Yuku free message boards

2. Discuss (a) dumping the tone ratings or (b) selecting a committee of maybe five veteran objective edgicians to decide on the tone ratings for each episode, after some "spirited" discussion in this thread by ALL participants. This will prevent "fandom" from continuing to overshadow tone ratings. I would suggest OTH, TLS, flf, Warrior, etc.

I strongly support this, as an Edgic viewer. In fact, I really think the guidelines for what is and is not Negative, Positive, etc. should be clearer cut, with an emphasis on the fact that it is about editorial intent and not your own personal opinion of the player, situation, etc. You can't say something like "Kathy is clearly NNNN in this episode because she quit. Quitting is something one should never do and therefore, it gives you an automatic N." While this is a valid opinion about the action of quitting, it does not capture what the editors intended for the audience to think and feel. This is really just a classic example, but there are many others. Underdogs are not automatically positive, betraying an alliance is not automatically negative, poor challenge performance is not automatically negative, etc. Do these things typically correlate? Yes, they do. But the focus needs to be on the editors' intent and that intent alone, IMO.

4. Do we really need the strategy ratings ?

I am currently messing around with creating a dataset so I can do some logistic regression analyses on Edgic and pretty much felt they weren't that useful, as of right now. If you look at the guidelines, they are quite broad and Parvati broke a couple this time around. I think strategy is likely to be an important aspect of the winner's character insofar as he/she provides the goods for the editors to play with. If a player isn't involved in a lot of strategy, how can we be sure they will not get a string of S0s? If a player pulls off a majestic strategic move in the early goings, why not give them that (resulting in an S3)? If a massive blunder by a single player is good TV and they end up winning, would they show it?

I think strategy is something that is typically correlated with winning the game, for sure. But I just don't know how one can predict the winner (by the merge no less) based on what is shown in terms of strategy. Whether or not a player gets strategic airtime seems highly circumstantial and I have doubts about the possibility of teasing apart the confounds within a "general strategy" barometer.

What about thinking about WHO gets credit for what events? If a group of players take someone down, which one is getting the editorial focus? If a massive blunder occurs, who takes the blame? If a player is constantly involved in such situations, but NEVER gets featured prominently, I would assume they are peripheral to the story and thus, not likely to win. I would also suppose that distractions are likely to get more than an expected amount of attribution. Winners likely get some balance.

Of course, this does not completely solve the issue. The number of possible players to attribute any given event to is variable. The number of times such an event occurs in the game is variable. If a player rarely has the opportunity to have such attributions made about them, what would their winner/loser edit look like? All in all, though, I think this may be a better way to attack the possibility that strategic content influences a winner edit.


4) The winner is very unlikely to receive an INV rating during the course of the season, and will likely receive at least a 2 visibility in Episode 1.

I think Edgic needs to more clearly lay out the guidelines for Episode 1 and the Merge. These two episodes are among the most talked about and cited in the debates that occur. "So-and-so had a weak episode 1." "This person had an amazing merge episode." Etc etc. You never hear someone say "Person X had a GREAT episode 4! What a clear winner's edit!" These episodes are of prime importance to so many people, so they should get more than a quick addition to guideline 4 (and nothing really at all for the merge). Along this vein, what about swap episodes? Are they also of prime importance in some way we haven't noticed yet? They are rarely cited in discussions, but are there patterns out there?

9) If two people are a 'known final pair' then the likelihood of them actually being the final two is slim. It would need to be a situation that cannot be avoided (i.e. Amber/Rob)

Drop this outright. Very circumstantial and not accurate a lot of the time.

Tone

PP) Extremely positive.
P) Positive.
No rating Static neutral
M) Dynamic Mixed neutral - If you think someone was noticeably P AND N within one episode.
N) Negative.
NN) Extremely negative.

I have argued before for this, but I think PP and NN are superfluous and create needless debate and arguments. These are much like INV in that they are practically guaranteed to eliminate a player. However, I feel that any player you use them on is almost always already eliminated by other factors anyway. In practice, would we ever have an Earl on our hands - a clear Edgic leader - have a PP or NN episode and then suddenly decide "No, he can't win now" with any sort of confidence? Such an event has not happened yet (before the episode in question, all NN or PP players were Edgically weak, at best, if I remember right), so how will we know if it's even important?

Thus, even in future use they would be VERY rarely used to actually eliminate a player. Even if they were, people would need to CAREFULLY consider whether or not the PP or NN came from events that were unavoidable. Judd's outburst was TV gold. He didn't win, but if he did, I would bet you anything they'd have shown it anyway, and you couldn't very well spin that as P or neutral unless Margaret had some really nasty footage out there.

Visibility Measures the amount of presence that a player had in an episode. Visibility includes confessionals, speaking time, face time, second person visibility and tribal council questions (TCQ).


Is it possible to find a way to include face time and SPV into the standard calculation of visibility done by Warrior? If they are part of visibility, they should be utilized in any quantification of a rating, shouldn't they? Face time is notably more difficult to do, but considering the way it is implicated so clearly in the UTR ratings, it seems it should be less open as to what denotes high and low face time (but it probably should be weighted less than the other three anyway).

Ratings

INV - Invisible

1) Irrelevant, no confessionals or tribal council questions in the early episodes, little or no confessionals or tribal council questions in the mid to late episodes.
2) The player's editing was insignificant to any character development.

You could probably alleviate a lot of the issues people have with this rating by making it more concrete. Why is INV the only rating that CHANGES in definition over the course of a season? Just make it consistent and clear - no visibility = invisible. Set a cutoff on Warrior's chart and set that as the point where a character gets so little visibility that they are invisible.

The only reason INV is such a controversial rating is that it is basically a stamp of death for a player. If a player gets an INV, they are virtually eliminated on the spot, barring some sort of anomaly that has thus far occurred exactly twice (Tina and Amber). If you make it clear cut, it is no longer something to dispute. I would suggest something like having less than half of the visibility factor required to get a 1 in visibility on the chart. If you are that low on the visibility scale, it is very difficult to argue for a UTR1. This would likely decrease the number of INVs given, but considering its weight, relative to the other ratings, I think that's a good thing.

Finally, that second descriptor applies to a large number of characters in a given episode, but is not indicative of invisibility. A player can be MOR and still get no character development. However, a player with no visibility cannot have character development. Again, the key is the level of visibility. No character development is an artifact of the low visibility and not something that actually indicates low visibility.

UTR - Under the radar

1) UTR means a character that is either underutilized or deliberately hidden (the trick is figuring out which). In story terms, they are kept out of focus. An UTR edit means the player may be seen, but they aren't seen doing or saying anything of major significance. Since showcasing players who do nothing isn't very entertaining, UTR edits do tend toward low visibility, but that needn't be the case.

Actually, the opposite seems to be a pretty prevalent opinion. Can you have an UTR5 character? I bet a lot of Edgic participants would say no (although I would disagree with them).

2) Players who receive significant face time, but little to no speaking time.

4) If the player gets some significant screen time, but doesn't do anything special, then you rate it UTR.

This contributes to what I have said above. Visibility is primarily calculated by Warrior's chart (which is one of the primary means used to rate visibility, it seems), which considers speaking time only. If a player receives little speaking time, they will register as low visibility. So if you use guidelines 2 and 4, how does the end of guideline 1 work?

MOR - Middle of the road

2) Someone who gets an odd positive and negative mix, but not in an Over the Top manner.

Is MOR a "Mixed" tone or is it a rating, here?

3) If the player did something that you feel makes them better that UTR, but not enough to qualify for CP, then you rate MOR.
4) MOR can be used to describe characters who were mostly UTR with a fair amount of CP.

These appear contradictory. MOR is either "in between" UTR and CP or "lots of UTR with some CP." Is it one or the other? Is it both? If it's both, does MOR accurately portray both ratings (are they functionally the same?)?

CP - Complex personality

4) If the player gets quality screen time and is either shown multiple times being strategic or shows a multi-dimensional side, then you go with CP.

This needs to be rewritten, based on the usual commentary throughout the seasonal discussions. Someone always steps in and says "Strategy does not mean a person is CP." Well why not? You say it does in the definition, don't you? (Of course, I agree that CP does not mean just strategy) However, the guidelines for it do not make that clear.

OTT - Over the top

2) OTT players are generally always edited in a Positive or Negative light.
3) A mixture of positive and negative editing ends up being neutral overall.

These could use a bit of clarification, given the M rating. Is OTTStatic possible, then, or does OTT necessitate M, N, or P?

4) Usually lacking in strategy, or strategy not expressed with much internal conflict - but they can have lots and lots of conflict with other people; no remorse if N and no "evil" strategizing if P - only focusing on loyalty or saving one's own skin

What dominates what in a situation where a character is OTTCP? That is, a player who is constantly strategizing, such that it dominates their character? Are they CP because they are talking strategy or are they OTT for being one-dimensional? This comes up often enough that I think it needs to be clearly addressed.

I also have a few comments about the "Rating" ratings in general.

Basically, I wonder if it's possible to rethink them or at least rethink their roles. They were originally, it seems, meant to represent both what we call "rating" and "visibility." UTR seems to be about low visibility, OTT and CP seem to be higher visibility, with MOR in the middle. Similarly, as I have alluded to, many of the ratings are linked to tone. OTT is rarely static, for example. In some cases, they exemplify all three: MOR seems to have been thought of as a Mixed tone, medium visibility person with little CP.

The 5 ratings (INV, UTR, CP, MOR, OTT) also seem to be thought of as exemplifying some sort of continuous scale. This is part of the whole MOR differential and rating average thing. What theoretical reason is there to think of them in such a way? What is this a scale of? Can a person increase on the scale by increasing their "Ratingness?" These calculations also, then, assume an interval scale - with each anchor being equidistant from each other. Is a jump from INV to CP the same (on whatever variable this scale exists on) as a jump from CP to OTT? It seems not - while CP to OTT is about insight and/or strategy, INV to CP would also require increased visibility. Although the MOR differential seems to aid in picking the winner, how does it work, statistically?

What I'm getting at here is that the constructs Edgic is working with in terms of "rating" are not orthogonal. That is, predicting something (the winner) based on scores on a variety of variables (tone, visibility, and "rating") makes attributions for that prediction difficult. We know that the winner is unlikely to get too many UTRs. But is that because the UTRs are the problem or the visibility? Or is it that they are NOT getting CPs? Originally, the concepts were the same thing, but now we can't be sure, as UTR does seem to contain some degree of visibility in its definition. In pretty much any of the 5 ratings, we imply some sort of visibility or tone.

In statistics, this is a bad thing. We want to know what variables are explaining our outcome. But when there is a high degree of correlation between our predictor variables, we aren't sure what the causal factor is. By working to isolate the factors involved, Edgic could, I think, become more precise and surer of what causes what.

So basically, is there a way to utilize the "rating" construct so that it minimizes overlap with tone and visibility? Is rating supposed to be qualitative (each rating is unique and unrelated to the others) or quantitative (increases on variable A and variable A alone will eventually move you from INV through the spectrum to OTT) or a combination of the two? For instance, I could buy that UTR and INV are distinct only in visibility. Or that OTT and MOR could be different only in tone. But I think these continuous scales need to be more clearly laid out, as well as the rationale for why (Although ideally I think Edgic could get the most bang for their buck by creating more qualitative constructs for "Rating").

All in all, I don't mean any offense by the ratings comments. It's meant to be constructive. I like reading Edgic and I want it to be as successful as possible. My background in statistics is what really inspired these suggestions, as I think they would be helpful to the overall aim of Edgic. Of course, altering the very foundations of Edgic ("rating") is unlikely to happen and I know that. If anything, I just hope that people consider the possibility that the ratings being used might be inefficient, contributing much of the same information that is gained by tone and visibility. The only one that really stands out is CP, which denotes, in my mind, basic character depth (which strategy just so happens to be a part of much of the time). It might actually be possible to solve the puzzle just by knowing that a person is generally average to medium-high on visibility, generally average or mixed on tone, and has a fair amount of "character development" in some way shape or form, and has significance in the Premiere, Merge, and possibly Swap episodes.

Anyway, hope this inspires some discussion that leads somewhere productive. If not, hell I know I'll keep reading anyway. Thanks for reading.

SFC wrote:

---

Just a tiny little idea.

If someone is ever 100% Edgically Eliminated on the voting site they get taken off the selectable winners for the rest of the season

I agree with this sentiment. I'd also take it to an individual level and say anyone officially eliminated by an Edgic participant during the season is off the table for that person's "official winner pick." If you're going to go out on a limb and say a person is DEFINITELY eliminated from winning, you should stick with it and mean it. Example: you shouldn't say Amanda is 100% done for the whole season and then, during the last few episodes, change your mind and officially pick her. If that's the case, you aren't going by the whole edit and are being swayed by only current episodes. Is it possible that you made a mistake in eliminating her early on? Perhaps. But if you missed the boat by so much that you fully and completely wrote her off, to pick her later on is just cheap, in my opinion. As many have said, there is nothing impressive about choosing a winner in episode 12.