Yuku free message boards

Kitty Pryde1 wrote:

---

The fact that CANDACE gets an OTT while neck-licker Coach gets an MOR is fucked up in my mind. It makes NO sense whatsoever.

Agreed.

astroline:
(and I've seen multiple posts saying "Well she was voted out, so who cares what her rating is?" - which is a HUGE problem that Edgic has)

Give me one good reason why that's a problem.

Why do we do Edgic? To determine a winner.
How do we do that? We look at the overall nature of the contenders' edits.
How do we keep all of that straight? We do ratings on individual episodes to keep track of data.

Data on players that have officially lost this season does not help us in the least bit in terms of determining who wins. Because the edit we are discussing only refers to boot episodes, this information also does not aid us in determining who losing players are (unless pegging a loser DURING the episode based on analysis of mounds of data is a priority for some people).

As I have acknowledged, the whole thing is fun for people who like to rate and debate. It's also a sort of "practice" for rating other players correctly. However, neither of these is so major that it could be termed a "problem" that some people don't care that much (really, 100s of ratings are done a year on viable players so who needs practice now?).

Collecting data is fun if you have tenure, but unless you do, you should work on your publications, so to speak. Edgic is not so well respected and powerful that it has nothing better to do than debate tone on a player that won't win anyway. The priority should be on enhancing the ability to pick a winner, IMO.

taks:
I'd really like someone with a more keen-eye for this who voted "P" (as many of us did) to possibly point out a specific? Or anything that would help?

Again, I second this. Where's the evidence that Taj's edit was P and not just that a general viewer P reaction existed?

KP1:
I gave Joe a one. The problem with visibility is that too many people just read what's on Warrior's chart and go with that.

That's not a problem, IMO. That's progress. Edgic has devised a way to take an editing construct and turn it into OBJECTIVE data. In the collection of data, objectivity is a GOOD thing.

If we just go by our own impressions and feelings, we introduce unneeded noise into the data. My favorite character this season is Jerry. I hope he wins. I also like Sierra, Tyson, and Brenden. I hate Coach. The rest of them are just sorta "bleh" to me. When I watch the show, I do it in large part because I enjoy the show. I like seeing my favorites on screen and want to boo the ones I hate. Thus, when they come on, my natural reaction as a viewer is to look closely and listen. I then process all of those ideas and integrate them into my opinions about the character. For the rest of them? From my standpoint who cares? So Sydney is talking about food... who cares? Stephen made a friend? Okey dokie, I guess. Woah woah Spencer is gonna blindside someone at next TC!? THIS I gotta care about!

Hypothetical ratings based entirely on my thoughts as I watch:

Jerry, Sierra, Tyson, Brenden, Coach: 5
Sydney: 2
Stephen: 3
Spencer: 4
Everyone else: 0

Does anyone rate THAT way? No. However, these implicit biases in the way we watch TV shows necessarily creep into our ratings. People might rate the ones they like/hate higher in visibility than is deserved and the ones they don't care about lower than is deserved. Imagine the hypothetical Joe fanatic after last episode: 2 seems much more reasonable. What we get is objective data + variance due to personal attentional biases.

Now, if we had a large, representative sample of Survivor viewers do this exercise, I would agree that variance based on personal viewer reaction would be inconsequential in the measurement (or at least less important) due to the sample size. However, because we are a small and unrepresentative group, the goal should be to eliminate this variance from the data in any way possible.

KP1:
If that's the case, then what's the point of voting for visibility at all? Why not just automatically give those visibilities to the contestants and have close calls be close calls?

I ask this question myself. Why not? You have an objective measurement of visibility, so why introduce variance? Why take unbiased data and allow noise into the equation?

This is not to say the objective measure is perfect (which is why Visibility voting is still somewhat needed). Ideally, I see Visibility as composed of:
Speaking time
Screen time
Second-person visibility (including that from Jeff)

Currently, the only thing that really gets measured is speaking time. I think SPV is something that COULD be worked in and I will continue to suggest it until I am told definitively to STFU and given the reason why it isn't, or until people agree (as an aside: the word visibility is part of the term! Why NOT count it!?).

Screen time is the bigger issue. This is often suggested as a reason for why people deserve more visibility than they get. However, because these things are NEVER quantified objectively, I feel like these suggestions are especially prone to attentional biases. Was anybody watching for the amount of non-speaking screen time that Joe got last week? Did anyone compare it directly and objectively to other players' non-speaking screen time? Hell no. So why is it that we can argue for the use of non-speaking screen time for calcualting the visibility of ONE player (usually being pushed as a winner candidate), but never even consider it with another? By doing that, visibility is being calculated differently for different players. Doing that makes visibility not only unobjective but also INCONSISTENT. That weakens the whole visibility construct.

Overall impressions and character development? I think that stuff belongs in the "Rating" rating. People will disagree, but personally I like clearly defined lines. A person with low impact as a character is UTR, not necessarily 1.

All of that said, I use Warrior's chart exclusively. Could it be more accurate? Sure. Will it ever be perfect? No. However, from a statistical standpoint, I think should be the standard and can't argue against it.

------------------------

Edited to add:

Official ratings:

Debbie - UTR2
Erinn - UTR2
Jerry - UTR2
Joe - UTR2
Sandy - UTR2

WOW!

Five toneless, 2 vis, primarily undeveloped characters according to the chart. I didn't get that feeling as watching the episode at all, myself. Did anyone actually vote this way for all 5 characters or was there a lot of variance for each that centered on this oddly coincidental setup?