29 September 2020 #BetweenTheLinesDotVote Analysis We Always Pretend The Data Is True If you're going to analyze anything you have to at least pretend the data given is honest, real, worthy. Sometimes the pretense is harder than others. It's okay, we'll find our way. I promise.
2) I have no idea if, like me, you are the kind of person who'd LOVE to go to the sausage factory. I'd do it in a heartbeat. I love eating sausage. I would love to know how its made. I'm that kind of guy. Well, this thread will be a sausage factory style thread. Okay?
3) Diligent student that you must be, you're going to want to read or review yesterday's post. In the coming charts, you'll see how yesterday's calls show up precisely on target in today's work. So, take your time, here's yesterday's work:
4) Please also note the chart above, taken from today's Dornsife posts. It is the one chart they publish daily that I give MOST credibility to. All that covered, here's my own presentation of the exact same data:
5) We're going to spend some time with the chart above. The first thing to note is a vast difference in scale. You've heard about liars, lying liars, and statisticians, right? The way you present the data is a HUGE part of how lying liar statisticians get away with their trade.
6) The Dornsife chart employs a scale with a high of 70 and a low of 30, for a full 40 percentage points range. Now, what does that do? It LOWERS volatility. Volatility = Change. When you wish to DE-EMPHASIZE change, you EXPAND the range you chart. Got that?
7) Now, look at my chart. It's range - which I have REDUCED for today - covers solely a high of 51.5 down to a low of 41.5. That's ONLY 10 percentage points. That is precisely 4 times EXPANSION of the differences as the data moves. My data shows up as 4X more volatile.
8) Why do you need to know these details? Why break your brain over them? The answer is actually quite simple. Truth is not just the data. It is also how the data is presented, but most of all, how YOU understand the data. You can be played. Or you can dive deep and understand.
9) In analyzing the data, my goal is to follow changes, and the more sensitive the change data is, the better for me. I also seek the MOST knowledge, believe it or not, from the LEAST data. You'll be hearing more about that from me in future posts. Be forewarned.
10) Now, let's look at today's overall chart again. It should be recognized that the difference between the two campaigns has been, by Dornsife's analysis, ONLY significant from 8/17 - 8/24. Every bit of their data, they say, is less than the margin for error, ever since.
11) The word "pretend" comes back, now. Why should we even care about data that is by their own definition "not significant?" Here, I return to my own approach. If I'm going to analyze data, my first step is to pretend it is true. I allow myself that pretense, and dive in.
12) With that pretense in place - I'll come back to that later! - what do we see in this chart? The most important thing are three different times that Biden has regained the lead over Trump. Biden starts out massively in front. Rises a little, and then begins to fall.
13) Once, twice, and now a third time, Biden retakes the lead. When charting two campaigns, the crossovers are the real story. That, and the longest term lines. Long term Resistance and Support. What's clear here, in Trump's long term rising Resistance Line.
14) Also, we see a current possible surge for Biden. In fact, let's look at his chart alone, right now.
15) I often explain that the lines I draw require three data points. Biden's channel resistance enjoys 4 points. But, his current surge line is above previous resistance and has an 8-day rise above it, previously. It appears he'll set a new resistance line soon. Wanna see it?
16) Now, let's return to the word "pretend." Allow me to propose my own impact. I believe the Dornsife and the LA Times people watch our work together here. I do. More, I believe they have honed the ability to shift the data in Biden's favor.
17) If I'm right, and if they've noticed that I found the one chart in all their data that actually favors Trump, they have - I propose - the ability to begin affecting that data. There are scores of ways to do so. Lying liar statisticians, and all that.
18) But what if they're right? That question must be asked. My answer then is is something like this. The Biden team has successfully mounted its attack against Judge Amy Coney Barrett. Their base is so against her, that it's showing up here at Dornsife's Expected Winner poll.
19) Let's see what we find in the Trump poll. I have to confess. I LOVE this chart! Yesterday, I did NOT know that Resistance had been found. I called for it. And, look at today. We have our 3 points.
20) Here the word "pretense" comes right back in. Me? I don't believe Trump got smacked down, and I do NOT buy this Resistance line. I think that foul play is showing its hellacious face. If I'm an LA Times executive, and I follow Scopelliti's posts, this is what I'd do.
21) I'd literally smack Scopelliti's work down. Let's look at that. I have clearly stated that FALSE polling was the 2020 strategy. Let's imagine that Dornsife, under the LA Times auspices, is doing everything in its power to project a Biden victory certainty.
22) If that's true, than every possible lever in the Dornsife world must be pulled to project a Biden victory. It's obviously an old story.
23) I certainly have not hammered out the details yet, but everyday I move forward. We are honest pollers, at BTL. We do NOT seek to affect the polling outcome. Rather, we seek only the truth.
24) Honest polling is driven my honest, neutral questions. I've shared them many times, but please, consider them again: 1) Will you vote? 2) Who for? 3) Who should win? 4) Who will win? 5) How do you identify: D,I, or R?
25) And honest analysis. That's what we're about. That's what I do. Please head over, and send everyone you know to:
Thread ends at #25.
Comments