Yuku free message boards

I just wanted to also jump in and agree that Rob tossing the clue in the volcano was not particularly negative from either an editorial standpoint or a logical "personal" standpoint.

First of all, nothing in the edit portrays this move as particularly foolish, overconfident, stupid, etc. Rob did something interesting. Show it. People who love Rob will love it. People who hate Rob will hate it. Bam!

Secondly, from more of a personal standpoint (because I know this is where some people vote from), I think it was actually a pretty smart move. Why?
- He already has the idol. Keeping the clue is of no use to him from a "find the idol" perspective.
- Throwing it in the volcano ensures that no one will find it at camp and wonder who put it there. It removes something potentially dangerous to Rob from the equation.
- Grabbing it in the first place ensures that none of the other people in the tribe get it. While Rob's first clue was practically useless, we all know that the clues slowly get more and more helpful. What do you suppose happens if the clue basically says "Climb this tree and you'll find it" and Grant manages to grab the clue? He could go to the tree, figure out it's gone, and then put two and two together, knowing the only other person with clues is Rob. While it is risky to grab it in the first place, if he succeeds he is better able to hide that he has the idol.

Obviously we are used to seeing such behavior as "stupid" or "arrogant," but the player in question usually has something to lose by doing so (Marcus is the prime example). In this case, he is maintaining his position and playing the game as safely as possible. If someone tones this situation N from a personal standpoint, I have to personally question their "CP" insight regarding how to play Survivor!

I think this also shows how this situation could have been edited in a directly P or N way. Perhaps Rob explained what I just did in more detail (it's Rob - very possible), which would have made the move seem like a really smart and proactive one. Instead, they cut it down to just the barebones of the reporting: "I don't need it so I'm gonna throw it in the volcano." On the other hand, perhaps someone in the tribe is wondering where the clue from that reward ended up. Phillip and Grant should know that one was there by now - what did they think about the fact that no one found it? Do they distrust Rob more for it? This sort of thing paints the move in a much more negative light, suggesting that Rob may have been better off letting them find it (even if I still don't think that's the best move).

Of course, we have no idea if either of these options were available, given the footage. I'm just saying that what we saw was a very simple version of all of the ins and outs of this situation. I would be highly shocked if they had absolutely no tools at their disposal to make the viewer see this move as either savvy or stupid using the magic of editing. Hell, even some heroic/bumbling fool background music would have worked.

cantthinkofaname wrote:

---

No, these are not 2 people. The top is if the boundaries of VIS are cut-dry. 17 is Stephanie and the | is the cut-dry boundary. The second one is the boundary being a little flexible to include 17 (Stephanie) in the 3-range. 

I think it should be flexible due to SPV but not for all cases, like when someone is right in the middle of a range, just for borderline people.

Ahh good to hear we agree on that, at least.

The second part is where most of the discrepancies come in. I'd like to attempt to demonstrate why this is problematic.

First, I want to clarify what I mean. What I'm talking about is the sort of thing where a player (like Andrea) is somewhat borderline 1/2 and it's being debated. Then someone says "Well she had SPV so she's obviously a 2 instead." This is the core issue I have. To demonstrate the issue, I finally broke down and tried to "crack the code" so to speak on the vis chart. Using the formulas of the vis chart, I think I can demonstrate the issue I have.

To start off, here's the official chart from this episode:

Player	Sentences	Confessionals	TCQs	VF	VIS
Rob	27	4	0	47	5
Andrea	2	0	0	2	1
Ashley	6	0	0	6	1
Grant	0	1	0	5	1
Matt	5	1	0	10	2
Natalie	7	0	0	7	1
Phillip	44	3	0	59	5
David	19	2	3	44	5
Julie	4	0	1	9	2
Mike	19	1	0	24	3
Ralph	17	1	2	32	4
Sarita	26	3	1	46	5
Stephanie	17	0	0	17	2
Steve	11	2	1	26	3
			Total VF	334	(M = 2.86)
			Avg VF	23.86
			Avg VF x2	47.71
			7.16	1 = 1-7
			16.70	2 = 8-17
			28.63	3 = 18-29
			42.94	4 = 30-43
				5 = 44+

A couple things - for simplicity I am doing these calculations under the following conditions:
- Vis is assigned strictly based on the numbers
- The 4-5 category has been dropped (part simplicity, part me not knowing the formula for that part)

So that's where we are starting. From there, imagine someone wants to argue that Andrea is a 2 based on the fact that she has SPV (not actually happening here but for the sake of demonstration). What they are essentially doing is inserting a new variable into the calculation of Andrea's numbers - SPV - and assigning some sort of weight to it. For the sake of demonstration, let's assume they implicitly consider an SPV to be twice the weight of a sentence. You would, in essence, be producing the following chart:

Player	Sentences	Confessionals	TCQs	SPV	VFNEW	VIS
Rob	27	4	0	0	47	5
Andrea	2	0	0	3	8	2
Ashley	6	0	0	0	6	1
Grant	0	1	0	0	5	1
Matt	5	1	0	0	10	2
Natalie	7	0	0	0	7	1
Phillip	44	3	0	0	59	5
David	19	2	3	0	44	4
Julie	4	0	1	0	9	2
Mike	19	1	0	0	24	3
Ralph	17	1	2	0	32	4
Sarita	26	3	1	0	46	5
Stephanie	17	0	0	0	17	2
Steve	11	2	1	0	26	3
				Total VF	340	2.86
				Avg VF	24.29
				Avg VF x2	48.57
				7.29	1 = 1-7
				17.00	2 = 8-17
				29.14	3 = 18-29
				43.71	4 = 30-44
					5 = 45+

Using this method, only Andrea's SPV is considered. She gets a bump for it, while the others' SPV is ignored. From a relative visibility standpoint, one is essentially calculating SPV where Andrea's = 3 and everyone else's = 0. If the others' SPVs were >0, they weren't considered. In any case, this gets the desired effect - Andrea is now a 2. Yay! In general, this didn't impact many other ratings. It pushed David's down one, though with the inclusion of a 4-5 category it would remain unchanged.

Of course, the big issue here is that certain levels of SPV for other characters are being assumed to be irrelevant. Why count it for Andrea if not for other people? Couldn't it also impact their ratings? Further, the fact that anyone's Vis group changed at all should be a sign that an approach like this cannot be taken without consideration of how it impacts the evaluation of other edits. Visibility is calculated based on where a person lands relative to the average vis x2. If SPV is taken into account for just Andrea, it STILL impacts the total amount of SPV, and thus the average.

So we know that for SPV Andrea = 3, Everyone Else = 0 is inaccurate. What other patterns could there be? Well one could also assume that SPV is equal for all characters (or very close to it). What would the chart look like in that case?

Player	Sentences	Confessionals	TCQs	SPV	VFNEW	VIS
Rob	27	4	0	3	53	4
Andrea	2	0	0	3	8	1
Ashley	6	0	0	3	12	2
Grant	0	1	0	3	11	2
Matt	5	1	0	3	16	2
Natalie	7	0	0	3	13	2
Phillip	44	3	0	3	65	5
David	19	2	3	3	50	4
Julie	4	0	1	3	15	2
Mike	19	1	0	3	30	3
Ralph	17	1	2	3	38	4
Sarita	26	3	1	3	52	4
Stephanie	17	0	0	3	23	3
Steve	11	2	1	3	32	3
				Total VF	418	2.93
				Avg VF	29.86
				Avg VF x2	59.71
				8.96	1 = 1-9
				20.90	2 = 10-21
				35.83	3 = 22-36
				53.74	4 = 37-54
					5 = 55+

Well that fucked everything up. Not only does Andrea no longer look like a 2, but now everyone's ratings are out of whack and the scales are all different. Essentially, adding an SPV variable where everyone receives approximately the same amount of SPV just pulls everyone closer together (their edits are becoming more similar). Fewer people end up on the extremes.

The main point here is that the picture is VERY different when you consider SPV for everyone vs. SPV for only one character.

Naturally, one could also predict that SPV is present for all characters, but varies randomly. I used a random number generator to assign SPV to every player except for Andrea (we know she got 3 because we were paying close attention). The following occurred:

Player	Sentences	Confessionals	TCQs	SPV	VFNEW	VIS
Rob	27	4	0	5	57	5
Andrea	2	0	0	3	8	1
Ashley	6	0	0	0	6	1
Grant	0	1	0	3	11	2
Matt	5	1	0	1	12	2
Natalie	7	0	0	0	7	1
Phillip	44	3	0	2	63	5
David	19	2	3	3	50	4
Julie	4	0	1	3	15	2
Mike	19	1	0	4	32	3
Ralph	17	1	2	1	34	3
Sarita	26	3	1	3	52	4
Stephanie	17	0	0	5	27	3
Steve	11	2	1	5	36	4
				Total VF	410	2.86
				Avg VF	29.29
				Avg VF x2	58.57
				8.79	1 = 1-9
				20.50	2 = 10-21
				35.14	3 = 22-35
				52.71	4 = 36-53
					5 = 54+

This is a picture much closer to reality and again, Andrea no longer looks like a 2 in this context. However, if these are the actual SPVs that people received, this formula brings you closer to their true visibility. For example, people who are low vis using the first chart would rise if they had average/above average SPV. People who are high vis under the first chart would fall if their SPV was average/below average. People who get SPV that is in concordance to their original vis score stay the same.

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

This more accurate picture is missed when we only cater to one player. The bottom line is that the only way "Andrea got SPV so therefore she's a 2" makes any sense is if you completely take Andrea out of the visibility context, boost her up with a separate variable, ignore the impact that a similar boost would have on other players, and then cram her back in there as if you didn't change the context of total visibility. It's a very Andrea-centric way to calculate what should be an Andrea-in-a-greater-context variable.

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

Now, on to a demonstration of why this does not need to be a "follow the chart no matter what" post. What if Andrea's edit truly is poorly captured by ignoring SPV? The following chart assumes that SPV and VF are highly positively correlated. SPV is assigned by taking the original VF, dividing by 10, and rounding it off. The correlation between SPV and VF is .99. Now say I modify this by inserting the observed 3 SPVs for Andrea instead of the 1 predicted by the formula:

Player	Sentences	Confessionals	TCQs	SPV	VFNEW	VIS
Rob	27	4	0	5	57	5
Andrea	2	0	0	3	14	2
Ashley	6	0	0	1	8	1
Grant	0	1	0	1	7	1
Matt	5	1	0	1	12	2
Natalie	7	0	0	1	9	2
Phillip	44	3	0	6	71	5
David	19	2	3	4	52	4
Julie	4	0	1	1	11	2
Mike	19	1	0	2	28	3
Ralph	17	1	2	3	38	4
Sarita	26	3	1	5	56	5
Stephanie	17	0	0	2	21	2
Steve	11	2	1	3	32	3
				Total VF	416	3.00
				Avg VF	29.71
				Avg VF x2	59.43
				8.91	1 = 1-9
				20.80	2 = 10-21
				35.66	3 = 22-36
				53.49	4 = 37-53
					5 = 54+

Here, Andrea's relative visibility has risen. By considering her SPVs in relation to others (in particular, having a high number of SPVs in comparison to people at a similar VF), we see that she actually does deserve to have a higher visibility than would be assigned without considering SPV. Of course, this also rightfully alters the ratings for other players, as this boost impacts the overall visibility context for all other players. The key is that those alterations only serve to make VIS more accurate, as more information has been taken into account.

As I have said, one does not need to go to these lengths to prove that point. As long as someone can demonstrate that a player got an abnormally large number of SPVs (or an abnormally small number of SPVs, I suppose) for their VF position, I am willing to be swayed. When we say "Andrea got SPVs" and that's it, the SPVs of other players is left in a magic black box. They could all be 0 (2nd example), they could all be the same as Andrea's (3rd example), they could all be random (4th example), they could be related to VF in a way that makes Andrea's Vis rise (5th example) or they could all be in perfect correlated harmony, causing Andrea's vis to stay exactly the same (example not shown). Show me what's inside the black box and I'm with you. Tell me the black box is unimportant and I'll occasionally burn an hour showing you why I think it is