DI Observations & Possible Implications

[thing85]

Alright, I'm not sure if this has been discussed before, but there are some things I've noticed in DI that lead me to some interesting conclusions.

Two things seem to pop out at me when I look at the DI hard graphs:

1) Contestants with higher DI usage (more votes recorded by DI) seem to be represented as lower than actual in the busy %. Perhaps this is due to the larger sample size of busy signals in comparison with other contestants.

2) Contestants with relatively low DI usage (less DI votes) seem to be extremely variable and/or represented as higher than actual in the busy %. Again, it's a sample size issue.

One contestant who is consistently high in DI votes is Gina. If you look at week 12, she received 24,000 DI votes (or vote attempts?) well above all contestants besides Blake (who had ~22,000 vote attempts). If you look at the busy signal % that week, she and Blake were among the bottom 2-4. We can assume that neither were actually that low as neither were in the B3 that week. Phil, who has consistently received a low amount of DI vote attempts (about 8,000 that week) was in the top 3 of busy % that week! He must've been doing really well! Well, not really. He was in the B3 that week.

Week 11: Gina has a relatively HUGE amount of DI vote attempts (~40,000) when the next highest contestant has only 20,000. And once again, you look at busy % and Gina is dead last. And yet again, she is not in the B3 (well, B2 that week). Taking our other extreme, Phil, we see that he actually has the LOWEST DI vote attempt count (~8,000). What do we find in the busy %? He's ranked #2! Now, he wasn't in the B2 that week, but the trend (DI vote attempts vs. busy %) continues.

What's the point in looking at this? It's not to show you how to look at DI and point out who's for sure going home. The vote attempts vs. busy % relationship is not solid enough to make conclusions about who is going home for sure. What this finding does show us, however, is when to notice a contestant on DI is falsely "safe" or falsely in trouble.

Let's look at this week: Sanjaya, Gina, Jordin, and Blake are high in DI voting (in that order). Phil is dead last. Busy % once again shows Phil as "safe," ranked 3rd. We know that after last week's very good performance, he was still in the B3. This week, IMO, he was much worse. DI once again has him safe, but with the trend I'm examining here, I beg to differ.

As for Gina, Jordin, and Blake? Surprise! Gina and Jordin are lowest in busy %. Blake is 4 or 5, and Sanjaya is 4th.

The outliers of this trend seem to be contestants who are actually doing really well or really poorly - in other words, regardless of the the number of DI votes, the public opinion is so heavily weighted in one direction that the busy % reflect the reality. So, regardless of Melinda's DI vote amount, she is consistently at the top of busy %. And, when for example, Brendon and Stephanie were each eliminated, both weren't particularly high in DI vote attempts, but still wound up at the bottom in busy %.

Closing notes:

- Like Phil, Chris R is low in voting attempts on DI. This could artificially inflate Chris R's DI ranking much like it has for Phil.

- Lakisha is an interesting case, as she is both low in DI votes and low in busy % (this week). Contrast this to the T12 week when she was still relatively low in DI votes but high in busy %. I'm not sure what to make of this.

- Remember that the underlying issue here is sample size: Low DI voting doesn't necessarily mean that a high busy % is false - it simply means the variability will be greater and can thus lead to artificially high or low busy percentages.

Also, remember to factor in my inherent bias in this. I'm making conclusions as the data seem to fit them, so you must take all of this with a grain of salt. I still believe that there is an added insight by examining this relationship.

I would appreciate any criticism and/or information that people can add to this.

Edit: I'm also not suggesting that Gina is safe - it's possible that her busy % rank is accurately reflected. I just wouldn't mark her as the one going home for sure (people, especially in other forums, seem to think she's gone for sure).

[smbruin22]

interesting this week.... if phil is in the bottom 2 or 3, then i think DI needs a major rework (or retirement)....

i am more of a believer in the early DI, although phil did well there too. but lakisha was the one who dropped alot as the hours passed (suggesting a "false negative" to me)