|
Yul
| Reputation: 1 | Group: | Prediction Writer | Posts: | 216 | Joined: | Feb 22, 2015 |
| Post #1: 15th Apr 2015 5:57 PM | |
Hi guys! So I was bored so I decided to run an experiment. I wanting to see if there were certain traits that were in opening statements for FTC that could determine a winner.
I looked at every FTC in sausage history. I analyzed it thrown IBM's Watson through this website: http://watson-um-demo.mybluemix.net/
The program analyzes how prevalent certain personality traits are by looking at one's writing. It demonstrates the writer's personality at the time the writing was made. It's interesting to thing about if this has an impact in how others perceive it, since it's difficult to express emotion/tone online.
The numbers represent %s. 0 meaning not at all and 100 meaning very much so. I'm gonna try to see if I can see any patterns or anything noteworthy, but obviously each FTC is situational.
Let me know what you guys think:
https://docs.google.com/spreadsheets/d/18PPGwoxPJ-kLFZlMOn2dTKb-M2rNtPDVbfF3BFXgI00/edit?usp=sharing
Note: The sample size of this is really small obviously, take this data for what you will. I don't want to make any statements about the data until I know more. | |
| | |
Host Lex
| Reputation: 14 | Group: | Hosts | Posts: | 1,713 | Joined: | Sep 16, 2014 |
| Post #2: 15th Apr 2015 6:16 PM | |
This is very cool | |
| | |
Host Brett
| Reputation: 20 | Group: | Hosts | Posts: | 823 | Joined: | Sep 19, 2014 |
| Post #3: 15th Apr 2015 7:01 PM | |
Rory has 0 imagination. He is not prone to challenging authority, and has a low intellect. He is cheerful, orderly, and cooperative. He is very agreeable, and known for his altruism.
Yes this sounds like Rory. | |
| | |
Jonas Otsushi
sushi master
| Reputation: 1 | Group: | Prediction Writer | Posts: | 80 | Joined: | Feb 23, 2015 |
| Post #4: 15th Apr 2015 8:15 PM | |
I just did a quick regression analysis of your data Igor and I couldn't find anything conclusive yet but you could be onto something legit. | |
| | |
Bob Dawg
| Reputation: 5 | Group: | Prediction Writer | Posts: | 181 | Joined: | Feb 21, 2015 |
| Post #5: 15th Apr 2015 8:23 PM | |
I hope it corrects typos | |
| | |
Yul
| Reputation: 1 | Group: | Prediction Writer | Posts: | 216 | Joined: | Feb 22, 2015 |
| Post #6: 15th Apr 2015 8:55 PM | |
|
I just did a quick regression analysis of your data Igor and I couldn't find anything conclusive yet but you could be onto something legit. |
I don't know what that is but thanks :D | |
| | |
Dr Sean
| Reputation: 0 | Group: | Prediction Writer | Posts: | 79 | Joined: | Feb 22, 2015 |
| Post #7: 15th Apr 2015 10:10 PM | |
Yeah I'm very curious if the winners displayed some sort of traits more than the runner-ups.
Also I'm Sean, not Jon (Season 5) :(
But this thing is pretty interesting. I put my opening statement into the thing to see what it said. | |
| | |
Host Brett
| Reputation: 20 | Group: | Hosts | Posts: | 823 | Joined: | Sep 19, 2014 |
| Post #8: 15th Apr 2015 10:49 PM | |
I wouldn't put too much stock into what Watson has to say tbh. I know the more samples it has, the better it is at this, but look at all the similar categories that get polar opposite results. Rory has no imagination but is artistically interested (Watson digging at Vern's youtube career). He's not adventurous at all, but he is heavy in excitement-seeking. It feels like a horoscope trying to cover all possible bases.
There are a few shared traits with almost all opening statements though. Imagination is low across the board. Ted is the only one who sort of has any.
But Watson once gave me 100 in extroversion so | |
| | |
Yul
| Reputation: 1 | Group: | Prediction Writer | Posts: | 216 | Joined: | Feb 22, 2015 |
| Post #9: 15th Apr 2015 11:31 PM | |
Right, confirmation bias can be a thing. It's important to sift through the bullshit. Opening speeches are always written a certain way so maybe. Every speech is persuasive writing. I feel like some of these categories are irrelavent, I just wanted to put all the information that was present. I feel like more samples would make it better. | |
| | |
Rory "THE CRUSHER" Freeman
| Reputation: 0 | Group: | Jury | Posts: | 971 | Joined: | Feb 21, 2015 |
| Post #10: 16th Apr 2015 1:28 AM | |
How is this determined.... makes no sense | |
| | |
Yul
| Reputation: 1 | Group: | Prediction Writer | Posts: | 216 | Joined: | Feb 22, 2015 |
| Post #11: 16th Apr 2015 1:07 PM | |
robots | |
| | |
Jonas Otsushi
sushi master
| Reputation: 1 | Group: | Prediction Writer | Posts: | 80 | Joined: | Feb 23, 2015 |
| Post #12: 16th Apr 2015 3:03 PM | |
I haven't gone home since I did a quick one yesterday, but essentially a regression will pointpoint how substantial one of these categories is in predicting a winner.
Statistics nerd-out:
|
The biggest determinant if Watson is an appropriate measure of is an R-squared statistic that is on a scale of 0 to 1. If it's 1, it means that the variables in question 100% predict your dependent variable (winning/losing). The one for this was around .3, which might seem low- but for a binary regression (which is when your dependent variable only has the values of 1 and 0) with subjective variables involved, it was actually pretty high. None of the variables showed much significance, but could very well be due to the sample size.
I think a better survey method might be collecting multiple question responses from each contestants, or to branch out past Sausage Island.
Either way, it's entirely too much work for a result that is impossible to emulate, but it's interesting nonetheless. It would be cool a cool addition to ORG edgic if we could pinpoint personality qualities that make someone a likely winner. Possibly based on confessional posts. |
| |
| | |
1 Users Viewing (1 Guests) |
|
|