I’m calling it NOW. The Democrat to Win, by 3 to 5 points. Why? The usual Trump factors, plus the latest: The Rexxon firing. A step too far, as he was one of the very few grownups left in the regime. Get out and VOTE, Dems. Most of us can’t, at least not today. Do your part, NOW.
2
I LOVE THE NEEDLE!!! It is by far the most accurate representation of what is happening in real time on election night. I have been checking the needle during every election since I found it. When I wanted to see what was really happening in the PA election tonight, I went straight to the needle. Unfortunately, it was off by the time I got there.
It probably is true that most Americans do not understand probabilities. That is definitely not a reason to cease using them.
If I recall correctly, 538 had Trump as something like a 15% chance to win right before the election.
That is roughly one in six, like rolling a 6 using a single die (in a game like Yahtzee, for example).
Obviously, it's something of a long shot, but it does happen.
I can't help it if many Americans are math-challenged and misinterpret information as a result. I wish it were not so.
But, keep the needle!
9
Here's a reason to dump there needle: it's annoying.
3
I will NOT be watching the needle.
1
if it gives people joy
I guess it's okay
2
I was so upset on election night that, more than once, I attempted to use my cursor to push the needle back onto the proper side. I hate the needle.
3
One of the fondest moments of my political junkie life was watching the needle move from Hilary to Trump on election night 2016.
Beautiful!
1
The needle traumatized me that horrible night in 2016 but brought me joy as it turned towards the blue in last December's special election. I am watching the needle again tonight.
5
I remember looking at that needle in the months before the election and thinking, huh? I like many others thought Hillary Clinton probably was in the lead, but I remember thinking that the crazy-high probability percentage levels it was pointing to were just not plausible, although I know that darned graphic contributed to my feeling positive about the outcome of the election up until the afternoon I voted. Talk about an epic fail.
So I'm not pleased about the return of the needle. I still feel traumatized.
That said, I agree with other readers who think the colors should be reversed. But I'm sure the decision to put them as they are was quite deliberate, so I don't expect that to change. Which gives me another good reason to avoid it!
The needle is just fun!
I like watching it on election nights.
3
I think I'd discard it via a needle exchange program. I'll stick to updated vote counts.
1
The problem is people don't understand how probability works. They see "90% of victory" and think well, no way that person can lose. But of course, there is still a chance the other person will win, it just isn't likely. I don't know why all those morons were calling the times biased for saying Hillary was likely to win. She *was* more likely to win. Doesn't and didn't mean trump couldn't.
4
Though most of our actions from crossing the street to eating out to turning on a light switch are predicated on success-estimate probabilities, they largely involve a non-conscious process of estimation regarding the results. However, when people consciously view a probabilistic situation, especially one they are invested in emotionally, rational probabilities tend to go out the window, the toggle mindset of YES/NO, WIN/LOSE taking over.
People assume that if the odds are 2-1 they'll win, then they'll win. I'd venture to guess that those of us who play poker were less shocked at the outcome of the 2016 election than were non-poker players.
2
I LOVE the jitter idea - a wonderful way to graphically show uncertainty. I wish it had been used in my statistics classes
1
I DO understand statistics, though I'm not an expert.
I greatly dislike the needle because it is misleading. It is misleading because it represents itself as a fine instrument (a needle), when actually it is a very very blunt one. A fine instrument(s) would include a lot more information, who has voted, what are the likely outcomes of uncounted precincts, etc. We need MORE information, not less, not information simplified; simplifying this information creates a lie of omission instead of a good communication tool.
3
"Yes, that night started with Hillary Clinton as a favorite, based on pre-election polls that had given Mrs. Clinton a clear but modest lead. But the needle makes election night forecasts, not pre-election forecasts. It had to start somewhere, and it merely inherited Mrs. Clinton’s presumed edge."
This seems like a lot of double talk. If the needle didn't make \forecasts/ for in the 2016 election, why did you allow it to be printed before election might?
In fact, what's wrong with making forecasts? We understand that the a weather forecast isn't guaranteed to be right.
What was such a problem in 2201 was just how poor the forecasting was. You say that thet 'polls had given a clear but modest lead'. It's quite likely that the polls we off from the beginning. Hence, the edge you gae to Clinton was more presumption than than data.
5
I assure you, you have not turned off the jitters -- I've had them non-stop since Election Night 2016.
10
Thanks, it reminds me I have to gas up on the way home. Who says it’s not useful?
1
This needle is associated with a great trauma of my adult life. On November 8, 2016, I sipped on whiskey soda watching the returns come in, confident that our country would do the right thing. Then, in the later hours, the needle began to lurch ever right-ward and with every update I had another whiskey until all was lost. I hate this needle.
14
Ditch the "needle." We need more journalism, investigative reporting and coverage of what voters want.
1
Most people do not understand probabilities and statistics, and even if we did, the polling and predictions are so often unreliable, it begs the question:
Isn't it rather irresponsible for the media to keep putting these out and sowing confusion?
This isn't a horse race.
2
Is the jitter a moving needle? Instead, show colored band(s) for the statistical ranges....
There is a danger that I'd like to point out in your predictions.
In earlier days, pollsters would make predictions of the outcome, i.e., Bush will get 51% of the vote, and it will result in some modeled electoral college vote.
But now the Times, and others, such as 538 and Huffpost, publish the odds, which is quite a different calculation, and turns a 52% vote margin into a 90% probability (I made those figures up, but you will recognize the directional correctness of them). Then the 90% figure becomes a headline all over the news, on TV, on internet, etc. This, in an era where the polls on which the figure is based are severely unreliable in the era of the cell phone.
Our populace is not particularly well educated on statistics, let alone basic arithmetic. In the popular view, 90% means it's already been decided, and it really isn't necessary to vote, or, if you're going to make the effort, you can go ahead and waste your vote on a third party or a write-in. I know plenty of millennials who did just that.
Wisconsin, Pennsylvania, and Michigan were determined by fewer that 80,000 votes, if I remember correctly, on a total vote of 137 million, which was a new low in turnout, percentage-wise.
I beg you to abandon your computation of the odds because it is actually influencing the results. I am sure you already realize this. So please, stop, and please tell the headlines why you decided to stop.
12
"I love the way you lie" by Eminem/Rhianna (Skylar Grey) captures my relationship with The Needle perfectly.
4
I'd vote to get rid of the needle. Whatever its logic, it encourages misleading conclusions. If your candidate is "very likely" to win, then it gives you permission to stay home, or "safely" vote for a third candidate of conscience, as many likely did in 2016. It discourages the fullest participation possible.
4
The needle doesn't show up until polls have closed.
1
No doubt you will continue to be needled about this.
3
I wish I had taken a course on probabilities or even statistics when in college. But I didn't. So, I understand a projection that, over the course of one year, on 74% of our daily walks--i.e., 270 mornings--my dog will bark. But I don't understand a projection that, on any given day, he is 74% likely to bark.
If Hillary were 55% likely to have beaten Trump, I wound understand that, if the race were run 100 times, she would win 55 times. But I have never understood what it means when the race if run only once.
So, I'll watch the needle and hope for blue, but I'm taking Nate's work on pure faith.
7
Though most of our actions from crossing the street to eating out to turning on a light switch are based on probabilities, they largely involve a non-conscious process of estimation regarding the results. When people consciously view something, especially something they are invested in emotionally, probabilities tend to go out the window, the toggle mindset of YES/NO, WIN/LOSE taking over.
People assume that if the odds are 2-1 you'll win, then you'll win. I'd venture to guess that those of us who play poker were less shocked at the outcome of the 2016 election than were non-poker players.
3
I watched the Needle during the Moore/Jones vote and, although I am a literature person, found it to be self-explanatory. I'll be watching again tonight!
1
What’s the point? The point is giving readers information they want. Should they want that information? This is what critiques are really asking. Take this situation: for many of us, the loss of Ms. Clinton was a tragedy that watching the needle helped us understand, and helped us accept the outcome. Furthermore, the slow reveal and movement of the needle also help us understand voters in unfamiliar states and districts.
On a completely different note, this would not work in my state. We vote entirely by mail, and usually don’t know the result of close elections for days because ballots post-stamped Election Day are valid.
1
After proving conclusively in 2016 that his predictions are not worth the electrons they are printed on, Nate spills more electronic ink explaining why polls, needles, applause-o-meters and other related shiny objects are merely intended to distract us from our impending electoral doom. Nate, we already figured that out.
13
Or you could just say, "I didn't understand before and I still don't understand now."
8
I have to quibble with the positioning of the blue and red sectors. I like to see needles turn clockwise rather than anticlockwise.
6
Interesting point. Especially in cases where folks must select between the two, it could make a difference.
The positioning is simply a result of Democrats being the party of the left in the US, and the GOP being the party of the right.
I have a Master of Science degree, and experience as a computer programmer, and I understand statistical probability relatively well as a result of that training. That said, I love the needle! I mean c'mon, lighten up everyone! It's a great way to engage the electorate in what is otherwise a pretty boring affair: counting millions of votes. I'll be watching the needle tonight, that's for sure.
42
The Bartlett quote "The greatest shortcoming of the human race is our inability to understand the exponential function" will doom us in not fettering our population growth and material consumption. To this I add the inability of the human race to understand probability will play at least an equal role.
18
Polling (further) lost its credibility in the botched Clinton-Trump predictions. Polling now mainly serves to discourage voter turnout by often giving the impression that a race is a done deal before anyone has even voted.
10
Except, of course, the result was perfectly within the margin of error for the polls, which were only off by about 2 points nationally. It was the interpretations by the news media, combined with their general innumeracy, that lead to widespread overconfidence in predicting a Clinton victory.
17
The polling was also accurate in who supported Trump and who supported Clinton; the demographics hadn't changed, but Trump won the electoral college. And Trump only hates polls when he's down in them. When he's up, he loves polls.
4
You are essentially saying that, because polling is inaccurate (wide margin of error,) the 2016 polling data are acceptable because they conform to the predicted inaccuracy.
If a popular election difference of ~3 million votes falls within your polling methodology's margin of error, the utility of that polling is limited. There have been many elections decided by fewer than 3 million votes:
https://en.wikipedia.org/wiki/List_of_United_States_presidential_electio...
Well, lets remember Clinton's own campaign had her trailing in the electoral college early in October before the debates. Everybody knew Clinton needed a 4%+ range to "cleanly" declare victory. The needle offers nothing new there.
2
Polling has lost what respect it had since it totally blew the forecasts for the November 2016 election. Voters have grown smarter, fewer are willing to share what their real thoughts are, and fewer landlines makes outreach to potential voters much more challenging.
4
The 2016 polls were pretty much dead on--well within the margin of error. What went badly wrong were most of the predictions based on them, such as that found here in The Upshot. Don't confuse polling results proper with pundits predictions from those results.
6
Appreciate the "shout out" to John King of CNN- he's been my "go to" guy for years now, as he 'duplicates' your wonkish ( & usually on the money) methodology in many ways.. Will be following you both tonight.
3
whats the whole point of the needle in the first place if it fluctuates critically right up til the last vote is counted? What value does it even bring to the table?
22
The article explains it well enough. The needle predicts what is likely to happen based on the information it has. As it gets more and more information, the model becomes more and more certain until "likely to happen" becomes "what happened". It is worthless, though, if you think "likely to happen" means "certainly will happen".
4
Your question implies that you think the needle should perfectly predict the outcome of the election from the moment it's turned on. If you actually read the article you'll find it explains why you can't have that, and why the needle was never intended as such a tool to begin with.
4
Why not add additional information.
Show the expected turnouts for each party and how their
vote is going and what voters remain to be counted.
5
Why not dress the needle with a band on each side like the hurricane track forecast? You could even shade the color from dense near the needle to faint far from it. This would convey the range of uncertainty and might convey the idea that farther from the needle is considered less likely. (I happen to know what a Gaussian distribution is.)
5
Maybe all we need is for it to be made clear that what the needle shows is how the current returns compare to the baseline estimate. "Doing better/worse than we expected at this point" seems clear to me, without all the asterisks necessary when saying "estimated margin" alone.
3
It does do this, with shaded areas indicating the likelihood of wins of a certain margin (fainter is less likely).
2
My question is why do election results need to be presented as a spectator sport in the first place? Why can't the electoral officials do the counting and report the final result when it's done?
32
Agreed. Actually, rather than a sport, it's an obsession. Campaigning, phone calls, elections, special elections, local elections, elections for the fire department, library, and PTA. It never ends 24/7/365. God bless Democracy, but there has to be a better way.
5
I think it's positive because it gets people involved in their democracy. You always have the choice not to watch, but it does add a more exciting element to the usually dry act of tallying votes. It appeals to younger people, who are more accustomed to "real time" accounts of things. The same young people with low turnout who could stand to get more involved in politics.
1
Given that we're dealing in probabilities, how can readers assess the relative accuracy of various probability models? Especially ones like the Needle, where the number jumps around all night but eventually ends up being 100% accurate with 100% of returns in, regardless of how well or badly designed it is?
2
The Needle is love
The Needle is life
All hail The Needle
11
It's a good explanation, but I would argue that if an infographic requires a nearly 1,900-word explainer...it's not very useful as an infographic.
52
Many people understand it just fine with no explanation. Many do not, so they provided an explanation. That sounds like a reasonable course of action to me.
8
This is only because introductory statistics are not taught to most high school students.
7
Sort of like the weather report... that says rain when it's sunny outside.. OK that is definitely possble. The stuff with Hillary was ridiculous. Hubris goeth before a fall.
Trump's actions which swing (swing state -- haha) may well be what makes a difference. Not so sure that tariffs can save the GOP in PA... while guns are going free. We'll see .. and frankly instead of saying that tariffs won't bring back jobs, a better strategy is to say that the NRA still has a lock on Congress and that young people who are denied alcohol are allowed really dangerous firearms.
Do not discuss global warming (this week), or school choice or healthcare.
1
A very fine explanation which will sadly be lost on so many who refuse to understand the basics of probability theory: predictions can always be wrong, otherwise it would be called certainty theory.
62