|
|
#122
05-09-2007, 04:40 PM
|
|||
|
|||
Re: The answer is 42
[ QUOTE ]
Vegan, I am very sorry but this is going to hurt [img]/images/graemlins/frown.gif[/img]. Everyone who wants to understand about DialIdol should open up a second browser window and play along with me. 1. Go to the DI website. 2. Find the orange bar in the upper right and click "DIHardGraphs". This brings you to a graphical display of the results by time period. 3. Set up the controls: Green - Set to week with 4 Idols. 50 intervals will do nicely. Orange - Select only the Jordin #1 and LaKisha #1 boxes at first. Later you can look at whomever you want. Blue - Select the Total box only. Red - Do not select any of these boxes. 4. If you did this correctly the graph should show two lines plotted at roughly 7-minute intervals. The left half is East Coast and the right half is West Coast. 5. This is a graph of the total number of vote attempts reported by DI users in each time interval (not cumulative). Looking at the East Coast we see that Jordin is much more popular with reporting DI users. Don't draw any conclusions from this as DI users often don't follow popular trends. 6. Now look at the West Coast side. Look at the LaKisha data point for 12:03 - 12:10 AM EDT. Hover your mouse over the dot and it will say "LaKisha Jones Interval Total, 12:10:07 AM, 25". That means that there were 25 vote attempts made and reported in the preceding seven minutes. Folks, 25 votes in seven minutes is the work of one or two modems. Add LaKisha #2 to the graph and the story is the same. The West Coast data for LaKisha is based on about three modems. Three modems is not a proper sample. Their high busy rate may reflect one DI user with bad modem hardware, local telephone switch issues, or lack of AI server capacity for his region. It could even be fraud against DI. You can see how there could be money to be made by manipulating the DI results with artificially created busy signals. The West Coast results are statistically insignificant. It's always been this way as you can verify by reviewing past weeks. DI reports from the left coast are rubbish. [/ QUOTE ] Post of the week. Very good analysis. 
|
|
#123
05-09-2007, 04:51 PM
|
|||
|
|||
|
Re: The answer is 42
[ QUOTE ]
[ QUOTE ] Well the numbers are in and if DI is any indication this week, Lakisha may survive. Apparently she made a big come back on the west coast and although she is still at the bottom it is very close. It's either her or Melinda. What a shame though because clearly it should be Blake. He was Gong Show material tonight. [/ QUOTE ] Vegan, I am very sorry but this is going to hurt [img]/images/graemlins/frown.gif[/img]. Everyone who wants to understand about DialIdol should open up a second browser window and play along with me. 1. Go to the DI website. 2. Find the orange bar in the upper right and click "DIHardGraphs". This brings you to a graphical display of the results by time period. 3. Set up the controls: Green - Set to week with 4 Idols. 50 intervals will do nicely. Orange - Select only the Jordin #1 and LaKisha #1 boxes at first. Later you can look at whomever you want. Blue - Select the Total box only. Red - Do not select any of these boxes. 4. If you did this correctly the graph should show two lines plotted at roughly 7-minute intervals. The left half is East Coast and the right half is West Coast. 5. This is a graph of the total number of vote attempts reported by DI users in each time interval (not cumulative). Looking at the East Coast we see that Jordin is much more popular with reporting DI users. Don't draw any conclusions from this as DI users often don't follow popular trends. 6. Now look at the West Coast side. Look at the LaKisha data point for 12:03 - 12:10 AM EDT. Hover your mouse over the dot and it will say "LaKisha Jones Interval Total, 12:10:07 AM, 25". That means that there were 25 vote attempts made and reported in the preceding seven minutes. Folks, 25 votes in seven minutes is the work of one or two modems. Add LaKisha #2 to the graph and the story is the same. The West Coast data for LaKisha is based on about three modems. Three modems is not a proper sample. Their high busy rate may reflect one DI user with bad modem hardware, local telephone switch issues, or lack of AI server capacity for his region. It could even be fraud against DI. You can see how there could be money to be made by manipulating the DI results with artificially created busy signals. The West Coast results are statistically insignificant. It's always been this way as you can verify by reviewing past weeks. DI reports from the left coast are rubbish. [/ QUOTE ] OK, well thanks SW. I have to say that analysis is quite impressive. At least I am not living in fantasyland any more but I am still hoping because your evidence shows that the west coast results are not reliable but it doesn't rule out that Lakisha could have made a come back in the west anyway. I know it's very improbable based on east coast results but I can dream, if even just for a few more hours.......
|
|
#124
05-09-2007, 06:04 PM
|
|||
|
|||
|
Re: The answer is 42
[ QUOTE ]
[ QUOTE ] Vegan, I am very sorry but this is going to hurt [img]/images/graemlins/frown.gif[/img]. Everyone who wants to understand about DialIdol should open up a second browser window and play along with me. 1. Go to the DI website. 2. Find the orange bar in the upper right and click "DIHardGraphs". This brings you to a graphical display of the results by time period. 3. Set up the controls: Green - Set to week with 4 Idols. 50 intervals will do nicely. Orange - Select only the Jordin #1 and LaKisha #1 boxes at first. Later you can look at whomever you want. Blue - Select the Total box only. Red - Do not select any of these boxes. 4. If you did this correctly the graph should show two lines plotted at roughly 7-minute intervals. The left half is East Coast and the right half is West Coast. 5. This is a graph of the total number of vote attempts reported by DI users in each time interval (not cumulative). Looking at the East Coast we see that Jordin is much more popular with reporting DI users. Don't draw any conclusions from this as DI users often don't follow popular trends. 6. Now look at the West Coast side. Look at the LaKisha data point for 12:03 - 12:10 AM EDT. Hover your mouse over the dot and it will say "LaKisha Jones Interval Total, 12:10:07 AM, 25". That means that there were 25 vote attempts made and reported in the preceding seven minutes. Folks, 25 votes in seven minutes is the work of one or two modems. Add LaKisha #2 to the graph and the story is the same. The West Coast data for LaKisha is based on about three modems. Three modems is not a proper sample. Their high busy rate may reflect one DI user with bad modem hardware, local telephone switch issues, or lack of AI server capacity for his region. It could even be fraud against DI. You can see how there could be money to be made by manipulating the DI results with artificially created busy signals. The West Coast results are statistically insignificant. It's always been this way as you can verify by reviewing past weeks. DI reports from the left coast are rubbish. [/ QUOTE ] Post of the week. Very good analysis. [/ QUOTE ] This actually works in vegans favor, lakisha had very little DI users voting for her, her numbers are insignificant west coast and east coast. Look at last week, again she had few voters so we couldnt get a good idea where she stood, and she ended up way understated. also look at New Mexico/Min/ND/NE/IA simply having activity will give you a good score even if you get few busy singnals. Having no activity definaltly hurts your score, and you become a wild card on where you actually stand. looking at the raw numbers last year the lowest was taylor #2 61k total Lakisha this year combined is 33k
|
|
#125
05-09-2007, 06:40 PM
|
|||
|
|||
|
Re: The answer is 42
[ QUOTE ]
looking at the raw numbers last year the lowest was taylor #2 61k total Lakisha this year combined is 33k [/ QUOTE ] do you have these T4 DI numbers from last year handy? post them please if you do.
|
 |
|
|
Linear Mode
