is inside information, which is most of the time. These numbers are for reading the game as it has been shown, not for wagering. What PROBST does not know.
Savu has a 55% chance to go to tribal council: worse challenge form this season.
| # | Player | If Savu votes | This episode | Raises the chance | Lowers the chance |
|---|---|---|---|---|---|
| 1 | Eric Macksoud | 17.9% | 9.8% | received votes last tribal, has received votes | |
| 2 | Carter Krull | 15.8% | 8.5% | less airtime | |
| 3 | Ori Jean-Charles | 15.1% | 8.1% | ||
| 4 | Cristian Chavez | 9.6% | 5.2% | less airtime | right side of last vote |
| 5 | Alexis Levine | 9.1% | 5.0% | right side of last vote, no votes last tribal | |
| 6 | Linnea Capobianco | 8.8% | 4.8% | right side of last vote, no votes last tribal | |
| 7 | Sharonda Cox | 8.8% | 4.8% | right side of last vote, no votes last tribal | |
| 8 | Kristin Flickinger | 8.6% | 4.7% | right side of last vote, no votes last tribal | |
| 9 | Rob Antonson, holds an idol | 6.2% | 3.4% | more airtime |
Toka has a 45% chance to go to tribal council: better challenge form this season.
| # | Player | If Toka votes | This episode | Raises the chance | Lowers the chance |
|---|---|---|---|---|---|
| 1 | Jenna Doore | 17.2% | 7.8% | received votes last tribal, has received votes | |
| 2 | Devin Way | 13.9% | 6.3% | wrong side of last vote | |
| 3 | Lewis Kelly | 9.9% | 4.5% | weak co-voting ties | right side of last vote |
| 4 | Thien An Nguyen (Thien An) | 9.0% | 4.2% | right side of last vote | |
| 5 | Mike Pinsky | 9.0% | 4.2% | right side of last vote | |
| 6 | Maggie Nestor | 8.9% | 4.1% | right side of last vote | |
| 7 | Patt Cannaday | 8.8% | 4.1% | right side of last vote | |
| 8 | Kilby Kilby | 8.5% | 3.9% | right side of last vote | |
| 9 | Brady Booker | 8.4% | 3.9% | right side of last vote | |
| 10 | Angelica Loblack (Jelly), holds an idol | 6.3% | 2.9% | wrong side of last vote |
"This episode" is the chance of going home across every tribe, a swap and an early merge. The column before it is the chance of going home if that tribe goes to tribal council. The last 2 columns name the features that move the number most, relative to the tribe. The model uses the voting record, ratings, airtime, idols and the co-voting graph. It does not use anything shown in the episode.
| Event | Chance |
|---|---|
| A tribe swap | 14% |
| The merge | 1% |
| 1 tribal council | 84% |
| 2 or more tribal councils | 12% |
| 2 eliminations | 17% |
| Episode | Chance of the merge |
|---|---|
| 3 | 1% |
| 4 | 1% |
| 5 | 3% |
| 6 | 55% |
| 7 | 28% |
| 8 | 7% |
| 9 | 1% |
| 10 | 1% |
| 11 | 1% |
Episode of the merge, across the simulated seasons.
Ana Sani went home in Episode 2. The model had Ana 14th of 19 at 4.9%. Its first pick was Jenna Doore at 10%.
The rest of the season is simulated 10,000 times from the state after Episode 2: who wins the challenges, who finds and plays idols, who the votes land on, when the merge comes. Angelica Loblack (Jelly) and Rob Antonson have the best odds at 9% and 9%. With 19 players left, an even share is 5.3%.
| Player | Tribe | Out next | Makes merge | Final 3 | Wins | Avg place |
|---|---|---|---|---|---|---|
| Angelica Loblack (Jelly), holds an idol | Toka | 3.5% | 86.4% | 26.1% | 8.9% | 8.0 |
| Rob Antonson, holds an idol | Savu | 4.3% | 84.2% | 25.2% | 8.6% | 8.3 |
| Devin Way | Toka | 7.6% | 77.2% | 20.6% | 7.4% | 9.6 |
| Lewis Kelly | Toka | 5.3% | 80.1% | 20.7% | 7.1% | 9.4 |
| Brady Booker | Toka | 4.6% | 80.6% | 18.2% | 6.3% | 9.6 |
| Kilby Kilby | Toka | 4.8% | 79.5% | 17.0% | 5.4% | 9.8 |
| Kristin Flickinger | Savu | 5.8% | 77.3% | 14.9% | 5.1% | 10.1 |
| Ori Jean-Charles | Savu | 9.2% | 72.5% | 14.4% | 4.7% | 10.6 |
| Linnea Capobianco | Savu | 6.1% | 77.1% | 13.4% | 4.7% | 10.3 |
| Sharonda Cox | Savu | 5.4% | 77.1% | 13.3% | 4.5% | 10.3 |
| Patt Cannaday | Toka | 4.9% | 78.8% | 13.7% | 4.5% | 10.1 |
| Eric Macksoud | Savu | 10.5% | 70.3% | 13.8% | 4.5% | 10.9 |
| Carter Krull | Savu | 9.3% | 71.3% | 12.8% | 4.3% | 10.9 |
| Alexis Levine | Savu | 5.8% | 76.4% | 12.8% | 4.2% | 10.4 |
| Jenna Doore | Toka | 8.7% | 73.9% | 12.6% | 4.2% | 10.7 |
| Maggie Nestor | Toka | 5.3% | 78.0% | 12.6% | 4.2% | 10.3 |
| Thien An Nguyen (Thien An) | Toka | 5.0% | 78.8% | 12.8% | 4.1% | 10.2 |
| Mike Pinsky | Toka | 4.6% | 78.7% | 13.4% | 4.1% | 10.2 |
| Cristian Chavez | Savu | 5.8% | 75.2% | 11.8% | 3.3% | 10.7 |
Avg place is the mean finishing position across the simulations.
| Event | Chance or count |
|---|---|
| Fire-making at the final 4 | 50% |
| A swap before the merge | 62% |
| Idols still to be found | 3.4 |
| Tribal councils where an idol changes who leaves | 6% |
| Episodes left | 11 |
Tribes, who is left, challenge ratings and form, and idols found and played all move in the simulation. The voting record and airtime stay as they are today. The weight on today's situation is halved each simulated week.
Across 40 past seasons, from the merge the eventual winner was in the model's top 5 60% of the time (43% by chance). Players given 13% won 11% of the time; players given 4.4% won 2.3%.
Co-voting through Episode 2, over 2 tribal councils, among the players still in the game.
No pair of players still in the game has voted together at 2 tribal councils yet.
Blocs are Louvain communities on the co-voting graph, the same method as the Allies section on player pages. Betweenness is how often the shortest path between 2 other players runs through this one. The forecast takes 5 of its features from this graph: the share of players a person has voted with, the strength of those ties, bloc size, membership of the largest bloc, and betweenness.
Each forecast was made before the episode aired, with the format and the immunity winner also forecast.
So far 2 players have gone home at tribal council. The model's first pick was right for 0 of the 2, and the player who left was in its top 3 picks for 1 of the 2. Picking at random from everyone still in the game would be right 5% and 15% of the time. The forecasts are least accurate before the merge, when many players are in the game and few have voted.
| Episode | Model's top 3 | Who went home | Model had them |
|---|---|---|---|
| 2 | Jenna Doore 10%, Angelica Loblack (Jelly) 8%, Devin Way 7% | Ana Sani | 14th of 19, 4.9% |
| 1 | Linnea Capobianco 5%, Aaliyah Puglia 5%, Thien An Nguyen (Thien An) 5% | Aaliyah Puglia | 2nd of 20, 5.0% |
Rank is among everyone still in the game, with the odds of each tribe going to tribal council included. Each week's forecast is rebuilt from the data on every refresh; this table is not edited after the fact.
Over 40 past seasons (584 boots, seasons 11 to 50, every week forecast with models trained on earlier seasons only), the first pick was right 18% of the time (11% by chance), the player who left was in the top 3 45% of the time (32%) and in the top 5 64% of the time (52%).
PROBST (PRobabilistic Odds for Boots, Swaps and Tribals), updated October 6, 2026 with data through Episode 2. Forecasts use only what had happened by then: ratings, votes, confessionals, idols shown, and the co-voting record. Models trained on 50 seasons. Method and backtests.