# Trump could win the election in a nowcast by FiveThirtyEight **URL:** \ **Category:** Politics & Elections\ **Created:** [July 25, 2016, 1:36pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846 "2016-07-25T13:36:00Z")\ **Posts on this page:** 20\ **Page:** 94
**Author:** ![CarnalK](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/carnalk/32/486_2.png) [@CarnalK](https://boards.straightdope.com/u/CarnalK)\ **Post date:** [November 6, 2016, 7:19pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1862 "2016-11-06T19:19:21Z")
Here is a graph from a couple weeks ago that 538 put out that shiws chances of a Clinton landslide/blowout. Yes it gives a good chance of happening. > **[Election Update: There Are 4 Ways This Election Can End — And 3 Involve...](https://fivethirtyeight.com/features/election-update-there-are-4-ways-this-election-can-end-and-3-involve-clinton-winning/)** > > Want these election updates emailed to you right when they’re published? Sign up here. It’s tempting to say that voters have made up their minds and that the pr… ---
**Author:** ![Merneith](https://avatars.discourse-cdn.com/v4/letter/m/57b2e6/32.png) [@Merneith](https://boards.straightdope.com/u/Merneith)\ **Post date:** [November 6, 2016, 8:02pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1863 "2016-11-06T20:02:42Z")
> [@DSeid](#): > > Bottom line is that deciding which assumptions should apply is key before the math applies. Silver is expressing opinions about which assumptions should apply that place him as an outlier. Thanks very much for spelling all this out! It made things much clearer. ---
**Author:** ![Riemann](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/riemann/32/3133_2.png) [@Riemann](https://boards.straightdope.com/u/Riemann)\ **Post date:** [November 6, 2016, 8:47pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1864 "2016-11-06T20:47:35Z")
There was a huge move on Betfair on the “email probe over” news. Trump from 21% to 16%. Is it that significant? ---
**Author:** ![JKellyMap](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/jkellymap/32/15880_2.png) [@JKellyMap](https://boards.straightdope.com/u/JKellyMap)\ **Post date:** [November 6, 2016, 8:56pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1865 "2016-11-06T20:56:32Z")
> [@Riemann](#): > > There was a huge move on Betfair on the “email probe over” news. Trump from 21% to 16%. Is it that significant? This surprises me. I didn’t think _anyone_ was waiting with bated breath to see what was on that computer (supposedly Weiner’s). I just took it as a news cycle that had its impact, and now, ten days later, we’re in the natural “recovery” phase – meaning, a slight uptick in Hillary’s chances from a few days ago, but finding out or not finding out what is or isn’t on that computer is immaterial. Perhaps there are a few Hillary supporters whose morale is slightly boosted by this, so now they might follow through on some local GOTV volunteering over the next couple of days, but if it weren’t for this they wouldn’t quite have bothered – but what’s that translate into, a few thousand votes scattered here and there? Whatevs. ---
**Author:** ![Riemann](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/riemann/32/3133_2.png) [@Riemann](https://boards.straightdope.com/u/Riemann)\ **Post date:** [November 6, 2016, 8:59pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1866 "2016-11-06T20:59:46Z")
It moved back a little to ~18%. I guess the volatility is a reflection of how nervous and uncertain everyone feels at this point. ---
**Author:** ![JKellyMap](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/jkellymap/32/15880_2.png) [@JKellyMap](https://boards.straightdope.com/u/JKellyMap)\ **Post date:** [November 6, 2016, 9:00pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1867 "2016-11-06T21:00:36Z")
> [@Riemann](#): > > It moved back a little to ~18%. I guess the volatility is a reflection of how nervous and uncertain everyone feels at this point. Yes – that makes sense. Thanks for the update. ---
**Author:** ![asahi](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/asahi/32/8693_2.png) [@asahi](https://boards.straightdope.com/u/asahi)\ **Post date:** [November 6, 2016, 9:28pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1868 "2016-11-06T21:28:27Z")
> [@JKellyMap](#): > > This surprises me. I didn’t think _anyone_ was waiting with bated breath to see what was on that computer (supposedly Weiner’s). I just took it as a news cycle that had its impact, and now, ten days later, we’re in the natural “recovery” phase – meaning, a slight uptick in Hillary’s chances from a few days ago, but finding out or not finding out what is or isn’t on that computer is immaterial. > > Perhaps there are a few Hillary supporters whose morale is slightly boosted by this, so now they might follow through on some local GOTV volunteering over the next couple of days, but if it weren’t for this they wouldn’t quite have bothered – but what’s that translate into, a few thousand votes scattered here and there? > > Whatevs. Basically it was 9 days of nothing, and now finally confirmation that Hillary probably won’t face any serious legal challenges coming from the FBI. Voters who aren’t already disillusioned with voting altogether can feel better about voting for Hillary but the damage, whatever it is, has probably been done. As I commented on the FBI thread, the real danger was never Hillary voters switching teams; rather, it was the danger of a suppressed vote. That danger still remains though it looks like democratic partisans have remained loyal. It’s independent middle class voters who could still tip this election. ---
**Author:** ![Riemann](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/riemann/32/3133_2.png) [@Riemann](https://boards.straightdope.com/u/Riemann)\ **Post date:** [November 6, 2016, 9:44pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1869 "2016-11-06T21:44:32Z")
> [@asahi](#): > > Basically it was 9 days of nothing, and now finally confirmation that Hillary probably won’t face any serious legal challenges coming from the FBI. Voters who aren’t already disillusioned with voting altogether can feel better about voting for Hillary but the damage, whatever it is, has probably been done. As I commented on the FBI thread, the real danger was never Hillary voters switching teams; rather, it was the danger of a suppressed vote. That danger still remains though it looks like democratic partisans have remained loyal. It’s independent middle class voters who could still tip this election. I don’t dispute this, but did any of that really hinge on what was maybe on this computer? It’s not as though it might have contained an email saying that Hills personally organized Benghazi as a false flag operation. We all kind of know that Hillary is a bit dodgy and sometimes less than honest but not some traitorous evil criminal, is that ever likely to change? ---
**Author:** ![Riemann](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/riemann/32/3133_2.png) [@Riemann](https://boards.straightdope.com/u/Riemann)\ **Post date:** [November 6, 2016, 9:55pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1870 "2016-11-06T21:55:11Z")
Here’s the Betfair link in case anyone’s interested, real-time 24hr betting market, it has cleared $150 million on the presidency. Prices are in decimal odds, so just invert the number to get an implied percentage (Trump currently 6.0 = 16.7%) > **[Betfair Exchange | Best Odds Online, Back and Lay Betting](https://www.betfair.com/exchange/plus/#/politics/market/1.107373419)** > > The Betfair Exchange allows you the opportunity to back and lay bets and Cash Out. We also offer Casino, Poker, Games and Bingo. ETA: they used to block the site from even being viewed by US IP addresses, but that seems to have changed. ---
**Author:** ![asahi](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/asahi/32/8693_2.png) [@asahi](https://boards.straightdope.com/u/asahi)\ **Post date:** [November 6, 2016, 9:55pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1871 "2016-11-06T21:55:32Z")
> [@Spice\_Weasel](#): > > Maybe there are legitimate concerns about the models Silver is using, but that HuffPo article was an unwarranted hit job and I don’t blame Silver for getting pissed. The media seem to be turning on him because they don’t like Trump’s chances, not because they have a more in-depth understanding of polling statistics than Silver does. I think the media are also making assumptions based on national polling and what they think they know about the individual states. All Nate Silver has been saying all along is that while he believes, like most others do, that Clinton will win, there is also a real possibility that the numbers aren’t telling the whole picture, and that a number of races could be close enough that this fact could produce a surprise on election night. Pennsylvania is a good example. Most people, including myself, believe that Hillary will once again take the Keystone State, just like all democrats have since Reagan. However, there really and truly a growing body of data that seem to show PA can no longer be taken for granted. This goes back to what Silver has referred to as the “unusual nature of Trump’s campaign.” And that unusual nature is owing to the fact that the state’s blue collar working class, which has been staunchly democratic platform of democratic support over a period of not years but decades, is showing signs of cracking, and in Trump’s favor. This is not just unusual, it’s potentially momentous and game-changing. Again, Silver isn’t really betting against Clinton’s odds of winning the state because she believes that women, people with college degrees, and minority voters will ultimately be a broad coalition that will help her one. But even within the minority demographics that have been Clinton’s strength, there are signs of lower African American turnout in key swing states, even as Hispanic turnout has surged to record levels. This is another potential headache. And if that wasn’t enough, again, in PA last time around, Johnson and Stein together had about 1 percent of the vote. In 2016, they hold a lot of voters in their camps, and how they end up voting nobody really knows. I think the problem that I am seeing in this debate over predictions is that people are relying too heavily on the quantitative analysis and not factoring enough of the qualitative aspects of this campaign that could ultimately impact the final math. You can look at trend lines and say Clinton averages +4, and that some news events push her down to +1 and others push her up to +6. I get all that. But there are real qualitative factors that move these numbers. They’re not set in stone, and trends aren’t going to predict with certainty who actually shows up and votes on election day. Hedging is the smart move, even if it’s to the chagrin of Huff Post. ---
**Author:** ![Fiveyearlurker](https://avatars.discourse-cdn.com/v4/letter/f/da6949/32.png) [@Fiveyearlurker](https://boards.straightdope.com/u/Fiveyearlurker)\ **Post date:** [November 6, 2016, 10:18pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1872 "2016-11-06T22:18:33Z")
> [@Riemann](#): > > Here’s the Betfair link in case anyone’s interested, real-time 24hr betting market, it has cleared $150 million on the presidency. Prices are in decimal odds, so just invert the number to get an implied percentage (Trump currently 6.0 = 16.7%) > > [https://www.betfair.com/exchange/plus/#/politics/market/1.107373419](https://www.betfair.com/exchange/plus/#/politics/market/1.107373419) > > ETA: they used to block the site from even being viewed by US IP addresses, but that seems to have changed. Is the “back all” column the relevant one? Just trying to figure this out! ---
**Author:** ![Sherrerd](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/sherrerd/32/8_2.png) [@Sherrerd](https://boards.straightdope.com/u/Sherrerd)\ **Post date:** [November 6, 2016, 10:46pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1873 "2016-11-06T22:46:25Z")
> [@Riemann](#): > > I don’t dispute this, but did any of that really hinge on what was maybe on this computer? No. What has changed is that Comey’s first letter gave Trump the ghost of a hint of a suggestion of something approaching Legitimacy. Now that’s been taken away. Trump and his surrogates will go on claiming that Hillary is about to be indicted, but their audiences will have lost any real-world rationale for convincing themselves that this might actually be true. They will want to believe that it’s true, but they won’t have any outside-the-bubble evidence (such as an FBI Director saying ‘the investigation is reopened’) to use as justification for the belief. They will have to go back to subsisting on pure fantasy. Many of them will go on doing just that. But they’ve lost the tool Comey’s letter had given them for use in trying to recruit those still clinging to reality (i.e. the ‘undecided voters’). ---
**Author:** ![Riemann](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/riemann/32/3133_2.png) [@Riemann](https://boards.straightdope.com/u/Riemann)\ **Post date:** [November 6, 2016, 10:56pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1874 "2016-11-06T22:56:04Z")
> [@Fiveyearlurker](#): > > Is the “back all” column the relevant one? Just trying to figure this out! The blue column is the “back” column, the pink column is the “lay” column. “Back all” and “Lay all” are just buttons that allow arbitrageurs to rapidly back or lay the entire field if the relative odds among different runners get out of whack (above or below 100%). Right now the back-lay spread on Hillary is 1.20 - 1.21. If you want to back her, you risk a dollar to make 20 cents; if you want to lay her, you risk 21 cents to make a dollar. So, the implied percentage chance for Hillary is 1/1.20 = 83.3% if you want to back her; 1/1.21 = 82.6% if you want to lay her. Likewise, Trump is back-lay 5.9-6.0, or back-lay 16.9%-16.7% ---
**Author:** ![DSeid](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/dseid/32/20194_2.png) [@DSeid](https://boards.straightdope.com/u/DSeid)\ **Post date:** [November 6, 2016, 11:06pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1875 "2016-11-06T23:06:57Z")
Silver early on when the FBI letter story broke [stated](http://fivethirtyeight.com/features/election-update-four-ways-forward-for-clinton-after-the-fbi-news/)that there was “probably just enough time left to measure the initial impact” but not enough time to “capture the reaction to the reaction …” Clinton will not talk this up at all. From her perspective the less said the better and the value is that it will force lots less said by Team Trump other than spluttering about how Comey is back to being a piece of dung. Negative news pressure on Clinton removed polls would be returning to the equilibrium of 4, if only there was a chance to measure them before election day. ---
**Author:** ![DSeid](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/dseid/32/20194_2.png) [@DSeid](https://boards.straightdope.com/u/DSeid)\ **Post date:** [November 6, 2016, 11:33pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1876 "2016-11-06T23:33:06Z")
> [@adaher](#): > > … Silver believes undecideds won’t break evenly, or at least that they might not. So if they do break mostly for one candidate, that’s a win for him regardless of who wins the election. … Agreed. I also do want to reinforce that I agree with those who think that Wang’s 99% is silly. But perhaps one should use the same averaging or using the media of the aggregators that one does for polls. Silver is an outlier in his uncertainty. We deal with an outlier poll like the LA Times/USC one not by ignoring it but by subjecting to inclusion as one of several that use different approaches. Probably that is the best approach here too. Yes, polls may be miscounting some demographics, underestimating how likely some are to vote and overestimating others, and _that_ does introduce some uncertainty. I personally believe the errors will on net favor Clinton but we will see. Meanwhile yes, Wang’s model has been predicting a C+4 +/-3 and a result significantly outside that, inclusive of a Clinton win by more than 7 as much as of a Trump win, would mean that he underestimated the uncertainty, and that Silver’s higher uncertainty was more correct (even if it turns out to be for the wrong reasons). FWIW \*\*Spice Weasel \*\* I do not think Silver has been “turned on” because he is saying something that we do not want to hear but because he seems to have been playing more to the media talking head have to keep it a horse race narrative motif. One more thing. 538 (not Silver) [did post something](http://fivethirtyeight.com/features/most-voters-havent-changed-their-minds-all-year/) pretty much in support of Wang’s perspective the other day that I had missed first go round. > [@](#): > > But how much have these story lines actually changed the race? That’s a hot topic these days, with some polls showing dramatic swings toward Trump while others have hardly budged. Writing at YouGov, Ben Lauderdale and Doug Rivers make the case that a lot of what we are seeing is driven by changes in who responds to surveys, not in who voters intend to support. If so, this presidential race might be more stable than you would think by looking at polls on a given day. … > > … The table below shows the responses for the 1,227 respondents who completed our October poll, with the rows showing their candidate support in January and the columns indicating candidate support this October. … around 80.7 percent of respondents gave the same answer in both surveys. … Notice another thing: There is very little movement between the two major-party candidates. 0.5 percent of respondents were with Clinton in January and are now with Trump, while only 1.5 percent of our respondents moved from Trump to Clinton. There is far more movement from not supporting either candidate to backing one than from supporting one major-party candidate to supporting the other. That, in turn, stabilizes the horse race to some degree: Voters who can be persuaded to switch sides are twice as influential, since they affect both candidates’ tallies simultaneously. … > > … the people who refuse to back a major-party candidate are lower-propensity voters, on average. … Pundits commonly call people who won’t commit to a major-party candidate “undecided,” but this data suggests that many of these folks have decided. They’ve simply decided to not support either Trump or Clinton. … ---
**Author:** ![asahi](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/asahi/32/8693_2.png) [@asahi](https://boards.straightdope.com/u/asahi)\ **Post date:** [November 6, 2016, 11:34pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1877 "2016-11-06T23:34:36Z")
> [@Riemann](#): > > I don’t dispute this, but did any of that really hinge on what was maybe on this computer? It’s not as though it might have contained an email saying that Hills personally organized Benghazi as a false flag operation. We all kind of know that Hillary is a bit dodgy and sometimes less than honest but not some traitorous evil criminal, is that ever likely to change? With her partisans, the story made no difference – we saw it as yet another politicized drama that attempted to undercut her legitimacy. But candidates need a voting bloc that will inevitably include people who are less enthusiastic – we need those less enthusiastic voters. It’s not just that we need them not to vote for Trump or Johnson, but we need them to go to the polls and actually vote for Hillary. Polls are nothing. What matters is how the electorate looks on election day and that can be different from what is reported by polls. The story never threatened to move voters to Trump (or if it did the number must be minuscule), but it has threatened to keep voters at home or to vote for third party protest candidates, neither of which are good results. In a race where it’s candidate one versus candidate two, maybe it doesn’t matter. But in a race where ‘consumers’ have ‘choices’ anything that dampens enthusiasm for a candidate is magnified. True, most voters know that Clinton had misused a private email server – that’s not news. But they thought she hadn’t acted criminally and believed that her presidency wouldn’t be impacted by a criminal investigation that could, perhaps, cast doubt on her legitimacy. When the FBI said they were opening the investigation again, those old questions about whether she would be able to finish her term came back again. People genuinely don’t want to have doubts about whether a president can function and complete a term in office. That’s how the story could have turned into (could even still be) a vote suppressor. Also, do the math. If you give Clinton a slight lead in the polls but Trump is 6-8 percent more likely to get his voters enthused enough to show up at the voting booth, that margin could easily be the difference between winning and losing. ---
**Author:** ![Happy\_Lendervedder](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/happy_lendervedder/32/8488_2.png) [@Happy\_Lendervedder](https://boards.straightdope.com/u/Happy_Lendervedder)\ **Post date:** [November 6, 2016, 11:58pm UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1878 "2016-11-06T23:58:54Z")
Slow day for new polls from toss-up states. Tomorrow’s gon’ be a humdanger. ---
**Author:** ![DSeid](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/dseid/32/20194_2.png) [@DSeid](https://boards.straightdope.com/u/DSeid)\ **Post date:** [November 7, 2016, 12:33am UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1879 "2016-11-07T00:33:06Z")
> [@Lantern](#): > > Thanks. For some reason he has rounded out the margins but it should still be sufficient to compare to the others models. Huffpo also has projected margins so we should be able to compare all the four main models. > > Can’t find any projections on Predictwise. Incidentally that site is such garbage when it comes to visual design. They have a unique dataset so I wish they would partner with a bigger site to present it properly. > [@Tzigone](#): > > That’s exactly what I was interested in. Not even just for swing states, though they’d be most important. A great source! Not margins but [The UpShot](http://www.nytimes.com/interactive/2016/upshot/presidential-polls-forecast.html?_r=0#other-forecasts) has a beautiful table comparing the different models probability predictions state by state. That’s 56 individual probabilities given. ---
**Author:** ![Tzigone](https://avatars.discourse-cdn.com/v4/letter/t/839c29/32.png) [@Tzigone](https://boards.straightdope.com/u/Tzigone)\ **Post date:** [November 7, 2016, 12:45am UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1880 "2016-11-07T00:45:47Z")
> [@](#): > > A great source! Not margins but The UpShot has a beautiful table comparing the different models probability predictions state by state. Thanks. I’ve seen this one. I don’t want who they call for the state for - I want percentage of votes for each candidate in each state. If one both sites say \>99% for dem win and one site says 54% of the vote will be dem and the other says 64% will vote dem and the state goes dem at 68%, then the 64% is clearly closer. That’s what I want - what percentage of vote believed to go to each candidate for each state. ---
**Author:** ![DSeid](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/dseid/32/20194_2.png) [@DSeid](https://boards.straightdope.com/u/DSeid)\ **Post date:** [November 7, 2016, 2:50am UTC](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846/1881 "2016-11-07T02:50:36Z")
Personally I see that as not the most important metric. What I want to know is whether or not aggregator A saying 85% probability of X predicts X about 85% of the time and predicts not-X about 15% of the time or not. But [this Excel file](http://election.princeton.edu/code/data/EV_stateprobs.csv) lists Wang’s state and district predictions using his median approach. The Democratic margin is the second column. The first is his today win probability and the last his election day one. (Now both the same.) Some fun comparisons. Percent is percent chance of Clinton winning. CA both at C+22.5 and 100% (PEC rounds and 538 states \>99.9%) IL PEC C+15, 100%; 538 C+13.5 98% (Really, even a one out of fifty seems high.) KY PEC T+22.5 0%; 538 T+19 0.3% NV PEC C+1 63%; 538 T+0.2 49.2% OH PEC T+1 37%; 538 T+2.3 33.5% VA PEC C+5 95%; 538 C+4.9 82.3% So on. Does it really seem like VA, a state they give the same margin on, has a nearly one in five chance of going Trump? Or is one in twenty more like it? [Previous page](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846.md?page=93) [Next page](https://boards.straightdope.com/t/trump-could-win-the-election-in-a-nowcast-by-fivethirtyeight/760846.md?page=95)