While Caleb is sitting in a better position probabilistically, Jackson seems to be better placed with the jury. The more times a players as voted with a juror, the more likely the juror will reward them with their vote at the FTC. #SurvivorAU

This is madness #SurvivorAU

I wouldn't have expected Caleb to be the frontrunner at final 4, but here we are. #SurvivorAU

Here's a random stat. I looked at how many tribal councils have had <6 people attending (excluding the final 5 onwards) • Out of 50 seasons, US #Survivor has had 96 (16% of TCs) • 36 have been in the New Era (32% of TCs) • #SurvivorAU has only had 3! (1% of TCs) and one of those may not count.

Dee wins immunity in Episode 6 but lost BEAST Elo. The model uses full placement ordering, not just win/lose. Dee won her group but placed 11th overall. Stephenie placed 1st (+128.7), Rick placed last (-108.8). Full recap: survivorelo.com/blog/us50_ep... #Survivor50

#SurvivorAU Redemption winning odds I think it comes down to whether Brooke can win the last few immunity challenges, potentially 5 IIC. If anyone can do it, it's Brooke. If not, Keeley is in the best position. #Survivor

Using {reticulate} and referencing R objects in Python with r.* and Python objects in R with py$* feels like God mode. #Rstats #Python

A little more on this. What stands out is the gap between 1st & 2nd. The seasons where there is a huge gap between 1st & 2nd, this is the top 5: 1. Boston Rob (S22) 🏆 2. Russell H. (S19) 🥈 3. Coach (S23) 🥈 4. Christian 👈 5. Savannah (S49) 🏆 2 winners, 2 runner-ups and Hubicki #Survivor

#Survivor50 CPH At this stage in the game, Christian has the 15th highest over-edit (2x means 2 times more than expected) 1. Boston Rob (S22) 2.7x 🏆 2. Yam Yam (S44) 2.6x 🏆 3. Russell H. (S19) 2.6x 🥈 4. Coach (S23) 2.5x 🥈 5. Boston Rob (S08) 2.5x 🥈 6. Rupert (S08) 2.5x ... #Survivor

I've taken a statistical look at the 'edit' for men and women across the 50 seasons. I've used a model to estimate how many confessionals to expect compared to how many received and shown how this differs for M/F, how it's changed over time, and where S50 ranks. #Survivor

#survivor50 confessionals/hr Christian #Survivor

This is an amazing site for #Survivor stats! It is incredibly detailed with stats across all aspects of the game and players, and a new take on a Survivor score. Check it out to see who is on the Survivor Mt Rushmore 🗻 🔗 www.survivor-reference.com Created by Chris Colton

#SurvivorAU confessionals/hr This pattern is wild #Survivor

#SurvivorAU Vote Network Who has voted with who, and how many times indicates alliance strength and is a strong predictor of who will win. Look for those with a large circle and the most votes - Simon and Keeley stand out

I honestly don't have a good feel for who is going to win this season from here. However, the data suggests there are two clear favourites: 🟣 Simon - 34% 🟢 Keeley - 26% #SurvivorAU

What episode did Mark find his idol? #SurvivorAU

At merge, the player that wins the first IIC immediately raises their threat level - "they're going to win every immunity challenge" tends to be the narrative. So, is it disadvantageous to win the first individual immunity challenge? Historically, not really. 🧵 #Survivor #SurvivorAU #stats

Merge is happening next week in #survivorAU, potentially the most important part of the game. It's critical to be on the right side of the vote. A player who is on the wrong side of the vote after merge has ~75% lower chance of winning the game. Here's a short #stats analysis 🧵 #survivor

I often get asked if the #survivoR #rstats package is also available in #python 🐍 Yes, it is! Check out Matt Stiles' Github for a script to read the data sets in as pandas data frames 👍 🔗 github.com/stiles/survi...

#Survivor50 E4 confessionals/hr Looking forward to an episode where Christian and Ozzy don't go to tribal. #Survivor

#Survivor50 E4 has the lowest IMDb rating of all time, currently at 4.0. But also 424 ratings, which is the second highest, second to S41E1, and it has only been 24 hrs. When it comes to polls/surveys, people are FAR more likely to leave a rating if they are unhappy. 1/2

So, I didn't know who Jac Brown was until this episode. Turns out he's pretty popular (with the players at least, not necessarily the viewers). Anyway, data is up. 🔗 github.com/doehm/survivoR #Survivor #Survivor50 #SurvivorAU

For those who love #survivor stats, you should check out @kevincsteele.bsky.social work on Reddit and his website. He's built an ELO-based model to rank castaways across all seasons. S50 is updated after each episode! 🔗 www.reddit.com/r/survivor/c... 🌐 survivorelo.com

I've shown that higher confessionals means a higher chance of winning the season. I've also wondered if higher confessionals also means a higher chance of being voted out next. So does this stack up? 1/ 👇 #stats #Survivor

#Survivor50 confessionals/hr #Survivor

#SurvivorAU confessionals/hr Mark has the 4th highest cph compared to the expected value after 9 episodes 1. Nick 3.3x 2. Luke 3x 3. Kat 2.8x 4. Mark 2.6x 👈 5. David 2.5x 6. George 2.4x 🥈 7. Mat 2.3x 8. Phoebe 2.3x 9. David 2.3x 🏆 10. Kirby 2.2x #Survivor

Data is up. Also, excellent news. Only two episodes of #SurvivorAU next week! #Survivor github.com/doehm/survivoR

Those savages better not kill the chickens straight away! I did a whole thing on this last season. Those in the final 5 will consume about 800 more calories by first farming the eggs and then slaughtering the chickens. It's a no-brainer. #Survivor #Survivor50 gradientdescending.com/you-win-chic...

Here's the full analysis of the impact of tribe swaps and being in the minority. Overall, it's not that bad. A player in the minority has a +35% increased chance of being voted out at the next tribal council than someone in the majority. #Survivor gradientdescending.com/are-the-chan...

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

#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

#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/

#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

#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

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.

#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

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!

Knowing that Alex Honnold climbed Taipei 101 while listening to Tool I think the only way to watch it is on mute with Tool playing in the background.

Quick chart of the 'Iconic' #Survivor episodes IMDb ratings. The distribution is only for episodes featuring at least one S50 cast member. - 8 episodes are above avg - 4 in the top 10% - S20E1 is the highest ranked episode (14/287) - S48E12 is the lowest ranked episode (278/287) #Survivor50

Here's another one from the John Wick universe - Ana de Armas from Ballerina. Made with geom_ribbon and magick::image_animate. #Rstats #dataviz

Here are some frames from the animation. The first is at 1% resolution of the original, the last is 100%. Kinda cool. #Rstats #dataviz

Having some fun recreating images in ggplot with the help of magick. This was made with geom_ribbon and stitched together with magick::image_animate. #Rstats #dataviz

I may regret asking this question, but should I watch #TheTraitorsUS, and is it worthwhile collating the data?