Looks like we've got two tribe swaps coming this week. This deserves a longer post, but I quickly looked into the chances of surviving the next tribal after the swap. tl;dr those in the minority are about 3.4x more likely to be voted out next. More to come.
#Survivor50 #SurvivorAU
Raw counts
#SurvivorAU confessionals/hr (cph)
Mark is clearly taking the lion's share with 2.2x higher cph than expected. It looks like there are fewer confessionals across the board this season.
#Survivor
#SurvivorAU S12 Week 2
Best plays this week:
🪆 The Russian Rouflint blow up
🦗 The "I'm Mat Rogers' brother" reveal
#Survivor
Raw counts
#Survivor50 confessionals/hr (cph)
I think I need to add to the model that if you 💩 your 👖 your get more confessionals! Christian has 1.6x higher cph than expected.
#Survivor
I really liked this challenge and I'm glad #SurvivorAU have challenges played in rounds e.g. the first to 3 or 5 points. US #Survivor has decided they don't need them anymore. I think the last challenge played in rounds was S40 Ep1 - By any means necessary.
I was a bit late setting this up, but the confessional timing and counting app is ready go for #survivor50 and #SurvivorAU for those interested.
danieloehm.shinyapps.io/survivorDash/
With all that considered, I don't think we'll see too much of them.
Next week, I'll show who we should keep an eye on.
6/
The number of confessionals per hour is correlated with the chances of winning. Higher cph, higher the chances
It's obviously very early in the game and anything can happen. At this stage higher cph can also mean a bigger threat and a more likely boot
5/
gradientdescending.com/do-winners-g...
A player that starts off with 0 confessionals, on average ends the game with only 0.9 cph.
No player has won with a cph lower than 1.7
A player with 1 or more confessionals, on average, ends with 2.3 cph.
Still not looking good for them.
4/
There's never been a winner with a cph lower than 0.5 after the first 3 episodes.
There's only been 1 player in AU history that has made the final tribal council with 0 confessionals in the first 3 episodes - Caroline (S9)
So far, it's not looking good for them.
3/
#survivorAU has a habit of giving some players barely any edit. Likely because they aren't that influential in the game, and the other storylines are more interesting.
What does this mean for Richard and Tez?
2/
#SurvivorAU Confessionals/hr (cph)
Those with a 🟠 have a higher cph than expected - that's Harry, Mark, Lottie, and Johnson.
Richard and Tez didn't receive a single confessional in 3 episodes. What does that actually mean? It doesn't look good for them 🧵
1/
#SurvivorAU Redemption
21 remain
2 idols in play
Loz has the most votes received
3 players on Barren have received at least 1 vote
#Survivor
Gutted about Kyle being pulled due to an injury. Not the way anyone should go out. He becomes the 21st player to be medically evacuated and the 5th to be pulled in episode 1.
#Survivor50 #Survivor
Since Ozzy didn't find the idol, he was gifted it, he perhaps should get slightly fewer cnfls, but I don't think it changes the outcome.
Those with a 🟠 have a higher cph than expected. Stephenie has the biggest difference between actual and expected and therefore the most 'over-edited', so far.
#Survivor50 Confessionals/hr (cph)
Oscar (who?) clearly had more screen time than other players (8.9 cph). Was he 'over-edited'? Ozzy fought for supplies, received an extra vote and an idol, and went to tribal council. Considering that, he actually received as many cnfls as we would expect.
1/2
I cancelled my vacation. I cancelled my doctor's appointment. I cancelled everything I could cancel, and now up to date with AU12 and US50. Data is on Git, and some stats to come!
🔗 github.com/doehm/surviv...
#survivorAU #survivor50
Typically, I'll be collecting the data a few hours after it airs in the US. I've been so busy I'm just about to watch ep1 - trying to avoid spoilers! I have the base assets pretty much set up now. If you're interested in contributing, send me a DM and we can chat more 🙂
Ok, so it looks like the #survivor wiki has put protections in place which means it blocks my requests and I can't automate the data collection. I'm going to rewrite the whole pipeline using other sources and AI models. It may take me a while to catch up given AU starts tomorrow.
*excited
Oh that's awesome!! Edited to know where it goes. He's right, the world needs to know!
Cheers! It's tough picking a winner before the start! I will say, from my model, Rick has the highest probability of winning for a non-winner. So, Rick is my winner pick based on data, but I'd love to see a Christian or Chrissy win.
It does. It's probably cleaner to remove her AU appearance, but it's also another data point.
2/2
The difference between their expected and actual number of confessionals is the strongest indicator of who is going to win. This is only relevant within season. So I'll be comparing these values to the S50 counts, which could be a stronger predictor. Stay tuned.
#Survivor50 cast % of confessionals received above their expected amount across their career
- Coach: Highest on +77%
- Jonathan: Lowest on -31%
- 5 players had fewer confessionals than expected
- 14 players above 10%
- Rick: Highest for a single season player on 51%
1/2
I'm very stressed out by this
This is kind of insane. I have to export 30.05kWh to recoup the cost of drawing 0.05kWh from the grid. Wild.
Got my new battery installed. The real time feed of data has me hooked!
