Posted on 01/30/2021 12:39:07 PM PST by grundle
1. You're using cumulative graphs that are great for purpose. By definition, a cumulative graph can never decrease as long as the data input is perfect. I don't get bent out of shape with a little imperfect chatter caused by zero change from one time to the next sampling time but these graphs are suggesting significant intervention.
2. The Likes line may be the only unaltered data of the three groups.
3. The Dislikes line suggests two separate interventions. First, around 0600 the is a single lump subtraction.
After this large subtraction, it appears like a separate, possibly automated, function is switched on that keeps the dislikes at a fixed level. When the dislikes increase to a fixed set point a constant number is subtracted from the total. The sawtooth pattern supports this possibility.
4. I almost want to say that the total views curve has some truth in it. You mention that there is some time delay between realtime data logging and display of results. This is accurate of course, but I have an alternate possibility.
Around 0400(ish) the cumulative count freezes for about 2 hours then starts back up at apparently the same number. Logically, what you would expect Is that after a pause, instead there would be a positive sloped line upwards to a larger number instead of a horizontal line restarting at the same value. This suggests that the entire total views line is fabricated.
The shape of the total views line is very much that of a polynomial function equation. I've worked lots with these equations fitted to data sets and it would be easy to set it up to automatically produce the before and after curves at the pause breakpoint.
5. What is up with a two hour pause? This makes no sense to me. With the notion that there is manipulation, the code could be manipulated in minutes via line editing or even faster if it was formalized by setting up a menu whereby you can just type in factors and hit the enter button. What's magic about 2 hours!?! Agree, disagree?
1. You're using cumulative graphs that are great for purpose. By definition, a cumulative graph can never decrease as long as the data input is perfect. I don't get bent out of shape with a little imperfect chatter caused by zero change from one time to the next sampling time but these graphs are suggesting significant intervention.
2. The Likes line may be the only unaltered data of the three groups.
3. The Dislikes line suggests two separate interventions. First, around 0600 the is a single lump subtraction.
After this large subtraction, it appears like a separate, possibly automated, function is switched on that keeps the dislikes at a fixed level. When the dislikes increase to a fixed set point a constant number is subtracted from the total. The sawtooth pattern supports this possibility.
4. I almost want to say that the total views curve has some truth in it. You mention that there is some time delay between realtime data logging and display of results. This is accurate of course, but I have an alternate possibility.
Around 0400(ish) the cumulative count freezes for about 2 hours then starts back up at apparently the same number. Logically, what you would expect Is that after a pause, instead there would be a positive sloped line upwards to a larger number instead of a horizontal line restarting at the same value. This suggests that the entire total views line is fabricated.
The shape of the total views line is very much that of a polynomial function equation. I've worked lots with these equations fitted to data sets and it would be easy to set it up to automatically produce the before and after curves at the pause breakpoint.
5. What is up with a two hour pause? This makes no sense to me. With the notion that there is manipulation, the code could be manipulated in minutes via line editing or even faster if it was formalized by setting up a menu whereby you can just type in factors and hit the enter button. What's magic about 2 hours!?! Agree, disagree?
It is.
There is a too large segment of the population who are swayed by such things as the ratio of likes to dislikes,the tabloids, People magazine,the View,etc.
Taking into consideration differences in criteria, perhaps the best general summary of this research indicates that 2.5 to 3 percent of the general population is mentally retarded.
Those are the folks that buy into that stuff.
Sue them. It’s official record.
You need a control group to show how a non-manipulated video responds to your collection.
Those graphs with a gentle arc interrupted by a vertical drop look suspiciously like the *ahem* election night vote tally.
The saw-tooth pattern following the drop would suggest YouTube is running a periodic cron task to keep “pruning” down the dislike count.
The deceit is everywhere.
In addition to a control group ... it would be nice to see the comparative timing and amount of reduction taking place in the zig zaggy portion of the graph.
brilliant. thanks for your work.
the patterns are telling.
expect more of the same.
“The timezone for the charts is UTC (London). Did you notice the huge drop at 06 hour (1 AM US Eastern)? Most working people go to sleep by that time. Coincidence? I think not.”
In other words they’re using Dominion software?
How do you know that they (or other ratbots) don’t manufacture the like numbers? They’ve shown they have no problem making up numbers like that - look at the phony-baloney election numbers for Beijing Obiteme.
It’s flatline across the bottom of the graph, not increasing.
Google has poisoned the Web
Considering how many keep voting for Democrats and Rinos despite ample proof of the lies, I say you’ve misplaced a decimal.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.