14 September 2020 #BetweenTheLinesDotVote Analysis Saving Dornsife? The Continued Question of Insignificance Wow! Look how busy I've been this morning. Frankly, I hate charts with TOO MUCH analysis represented. I don't hate the chart below. Still too much, right?
2) If you go slowly with me, though, over the chart, you might end up agreeing that this time, more was actually needed, whereas normally we wish to employ the less is more philosophy. Let's start with the Insignificant Difference Line (IDL). That's a better name for it!
3) The Dornsife geniuses (and they really are!), simply did an abysmal job naming this thing of theirs. To call it an Insignificance Area simply misleads. Nothing could be more significant than whether the two campaigns are too close to call, and we can't quite predict.
4) Done rightly, this factor would be the greatest testament to the integrity of their method. The name they've chosen indicates the opposite. Our numbers don't matter, they're saying, I imagine unwittingly. So, and I think I'll stick with this, I'm calling the IDL.
5) If you care to watch Dornsife's output with me, this may end up being one of the most significant elements of our analysis. On that basis, I've decided to include an image like this every day that I choose to analyze Dornsife. Check this out.
6) So what you see on today's great chart is focused on that factor, more than any other. First, I list out where Dornsife originally assessed themselves. You see this over on the left, with their original numbers from 17 August, at 42.67 and 49.29 as the lower and upper limits.
7) Next you see that, for the first time, I've mapped out the full trajectory of their current upper and lower limit trend lines. The great test I offer Dornsife is to allow the candidates to arrive in the IDL area. My great attack is that they won't do it. I hope to be wrong.
8) In science, a predetermined outcome is the worst possible attack that can be made. I hate digressing here, but I must offer an example. The entire Climate Change question has nothing to do with data, data truth or distortion. It has everything to do with an agenda.
9) The predetermined outcome of Climate Change Pseudo Science is found in the term "Anthropogenic." It's just a latinized version of the term "man made." The predetermined agenda is that man is causing the rise in temperature, assuming you accept that data.
10) There was a great hoax over the fact that white men's craniums were larger than black men's, proving racial superiority. You can check out the history on that hoax, here, if you're interested. jstor.org/stable/pdf/108…
11) The bottom line was that the measurements ended up being driven by the scientist's racism, his personal beliefs caused the manner in which he recorded his data and "proved" his position. This is the profound nature of my potential attack against Dornsife.
12) If my sad theory is correct, they started out with the outcome of Biden's victory as their starting point, and, it was their Insignificance Range where they found their avenue of false proof, supporting their desired outcome. As I always say, I hope to be wrong.
13) In all my wanderings around their website, I have never found an explanation of why their margin of error (MOE) is so much larger than the standard claim of most pollsters, which runs between 3 & 3.5% or so. No clue. Yet, doesn't that begin to look like play room?
14) We'll talk about this more when we look at the Trump campaign's current data. Here, big picture, I do NOT - I say again - need Dornsife to call for Trump's victory in order to believe in them. I just need to break the prediction that his numbers are always below IDL.
15) Turning to Biden's numbers, you can plainly see here that his support line, as I call for it, stretches through the entire campaign, and never falls beneath the upper limit of the IDL. Understand, any formation can be broken at any moment, and all will eventually be.
16) I have to admit, one of the things that makes me love Dornsife is that I'm able to find interpretive formations such as Biden's flag form, shown above. That he's hit and bounced off his resistance line throughout the polling period so far strengthens that line.
17) If he hits, and bounces upward off of his support line, we'll then see it strengthen too. The funny thing about this flag form, that I did NOT see back in 2016, is that it might actually be sustained from start to finish throughout the entire election period. Wow.
18) Then, if you look at the virtually parallel relationship between Biden's support and the upper limit of the IDL, you see where Dornsife is laying down its stake. If that holds, and then, on 4 November he loses, they will have humiliated themselves.
19) Before turning to Trump's chart, I'll say this again, one more way. I can't buy that Biden is that much ahead, and that Trump has no chance to win, which is what their numbers are shouting. Can't buy it. They may be right, but it looks hinky to me.
20) So how might they win back my full support? NOT by putting Trump ahead of Biden. Rather, if they allowed the IDLs to get transgressed, and then sent the data back the other way again, I'd begin to have less doubt. And, let's call for drama! Trump rises over Biden, then falls.
21) If, after allowing Trump to rise over Biden, and then, Biden retook the lead, I'd be quite tempted to believe they did NOT tilt the scales the way I'm fearing they are, right now.
22) Here are Trump's numbers for today. Maybe I'll have to go find it, but this reminds me so much of the 2016 cycle. It just feels like Trump to me. And I want to play with that thought a bit. Follow me out.
23) Let's put forward another theory, a counter-theory. Let's imagine that the Dornsife people continue to be honest scientists, albeit obviously affected by their funding source, the LA Times. That pernicious influence accounted for, imagine they're doing their best anyway.
24) As we've analyzed extensively, they've changed the fundamental nature of their method. So, imagine this much more. The formations are right. That is, Biden's numbers are moving in an accurate manner. Trump's also. The only thing is, due to bias, they're inverted.
25) As a biased Trump supporter, I have long called for a massive landslide in November, for Trump. I won't get into any of the fundamentals, today. I am simply confessing my bias, and my hoped for call. That said, I still have my "what if?"
26) What if Dornsife is polling well, just having inverted who's on top, who's on bottom? It's readily possible. That, in fact, is how I'm going to analyze today's Trump numbers. And the first thing to notice is velocity and momentum.
27) The first thing to note is his widening channel. Resistance rises at a higher slope that support. A wide open channel is the very best possible formation we might ask for. Now, look closely at exactly HOW Trump bounces upward, off his support line.
28) Oh, my goodness, I have to go draw you another chart to illustrate this even more significantly. Oh, holy cow, I have a couple of consulting sessions coming up, now, and I have to break off analysis for the moment. I'll be back later this afternoon, and finish up!
Thread interrupts at #28, and will resume later this afternoon.
29) Joe's chart form, a very traditional, if somewhat extended flag form from today, has stood solid for 36% of the days of the contest period. Both its top slope, and its bottom slope appear to be be 100% in tact. Not to mention the underlyinging IDL. This is phenomenal!
30) It was one of the things I most profoundly noted upon completion of my 2016 analysis, that there were lines that might seriously be drawn across the entirety of the campaign. Here's their final visual post.
31) As you can see, it was the two competing resistance lines that told the entire story in 2016. Even now, it amazes me that an entire election might show itself revealed so powerfully on just two lines.
32) I haven't decided to redraw 2016's lines in color, the way I do them now. Perhaps I will, but for now, just look at the retrospective clarity of their lines. Amazing. I don't know when we'll find lines with that much credibility and force. We may not this season. We'll see.
33) I hasten to remind you that I called for Trump, by Dornsife's numbers and my charts, on 15 October 2016. You know, just looking at these wonderful, old charts, I realize I'm going to have to redraw them in color. I shall. But the thing is, try to imagine just two lines...!
34) Oh my again. This is what I've been looking for, but couldn't remember where to find it. Look at this zoomed in image! Note the phrase "95-percent confidence interval." They have inverted their representation. BAck then, the entire campaign occurred within it.
35) As you can tell, I'm going to have to really let all this sink in. 2016, 95% area of confidence covers the majority of the election. Now, in 2020, we're given this new approach, an Insignificance Range. I'm amazed it took me this much time to see it. More to come...
36) Thread ends at #35.