What I would love to see is they stock the bag at the start of the game with 3S/15NS (for an 18 player season) and not restocked for the entire game so the odds actually change with each play. With each 'not safe' the sitd becomes more threatening.

If you know of a post/thread/clip that has looked into this, link it here. I'm curious what they said.

Hope this made sense. It's something better explained in person.

I'll share the code with anyone who asks. Cheers.

11/

You could think of a bunch of reasons why the 1st juror voting for the winner is above avg but whatever logic fits the narrative also needs to hold for the 4th juror but in the opposite direction, and let's be honest, it all seems fairly arbitrary.

10/

The 4th is interesting because it's in the other direction - they have voted for the winner 4 fewer times.

So, all up there's not a lot of evidence to say there is something special about the 1st juror. These results are saying that what we've observed is pretty ordinary.

9/

Here's that table again. I have repeated this for the 1st-8th juror. Half are above the mean and half below, or exact. But you can see all lie within the 95% CI. The 1st and 4th jurors are the most significant ones.

8/

While this has happened more times than expect you can see from the distribution 40 is within the range we would expect. If 40 was in the tails, there may be something there.

7/

Across 48 seasons (excluding S32 where the juror was removed), if there was no 1st juror effect, these are the values we can expect (see chart).

Median: 36
Observed: 40

We have seen the 1st juror vote with the group who voted for the winner in 4 more seasons than expected.

6/

There are a few ways to look at this, the easiest is simulation.

Method:
- Simulate the number of votes received to win e.g. when there are 7 jurors, a winner can receive 4-7 votes. This is based on the history (below).
- Simulate if the 1st juror was part of this group.
- Do this 4000 times

5/

These are all dependent relationships we need to control for.

What is independent? Each season. We also need to consider the number of votes a winner receives.

So a better question is: Is the 1st juror part of the GROUP that voted for the winner?

4/

The trick with this analysis is:
- A juror can only vote for one finalist and not the others
- The winner is decided by the jurors votes, who receives the most wins
- Jurors tend to vote as a group. It's an incorrect assumption to treat a juror's decision as independent

3/

To start off with, there have been 48 seasons where the 1st juror voted (excluding S32 where the juror was removed).

40/48 seasons (83%) where the 1st juror voted for the winner

Seems like a lot, but is it a significant difference from what we would expect? Or within variation?

2/

Someone reached out to me for an analysis on whether or not the 1st jury member is more likely to vote for the winner. I understand there was some discussion about this, but I haven't seen/heard anything, so coming in fresh.

Here's what the numbers mean 🧵 1/

#Survivor #Survivor50

At the conclusion of the season I am going to include a tribal_council_id to handle these cases.

Let's just hope this doesn't happen again though.

8/

Having two tribals maintains consistency. The only problem is any code you may have will count this as two tribal councils when really it was only one, but that's a trade-off I'm very willing to make. The alternative is way worse.

7/

The reason for this is recording the vote as 'Chrissy and Coach' is code breaking. It makes it difficult to count how many votes someone received, mapping vote_history to boot_order, castaways, boot_mapping, and challenge_results.

6/

So, for the vote_history table I have recorded two separate tribal councils that occurred simultaneously. The first one voted Chrissy out and the second Coach.

5/

The boot order, for reasons unknown Chrissy had her torched snuffed first so she is considered the 11th voted out followed by Coach, 12th voted out, even though they were voted out at the same time.

4/

Cirie went to Exile Island but immediately returned, and her activity there was more like a Journey. So, I've recorded it as going on a Journey. The penalty for not completing the challenge is Exile. Completing the challenge, return to the game with immunity.

3/

The Immunity and Reward challenge is recorded as a Team challenge won by Joe and Tiffany, not individual immunity. I don't know if the other records are going to treat it as Individual Immunity, but I think it's more consistent for it to be considered a Team Immunity.

2/

Back from a few nights camping and caught up with the latest episode of #Survivor50 - Double the Demise, Half the Fun... or something like that.

The game/twists aside, this episode sucks from a data perspective. Data is up, and here are some release notes 🧵 1/

💾 github.com/doehm/survivoR

Complete #SurvivorAU Redemption (AU12) data has been delivered. As per usual, any corrections needed, please log an issue on Git and I will correct them asap.

💾 github.com/doehm/survivoR

#Survivor #Rstats

Here are the final votes with the jury. It almost played out as expected. Only Cameron voted differently.

#SurvivorAU

All models are wrong, but some are useful, and this might be turning out to be one of the useful ones #SurvivorAU

Need to review in a week or so, but #SurvivorAU Redemption is the 5th highest rated (IMDb) season and higher than all US seasons.

𝟷 𝚂𝟷𝟷 𝟿.0𝟹
𝟸 𝚂0𝟿 𝟾.𝟼𝟾
𝟹 𝚂𝟷0 𝟾.𝟼𝟽
𝟺 𝚂0𝟻 𝟾.𝟻𝟾
𝟻 𝚂𝟷𝟸 𝟾.𝟻𝟷 👈
𝟼 𝚂0𝟺 𝟾.𝟺𝟹
𝟽 𝚂0𝟽 𝟾.𝟺𝟸
𝟾 𝚂0𝟾 𝟾.𝟸𝟿
𝟿 𝚂0𝟹 𝟾.𝟸𝟻
𝟷0 𝚂0𝟼 𝟾.𝟸𝟺
𝟷𝟷 𝚂0𝟸 𝟾.𝟷𝟾
𝟷𝟸 𝚂0𝟷 𝟽.𝟽

A live reunion with a theatrical vote delivery was exactly what I needed. It's the perfect way to end a great season.

Not only that, the next season being set in Malaysia is a breath of fresh air. I haven't been this excited for a new season in a while. Well done #SurvivorAU

That makes sense. Given how varied Survivor is, how the game is set up, and the dependent relationships between players, it's super difficult to have a perfect system.

I don't know a lot about the ELO system. Why is it centered around 1500?

Do you have a histogram of the strategic and challenge scores? Keen to see how they are distributed.

after the latest episode, Cirie is up to #2 all time in strategy elo (SHALLOW). still a long way off the model's namesake for #1.

survivorelo.com/players/ciri...

Same show, different lens. Don't take this as gospel, this is just the measurable aspect of someone's current position in the game based on the data. There are a lot of other aspects that aren't captured in the data. But this view is useful to inform how things are shaping up. It will change.

It's a bit early for the probabilities of winning, but it's still useful to see current standings.

My take: Rizo climbs the ladder fast. Let's see how this evolves over the next few tribals.

Also, data is up: 💾 github.com/doehm/survivoR

#Survivor50