How To Win An Election

Excerpted from “How To” by Randall Munroe. Copyright © 2019 by Randall Munroe. All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.

Winning elections is hard. The truth is, people are complicated, there are a lot of them, and no one is ever 100 percent sure why they do what they do or what they’re going to do next.

But if your goal is simply to win an election, then as a general rule you should be for things that voters like and against things they dislike. To do that, you’ll need to figure out what the voters like and dislike. One of the most popular tools for figuring out what the public thinks is opinion polling — talking to a bunch of people, asking them what they think and tallying up the results. The website FiveThirtyEight has conducted an exercise in which they had professional speechwriters write a speech that simply pandered as much as possible — only making statements that most voters support, to pander either to one party or to the electorate in general.

But what do we agree on the most? If your goal is simply to be in favor of popular things and against unpopular things, what should you campaign on? What are the least controversial issues in the country?

To help figure this out, I reached out to Kathleen Weldon, director of data operations and communications at the Roper Center for Public Opinion Research at Cornell University, to commission a poll of their polls. The Roper Center maintains a tremendous database of opinion polling data — over 700,000 polling questions spanning almost a century of opinion polling, collected from virtually every organization that has ever conducted a public poll in the United States.

I told them I was looking for the most one-sided questions in their polling database — the questions where virtually everyone gave the same answer. In a sense, these would be the least divisive issues in the country.

The Roper research staff sifted through their database of 700,000 questions and assembled a list of those questions for which at least 95 percent of respondents gave the same answer.

It’s pretty rare for that many respondents to agree on anything in a poll. A small percentage of respondents will often choose ridiculous answers because they’re not taking the poll seriously or because they misunderstand the question. But one-sided questions are also rare because no one bothers to conduct polls on uncontroversial topics unless they’re trying to prove a point. Since everything in the Roper database is something that some person or organization bothered to commission a poll to ask, it means it’s at least potentially controversial, if not actually so.

Here is a selection of the most one-sided issues in the history of polling. If you want to run for office, these are views you can safely espouse, secure in the knowledge that at least one scientific survey puts the people squarely behind you:

  • 95 percent disapprove of people using cell phones in movie theaters. (Pew Research Center’s American Trends Panel Poll, 2014)
  • 97 percent believe there should be laws against texting while driving. (The New York Times/CBS News Poll, 2009)
  • 96 percent have a positive impression of small business. (Gallup Poll, 2016)
  • 95 percent believe employers should not be able to access the DNA of their employees without permission. (Time/CNN/Yankelovich Partners Poll, 1998)
  • 95 percent support laws against money laundering involving terrorism. (Washington Post Poll, 2001)
  • 95 percent think doctors should be licensed. (Private Initiatives & Public Values, 1981)
  • 95 percent would support going to war if the United States were invaded. (Harris Survey, 1971)
  • 96 percent oppose legalizing crystal meth. (CNN/ORC International Poll, 2014)
  • 95 percent are satisfied with their friends. (Associated Press/Media General Poll, 1984)
  • 95 percent say that “if a pill were available that made you twice as good looking as you are now, but only half as smart,” they would not take it. (Men’s Health Work Survey, 2000)
  • 98 percent believe adults should watch swimmers rather than reading or talking on the phone. (American Red Cross Water Safety Poll, 2013)
  • 99 percent think it’s wrong for employees to steal expensive equipment from their workplace. (NBC News Poll, 1995)
  • 95 percent think it’s wrong to pay someone to do a term paper for you. (NBC News Poll, 1995)
  • 98 percent would like to see a decline in hunger in the world. (Harris Survey, 1983)
  • 97 percent would like to see a decline in terrorism and violence. (Harris Survey, 1983)
  • 98% would like to see an end to high unemployment. (Harris Survey, 1982)
  • 95 percent would like to see an end to all wars. (Harris Survey, 1981)
  • 95 percent would like to see a decline in prejudice. (Harris Survey, 1977)
  • 95 percent don’t believe Magic 8 Balls can predict the future. (Shell Poll, 1998)
  • 96 percent think the Olympics are a great sports competition. (Atlanta Journal-Constitution Poll, 1996)

You can use this list to assemble a campaign platform. For example, you could stand firmly against hunger, war and terrorism; for friendship and small business; and against texting while driving. You could support laws that ensure doctors are properly licensed and oppose allowing other countries to invade.

On the other hand, if you wanted to lose an election as spectacularly as possible, this list could be even more helpful as a blueprint. By taking the opposing position on each issue, you could potentially run the most unpopular political campaign in political history. You’d probably lose, but who knows!

The Saints And Steelers Lost Their Franchise QBs. Can They Still Make The Playoffs?

On Sunday, injuries at the NFL’s most valuable position — to two of history’s most prolific passers — radically altered the landscape of the season.

Sean Payton is about to coach only the second game of his NFL career without a healthy Drew Brees, who is out six-to-eight weeks with an injury to his right (throwing) thumb that will require surgery. With Brees at the helm, Payton’s record is 119-72, and the pair has the second-most wins of any QB-coach combo in NFL history. Now, Payton’s quarterback is expected to be Teddy Bridgewater, who last started an NFL game that mattered in the 2015 season.1

In Pittsburgh, the situation is even worse: Ben Roethlisberger will undergo season-ending surgery on his throwing elbow. Steelers coach Mike Tomlin is 115-60-1 with Roethlisberger starting, the fourth-most wins of any QB-coach combo. Tomlin has had to guide his team through Big Ben’s missed time before, going 10-8 in games without Roethlisberger. But those previous fill-ins included veterans Charlie Batch, Byron Leftwich and an over-the-hill Michael Vick. Current backup Mason Rudolph, a third-round pick in the 2018 NFL draft, has never started an NFL game.

What can we expect from each team with their current backup QBs? The Saints still have a coin-flip’s chance to make the playoffs, according to FiveThirtyEight’s NFL prediction model. But the Steelers’ playoff chances are now practically nil — down to just 8 percent.

Even with Roethlisberger, the Steelers had started the season so poorly that, had he not gotten hurt, they still would have been projected for just 7.54 wins, or a record of roughly 8-8 — likely falling short of the postseason again. With Rudolph under center, the Steelers’ outlook drops by 2.47 expected wins. Because Rudolph has not yet started a game, the model projects his performance based purely on where he was drafted (the 3rd round) — which could be too bearish an estimate. But as things sit, Rudolph is expected to perform at only 36.6 percent of Roethlisberger’s Elo level, a rolling average of recent performances that incorporates both passing and running.

On paper, the Saints are in much better shape with their backup in Bridgewater, who is expected to perform at 56.5 percent of Brees’s level.2 With Brees starting every remaining game, the Saints would have been the seventh-strongest team in the league by Elo, with an expected win total of 9.62. With Bridgewater in, the Saints drop to 21st overall — and he lowers their projected win total to 8.93 for the full season.

But Bridgewater’s performance is almost impossible to project, given that he’s not the same player after recovering from the “horribly grotesque” injury that he sustained in 2016. Though the sample size is small — only 53 passes to date — Bridgewater’s returns as a Saint have been poor: He has notched nearly 2 yards fewer per pass attempt than as a Viking and has performed worse both by passer rating and ESPN’s QBR.

The Saints could turn to a Sean Payton favorite, quarterback and Swiss Army knife Taysom Hill, who receives zero Elo points because of his lack of NFL starts and his undrafted status. Payton, though, sees the next Steve Young. Hill is already 29 and in his third season of being groomed by Payton, but Young didn’t begin his Hall of Fame-caliber play as the 49ers’ primary starting QB until age 30. So if it’s ever going to happen for Hill, why not now?

For the Steelers, although Roethlisberger says he’s returning from this injury to finish out his contract, there are no guarantees. Eerily, the playing career of the last great Steelers quarterback, Terry Bradshaw, ended with an almost identical injury in the last game of the 1983 season. And Bradshaw also said he’d come back after his surgery, but he never did. He was 35. Roethlisberger is 37.

But what if the Steelers were able to get a veteran quarterback for next-to-nothing in a trade? And what if that player is a two-time Super Bowl champion, like Roethlisberger? The Giants are finally turning the page on the Eli Manning era, and it would be far less awkward for them to hand the team to rookie Daniel Jones with Manning gone than with him holding a clipboard. Elo projects Manning at nearly twice the value going forward this year as Rudolph, and he would be the AFC North’s third-best quarterback today in terms of Elo, ahead of Andy Dalton.

It appears, though, the Steelers lack the cap room to trade for Manning or even a modestly paid veteran QB. Pittsburgh sent a first-round pick to Miami on Monday for cornerback Minkah Fitzpatrick, so it doesn’t appear that the Steelers are intent on tanking. They’re probably just stuck with the uncertainty of a QB prospect who has yet to start a game.

Neil Paine contributed research.

QB Injuries And Absurd Replay Reviews Dominated The NFL’s Week 2

sara.ziegler (Sara Ziegler, sports editor): Aside from a Jets-Browns matchup that I know we’re all excited about, Week 2 is in the books. And boy, did it have it all. Inexplicable replay decisions! Calls blown dead that should not have been! Fires on the sidelines!

So let’s get into it. How about the replay situation? Who could have possibly foreseen this being a huge mess?

joshua.hermsmeyer (Josh Hermsmeyer, NFL analyst): It’s funny — I was watching Tony Dungy on the Sunday Night Football halftime show going over plays that he deemed clear and convincing evidence of pass interference, and I disagreed with every one of them.

neil (Neil Paine, senior sportswriter): 😬

Salfino (Michael Salfino, FiveThirtyEight contributor): The cost of replay is being able to live the game you’re watching in the moment. And since replay is far from perfect and can never be perfect, why do we tolerate it?

neil: Because the alternative is missed calls! Except now you can inject questionable calls where there originally were none.

Take Sunday’s Seahawks-Steelers game. Trailing in the fourth quarter, Pete Carroll was able to successfully challenge this no-call, and it set up a go-ahead score. In real time, that didn’t look like pass interference. But in slow motion it did, and that is one of the big gray areas for this rule: How much contact is allowed before it becomes a “clear and obvious” hindrance to the receiver? In slow motion, things tend to look more “clear and obvious” than perhaps they were at game speed.

sara.ziegler: I’m reminded yet again that football is so arbitrary in so many ways.

Salfino: If you’re a coach, how do you know when to challenge? There is some kind of contact on most contested catches, if we’re going to Zapruder things.

neil: Yeah, so do you basically save it for any crucial incompletion in traffic late in a game, and just challenge no matter what on the off chance it works?

Salfino: Think of how bizarre things are in the NFL now, when the most game-changing events in games often involve replay rather than the live action on the field.

joshua.hermsmeyer: The live action has its own set of flaws, though. I think there’s a fundamental principle of fairness that people also enjoy and want in a sporting event.

The refs, as we saw Thursday night, aren’t even all that great at spotting the damn ball.

Salfino: I guess it depends on whether you view refereeing a game as something that should be almost automated and perfect, or whether it’s just an organic part of the game with its variation in performance, just like the players. I totally get wanting 100 percent justice on the field, but it seems like it’s just never going to happen, for structural reasons. It’s not like the system can really be improved. It’s all one step forward and at least one step back.

joshua.hermsmeyer: The replay system and refs in general certainly aren’t helping Sean Payton and the Saints.

neil: That much is for sure, haha.

sara.ziegler: I don’t want to hear about it from the Saints, ever.

neil: To err is Favre-ian. Or referee-ian.

sara.ziegler: 🤣

Salfino: It’s incredible that the Saints again got screwed by a rule that made it impossible to fix the injustice we just witnessed on the field.

joshua.hermsmeyer: Al Riveron now posts an instant home office analysis on Twitter. They cut out the part where Cam Jordan runs it all the way back for a TD.


neil: Can’t they just, as a rule, err on the side of waiting to blow the whistle?

Let it play out, and if you have to bring it back, bring it back.

sara.ziegler: That was incredible. I thought that’s what they’re told to do?

Salfino: If you wait for the whistle, you’re going to have guys getting blasted on many plays that should be dead.

joshua.hermsmeyer: If it’s close, they are told to let it play out on the field, but in practice it seems that rarely happens.

sara.ziegler: So on replay, what are our options? What can — or should — the league do?

joshua.hermsmeyer: I think the smartest thing to do — given the mountain of evidence that what’s driving a lot of the issues is the faulty original ruling on the field, and the deference given to those calls by the rule book — is to not privilege any evidence before review.

Salfino: I really want to go back to watching live action. I’d just scrap the entire thing. It was done before. Replay just can never be perfect, and perfection is its reason for existing.

neil: One of things coaches and commentators always beg for is consistency. Especially with regard to the PI challenges, which coaches are still trying to figure out, make what is “clear and obvious” consistent from week to week. Maybe this is just based on a few plays, but on Sunday it seemed like it was easier to overturn a PI non-call than it had been in Week 1.

Salfino: Short of scrapping it, I’d make the booth responsible for all reviews. There should be no limit on things. No strategic component to it.

sara.ziegler: The limit is frustrating, for sure, given how arbitrary it all seems now.

I guess it comes down to what the goal is. Is it to get the call perfectly right, every time? If it is, then the booth needs to be a lot more involved.

I don’t think that is the goal, FWIW, but I’m not sure exactly what the goal is.

neil: The goal is to avoid media and fan criticism. And it always fails.

sara.ziegler: LOL

joshua.hermsmeyer: Yeah by that measure, just delete the account.

Salfino: I thought overturning the Stefon Diggs touchdown was absurd. And there was an outbreak of questionable calls on Sunday besides that, most for very ticky-tack reasons.

sara.ziegler: Don’t even get me started on the Diggs TD.

Salfino: By the way, what the heck is wrong with Kirk Cousins? That game-losing interception was such a horrible decision. It was first down. Cousins is conservative when the situation calls to be desperate and desperate when conservatism is warranted. Sorry, Sara.

Also, Diggs doesn’t take off his helmet on what should have been the second TD if he doesn’t get robbed of the first TD, I would bet. That’s bad by Diggs, but it sure didn’t feel like justice was carried out on replay.

sara.ziegler: Yeah, that was just a whole mess.

neil: The Vikings really just did Vikings things to lose that one.

sara.ziegler: All right…..

neil: (Sorry.)

sara.ziegler: LOL

joshua.hermsmeyer: The Vikings running back is leading the league in rushing yards, so clearly these kinds of losses are just variance.

Run to win!

sara.ziegler: Dalvin Cook is a GENERATIONAL TALENT, Josh.

joshua.hermsmeyer: Glorious.

neil: 75-yard runs will really pad the totals.

sara.ziegler: ANYWAY.

Let’s move on to the other big issue of the weekend: injuries. We had two big injuries to older star quarterbacks — Drew Brees and Ben Roethlisberger. Roethlisberger is out for the season with an elbow injury; Brees will have thumb surgery and miss at least six weeks.

Salfino: Yeah, there was video on the sideline of Brees trying to pick up the ball. Imagine trying to do that without a thumb, and that’s how that went.

sara.ziegler: Ooof

joshua.hermsmeyer: There was something deeply sad about that footage. He tried to play it all off, and just walked away with his head down looking at his hand.

neil: One of the cool new features in our quarterback-adjusted NFL Elo prediction model is that we can quantify the effect of losing a star QB. And these are very damaging injuries. Roethlisberger and Brees are currently the fifth- and sixth-best starting QBs in our model, trailing only Patrick Mahomes, Dak Prescott, Tom Brady and Matt Ryan.

Our model instantly had the Steelers’ playoff odds dropping from 31 percent to 8 percent with the news about Roethlisberger:

The effect of Brees’ injury was a little bit less, because he could potentially come back for most of the second half of the season. New Orleans’s playoff odds dropped from 58 percent to about 51 percent, with Tampa Bay being the primary beneficiary of that dip. But in both cases, we’re talking about drop-offs that cause these teams’ chances of winning their next game to fall by 10 to 15 percentage points with the backup having to start.

sara.ziegler: Are injuries what will finally do these guys in?

neil: Injuries do seem to be what spells the end for old, productive QBs.

Salfino: It seems like Roethlisberger’s injury was likely wear and tear, while Brees’s was a fluke and not age-related. But playing through the injury, if that’s what Brees opts to do, is likely going to be much tougher at his age than it would have been at his peak.

joshua.hermsmeyer: Brees looked like he lost his fastball — or what was left of it — last year. There was a phantom/unreported injury in November where he took a big hit and stopped throwing deep. He didn’t appear to have much zip on it this year either. He attempted just one pass over 20 yards in Week 1. So I guess my money would be on Brees being the most impacted by this accumulation of injuries. He’s also older than Roethlisberger.

neil: People like our friend Bill Barnwell were already speculating before the season about Brees’s performance potentially collapsing this season after his struggles late last year.

Salfino: Do you go to the unknown with Taysom Hill or the known with Teddy Bridgewater? I would do the former: Try to inject athleticism and explosiveness into the position and worry less about floor.


The drop off from Brees to Bridgewater was steep.

I think trying Hill — if Payton really believes he has some “Steve Young” in him — is the right move.

neil: Here’s a little bit more on the drop-offs between Brees/Roethlisberger and their backups. According to our model, replacing Brees with Bridgewater (who is roughly as good as Indy’s Jacoby Brissett, so the 29th-best starting QB in football) knocks the Saints’ Elo rating down from seventh in the league to 20th.

It gets worse for Pittsburgh. Replacing Roethlisberger with Mason Rudolph, who rates significantly lower than Ryan Fitzpatrick (the 32nd-best starter in the league), drops the Steelers from 17th in Elo to 31st, ahead of only the Dolphins.

joshua.hermsmeyer: Wow.

Salfino: I actually thought Rudolph looked good against Seattle. But it’s always tricky to judge the backups in the game when they are called upon because they play, I think, more free and easy since they had no idea they would be playing. The defense also doesn’t really know who they are. The test for Rudolph will be next week. He has draft pedigree to some extent.

sara.ziegler: Once again, BAN INJURIES.

Let’s wrap things up with a new game I’m calling Good Team/Bad Team. There’s a bunch of teams at 2-0 right now and a bunch at 0-2. But are they actually good/bad?

Starting with the 2-0 teams, I’ll name a team, and you tell me if that team is actually any good. Ready?

Salfino: I like this.

neil: Let’s go!

sara.ziegler: Let’s start with Buffalo! Did you know the Bills are 2-0?

joshua.hermsmeyer: The Bills are an eight-win team until Josh Allen gets injured on a scramble.

Salfino: The Bills are a bad team. They got lucky in Week 1: They played a QB with mono and then had the QB of the opposing defense (C.J. Mosley) go down, before which they hadn’t scored.

neil: Yeah, I agree with y’all. They may have swept the State of New Jersey in Weeks 1-2, but I’m not sure they’re actually any good.

Salfino: (OK, the Bills are not good but maybe not bad either; also, are the Giants tanking?)

sara.ziegler: The eternal question: Are the Giants tanking?

neil: Giants: Bad team.

joshua.hermsmeyer: Are people actually asking this question?

Salfino: Eli is 2-16 now in the first half of seasons from 2017 to 2019.

joshua.hermsmeyer: GM Dave Gettleman cannot abide a tank. They are just bad.

sara.ziegler: I think we know the Giants are bad, so we don’t really even need to discuss them.

What about San Francisco?

Salfino: San Francisco is good. I like their offensive coaching and play calling. They have very good offensive players — Matt Breida and Raheem Mostert are super-talented backs, especially Breida, who has some Barry Sanders in him. They look like they got it right with Deebo Samuel. The defense looks OK. I say 10-6.

neil: Good team…? The Niners currently rank third in EPA per game — granted, they beat a couple of mediocre teams (Tampa and Cincy) to get there. But Cincy held its own in Week 1 at Seattle! And Tampa won in Week 2 against Cam Newton and the Panthers!

joshua.hermsmeyer: The Niners were impressive in Week 2, but I’m not sold. It seemed like all the scheme stuff worked to perfection, and they didn’t need to really lean on Jimmy Garoppolo. In the first half of Week 1, they looked like the Niners we’ve seen the past two years, which is closer to my prior. So I’m gonna be the pessimist and say third in EPA/game is a bit of a mirage and that they aren’t a playoff team.

neil: San Francisco’s situation would also be rosier in a different division. The Rams and Seahawks are tough competition — our model thinks all three teams win double-digit games.

sara.ziegler: One more 2-0 team: Dallas!

joshua.hermsmeyer: Good.

neil: Good team! I wrote last week about how much it would help Dallas if Prescott became more consistent as a passer, and after two straight strong performances to start 2019 (on the heels of even more to close out 2018), it appears he may have turned a corner.

After that first interception, it seemed like “here we go again…” But he has been great since then.

Salfino: They’re good. For all the joking at Jerry Jones’s expense about how he takes over the team and doesn’t let the football people rule, they draft great every year. However they are doing it is working. Where is the weakness on Dallas? I don’t see one.

neil: With Philly being kinda all over the place so far, Dallas seems like clear NFC East favorites right now.

(Philly has actually been all over the place for like two years now…)

joshua.hermsmeyer: I think perhaps the biggest danger is that Kellen Moore runs out of plays three-fourths of the way through the year, and the league catches up. He’s never done a full season of this, and it seems to have happened to Sean McVay last year.

Salfino: Philly had such a spell of injuries on Sunday and still could have won if Nelson Agholor hadn’t dropped that touchdown pass down the sideline in the final two minutes. Carson Wentz often lacks pocket awareness and takes shots like Michael Vick used to take. If they can stay relatively healthy, I think the Eagles are at the Cowboys’ level.

sara.ziegler: OK, let’s move on to a few 0-2 teams to ask whether they’re actually as bad as they seem.


Salfino: Bad.

joshua.hermsmeyer: Not as bad as they seem. They’ve been surprisingly good in the passing game, and that will earn them a few unexpected wins down the road.

neil: Not sure. Case Keenum has actually been OK-to-good so far. And their losses were against two teams we just said we considered good: Dallas and Philly.

(Anybody else VERY confused to see Keenum as No. 8 for Washington, and think it’s Cousins?)

joshua.hermsmeyer: Hah, yes!

Salfino: I agree that Washington has been decent on offense, but Keenum is going to be benched when the losses pile up. Dwayne Haskins reportedly is not remotely ready, but the fans and owner will demand him. So I guess I’m building seven-plus Haskins starts into the “bad” call.

Keenum is better than Cousins. (ducks)

neil: LOL

sara.ziegler: Hahaha

joshua.hermsmeyer: What a lukewarm take after Cousins’s performance this week, Mike. Shame.

neil: Our QB model agrees! (Barely.)

Salfino: Cousins is actually losing games now.

sara.ziegler: OMG, we can’t talk about the Vikings anymore, I’m sorry — I just can’t take it.

How about 0-2 Carolina?

joshua.hermsmeyer: What’s wrong with Cam Newton?

neil: I wish I could defend him and them. But he’s getting outplayed by Jameis Winston at this point. At home!

Salfino: I worry that Cam is broken down and forever Clark Kent now. He’s taken such a beating. We love the running QBs, but there is a price to be paid, and Cam may not be functional in his 30s because he can’t execute at an NFL level in structure in the pocket — he’s not a good enough passer.

joshua.hermsmeyer: If you’re right, then 0-2 is very, very real.

sara.ziegler: LOL

One more 0-2 team: Pittsburgh

neil: With Big Ben out, they’re in huge trouble. I mean, they were probably already in huge trouble anyway.

Salfino: It was ridiculous to think an offense could withstand the losses of Le’Veon Bell and Antonio Brown in consecutive years. JuJu Smith-Schuster is not a true No. 1 and is being thrust into that role without even a competent No. 2 receiver.

Pittsburgh is bad.

joshua.hermsmeyer: Agree with all of you. JuJu has been a hobby horse of mine as well. His efficiency last year was buoyed by the presence of AB and being asked to run routes near the seams and in the middle of the field — the best places to pass the ball. Without their starting QB, they are bad.

Salfino: And the thing about Ben aging — you don’t get the sense he’s focusing on diet and yoga. There is no BR7 program, I suspect.

neil: 🤣

Check out our latest NFL predictions.

Sex, Speed And The SEO Migration of Romantic Depot to WordPress

Romantic Depot operates six, soon to be seven adult stores offering sex toys and lingerie, in the New York City area.  Their flagship stores are in Manhattan and the Bronx.  They’ve been around for sometime and over the years their website aged along with other businesses seeking to help drive local foot traffic.

With the move to mobile devices in full demonstration the old site was not responsive.  This lead the owner of the chain to build a new site that was mobile friendly and it lead to Ultimate SEO‘s involvement overseeing the process of migrating to this new site without hurting the site’s strong local SEO presence.

Romantic Depot does have an impressive keyword positioning presence in the New York City area. Even nationally they are on page 2 of results for “sex shop“.  The goal was to ensure a smooth transition to the mobile site while maintaining the SEO that had been built over the years.

301 Redirects

The new site was largely a 1 to 1 ratio.  The html static page manhattan.html went now to /manhattan/ on the WordPress site.  We placed in any one off redirects, a redirect that took any url that ended in .html and would return it without the html and a redirect for the index.html homepage to come back with the WordPress homepage at /


Backlinks are the life blood of a sit’s ranking and it was important to ensure that those would be maintained with relevant content as well.  Using we collected all of the backlinks and their existing targets and ensured those had rules as well.  While the site’s backlinks were in the tens of thousands it quick came down to a few hundred target urls that needed to be accounted for to maintain SEO.

Site Speed

Most of the work involved in preparing for the migration was speed performance in nature.  The new site when tested on was loading in 12.6 seconds with over 400 http requests. We targeted a 3 second load and through we were able to utilize a CDN that brought the site closer to users as well as offered other benefits.  Cloudflare alone brough the site load time to about 7 seconds.

We further limited content that could be on other pages for those other pages such as Google maps to the location homepages. Instituting lazy load ended up being the primary aspect of speeding up the site.  Image optimization was also completed and a move to PHP7.3 from 5.6.  Merging CSS and JS files also worked to reduce the requests.

SIte Speed FOr SEO MIgration

SIte Speed FOr SEO MIgration

With our work to provide a faster site complete the migration was completed and load times on the site are under 3 seconds for mobile users.

Multi Domain Strategy Consolidation

During the migration I also mentioned the value of building one brand. was the site used as the online store for the chain.

The problem that arises from multiple domain strategies is the segmentation of resources and confusion it can cause to Google Analytics.  An easy eample of this is the bounce rate and pageviews metrics are actually hurt on the primary domain.

Consider this… a person searches romantic depot on Google.  The first result is their site, likely the person is going to want to know what items might be at the store.  Once the page loads they find the link to the store, maybe even before the page loads.  Clicking that link they are now taken to a new domain.

That visitors actions would have counted as a bounced visitor.  See when someone goes to your site and immediately leaves for another site that signals to Google that what was on that original site wasn’t what the searcher wanted.  To prevent future searchers from going to a site that people leave directly after going to it they might increase the position of other sites to try and correct for this in the future.  That ultimately means the top spot position for the keyword is being hampered by the site’s structure.

Further Google sees that a person wasn’t even interested enough in the site to look at a second page, they just left.  In realty the second site is part of the same overall topic or brand its just that Google doesn’t necessarily understand that.  An artificially inflated bounce rate and lower page views are all that the first site is getting and the second site is losing out as well as most of the marketing is surrounding the first site’s address…backlinks, social mentions and such.

Lower Bounce Rate

The illustration above shows our page views of the main root domain.  Guess when on the graph the site was rolled into the main domains … late July.  The thing is, the traffic isn’t any greater its just not split up anymore.  The homepage link to the store is now going to a subfolder of the same domain, its helping by acting as another page view rather than hurting the site as a bounce.

The keywords and authority of this additional site were better utilized under the main domain and this site was migrated to a subfolder /store as a separate WordPress site.

Thats important the site was migrated as a separate site under the original.  This was done for multiple reason and it creates its own set of unique challenges but we’ll discuss that later in a future post.

Consolidated Backlinks

The consolidation of the sites further helps with SEO because after we migrated we put into place redirects from the store’s domain to its new home within the subfolder.  That means all the backlinks now combine to help one site.  Lets consider the following illustration…

Domain A: DA 30 Backlinks: 10,000 Referring Domains: 1,000

Domain B: DA 30 Backlinks: 9,000 Referring Domains: 900

Competitor: DA 35 Backlinks: 13,000 Referring Domains: 1,300

Lets assume everything else is the same…we’d expect then that Competitor will rank higher on Google Search.  But if we combine Domain A ad Domain B.

Domain AB: DA 40 Backlinks: 19,000 Referring Domains: 1,900

Everything else still the same….Domain AB will now rank higher than the Competitor.

Hits: 0

The Third Democratic Debate In 7 Charts

For the first time this cycle, there was just one debate night, and only 10 candidates made the cut — so now we’re trying to make sense of what happened when the front-runners shared the stage. In recent weeks, the polls have shown a top tier of three to five candidates, with former Vice President Joe Biden leading, Sens. Bernie Sanders and Elizabeth Warren tied for second, and Sen. Kamala Harris and Mayor Pete Buttigieg of South Bend, Indiana, a distant fourth and fifth — but did that change last night?

Some candidates from the lower-polling tiers had strong performances — former Rep. Beto O’Rourke delivered an impassioned speech on gun violence and Sen. Cory Booker spoke nearly as much as Biden, though Booker is only polling at 2.1 percent on average (based on 21 debate-qualifying polls). But as you can see from the FiveThirtyEight/Ipsos poll conducted using Ipsos’s KnowledgePanel, the overall picture hasn’t shifted much yet — although Warren does seem to have done the most to boost her campaign. Here’s what we’ve learned so far about what viewers made of the debate and the candidates’ performances:

Which candidates wowed the crowd?

First of all, how did viewers in our poll think the candidates did on Thursday night? To answer this, we compared debate-watchers’ ratings of the candidates’ performances to their pre-debate favorability scores1 to see if any well-liked candidates failed to impress or if anyone got high marks despite lower favorability. By this measure, O’Rourke and Warren were the biggest standouts, though Buttigieg and Booker also made a positive impression. But Biden and former Cabinet secretary Julián Castro — who memorably clashed — got the lowest scores relative to their pre-debate favorability.


Who gained (or lost) potential supporters?

Another way to assess who won last night’s debate is to see who convinced more voters to at least think about voting for them. Most candidates saw some change in the share of likely Democratic primary voters who were considering supporting them, though not all changes were positive. Warren, for example, saw the biggest increase in voters who were considering her — almost 4 percentage points, while Harris lost more than 2 percentage points of potential support. But for most candidates, the numbers stayed pretty much the same as they had been before the debate. Even for those whose debate performance stood out — like Biden and Castro, who got relatively poor grades, or O’Rourke, who got a strong rating — there was little change in how many likely primary voters said they were considering voting for them.

Which candidates appeal to the same voters?

With many voters in our poll still considering multiple candidates, we were also interested in examining which candidates share potential supporters. Perhaps unsurprisingly, the two candidates who are being considered by the most voters in our poll — Biden and Warren — also tended to draw a high proportion of other candidates’ supporters, too. Seventy percent of Buttigieg’s supporters are also considering Warren, for example, while 65 percent of O’Rourke’s supporters are also considering Biden. Although many respondents in our survey said they were considering Sanders, fewer of his supporters are considering supporting other candidates. In fact, Biden and Sanders had the most exclusive supporters — 24 percent of Biden’s supporters and 18 percent of Sanders’s supporters aren’t considering any of the other candidates who participated in the debate.

Who made a positive (or negative) impression?

You can also look at the change in candidates’ favorable and unfavorable ratings to understand who got people feeling more positively about them (or perhaps gained unwanted notoriety). So after Ipsos polled voters before and after the debate, we calculated the change in candidates’ net favorability (favorable rating minus unfavorable rating). O’Rourke may not have picked up many potential supporters, but he did improve his net favorability rating by more than 8 points with his debate performance. Castro, meanwhile, took the largest hit, dropping 6.7 points in net favorability, which could be related to his heated exchanges with Biden.

More people like O’Rourke, but Castro lost ground

Change in net favorability for candidates in a FiveThirtyEight/Ipsos poll taken before and after the third Democratic primary debate

Net favorability
candidate before debate after debate change
O’Rourke +23.9 +32.5 +8.6
Warren +48.5 +56.0 +7.5
Klobuchar +8.1 +13.9 +5.8
Buttigieg +32.2 +37.7 +5.4
Booker +26.7 +32.1 +5.4
Harris +31.3 +35.2 +3.9
Yang +14.8 +17.4 +2.6
Biden +45.7 +45.5 -0.2
Sanders +44.0 +43.7 -0.3
Castro +19.7 +13.0 -6.7

From a survey of 4,320 likely Democratic primary voters who were surveyed between Sept. 5 and Sept. 11. The same people were surveyed again from Sept. 12 to Sept 13; 2202 responded to the second wave.

Who spoke the most?

Though respondents to the FiveThirtyEight/Ipsos poll thought Biden’s debate performance was less impressive than Warren’s, it wasn’t because he didn’t get a chance to talk. Of all the candidates on the stage last night, Biden had the highest word count, with over 3,000 words spoken. Booker and Warren, the next two most prolific speakers, were about 600 and 750 words behind, respectively.

Who held the floor?

Number of words candidates spoke in the third Democratic debate

Candidate Words Spoken
Joe Biden 3,363
Cory Booker 2,769
Elizabeth Warren 2,616
Kamala Harris 2,369
Julián Castro 2,104
Pete Buttigieg 2,054
Amy Klobuchar 1,933
Bernie Sanders 1,891
Beto O’Rourke 1,714
Andrew Yang 1,546

Excludes words spoken in Spanish

Source: Debate Transcript via ABC News

Booker’s place as the second-most-prolific talker is even more impressive considering that he’s polling in the low single digits. (The polling average is based on 21 debate-qualifying polls released between June 28 and Aug. 28.) Castro also spoke more than anticipated given his polling average (1 percent), holding the floor for longer than both Buttigieg and Sanders. Sanders, in fact, had the second-highest polling average going into the debate, but was third from the bottom in words spoken, beating out only O’Rourke and businessman Andrew Yang.

Who mentioned Trump?

In addition to counting the words spoken by candidates, we also tracked the number of times each candidate mentioned President Trump by name:

Who talked about Trump?

How often Trump was mentioned by candidates participating in the third Democratic debate

Candidate Trump Mentions
Kamala Harris 11
Julián Castro 7
Cory Booker 5
Bernie Sanders 5
Beto O’Rourke 2
Andrew Yang 2
Joe Biden 1
Pete Buttigieg 1
Elizabeth Warren 1
Amy Klobuchar 0

Source: Debate Transcript via ABC News

Harris was the clear leader, mentioning Trump 11 times, though as a group, the candidates talked about Trump considerably less often than they did in either night of the second debate. And some of the candidates who spoke the most, such as Biden and Warren, seemed to avoid Trump, each mentioning the president only once. Sen. Amy Klobuchar, meanwhile, got through the whole debate without saying Trump’s name even once.

So did the single-night debate change the game? Thursday night’s debate drew about 14 million television viewers, which is more viewers than both nights of the second debate, but still slightly fewer than those who tuned into watch the first debate. And if our poll with Ipsos is indicative of voters’ reactions, then the needle didn’t move all that much. But for those of you who preferred the two-night approach, you might be in luck — the fourth debate, set for Oct. 15 and potentially Oct. 16, might be split across two nights, since at least 11 candidates have qualified so far. (The Democratic National Committee hasn’t yet confirmed what it will do, however.) Either way, we’ll be here live blogging, so stay tuned!

Do you want even more debate coverage?

Cool graphics from other sites

And here’s more great post-debate analysis:

Finally, check out the rest of our debate coverage:

Additional contributions from Aaron Bycoffe.

Charles Leclerc Has Gone From Ferrari’s Future To Its Present

In recent years, Formula One’s defining rivalry has seemed clear: Lewis Hamilton of Mercedes (and before that, McLaren) vs. Sebastian Vettel of Ferrari (and before that, Red Bull). The two have combined to win nine of the past 11 world driver’s championships, with the only interlopers being Jenson Button in 2009 and Nico Rosberg, another of Hamilton’s most bitter rivals, in 2016. Hamilton and Vettel finished 1-2 in each of the previous two seasons’ standings, so the path to the F1 title seemed very likely to go through them in 2019 as well.

This year, though, only one of the two has held up his end of the bargain. Hamilton currently leads the championship, 63 points clear of Mercedes teammate Valtteri Bottas. But instead of vying for the title as well, Vettel is all the way back in fifth place, 115 points behind his English archrival. The German hasn’t finished so low in the standings since 2014, his tumultuous final season with Red Bull.

With the recent breakout of Vettel’s Ferrari teammate, Charles Leclerc — the 21-year-old rising star who just grabbed his first two F1 career wins in back-to-back races — there are legitimate questions as to which driver Hamilton should be more concerned about as a rival going forward. In fact, Leclerc has done so well in his second F1 season that the four-time world champ Vettel might not even be the No. 1 driver on his own team anymore.

That this is up for debate speaks volumes about Leclerc’s meteoric ascent. The Monégasque phenom was a teenager in Formula Two just a couple of years ago, impressing as a test driver but failing to nail down a seat at a prestige team for his rookie F1 season. Leclerc joined Alfa Romeo-Sauber for 2018 instead and flashed his potential with a very strong qualifying performance against teammate Marcus Ericsson, whom he started ahead of on the grid 17 times in 21 races (81 percent) last season. But on race days, Leclerc finished slightly worse than he started,1 beat Ericsson only 57 percent of the time (after adjusting for grid position)2 and notched only 39 points in the championship, finishing a distant 13th.

(Besides, Ericsson wasn’t exactly the toughest competition; he didn’t secure an F1 ride in 2019 and currently competes in the IndyCar series.)

Still, Ferrari saw the talent evident in Leclerc’s performance and pegged him to replace folk hero Kimi Raikkonen as Vettel’s No. 2 going into 2019. But Vettel’s place atop the pecking order seemed secure. He had spent most of the previous four seasons driving circles around Raikkonen, beating the 2007 world champion in 67 percent of qualifying runs and 73 percent of races. Although he couldn’t quite outduel Hamilton, finishing an average of 101 points behind him in the overall standings during those seasons, Vettel was still very competitive — and hopeful that a new 2019 design package would bring the Prancing Horse its first championship (as either a constructor or for its drivers) since the late 2000s.

The early returns seemed like business as usual. Although Mercedes opened the season with an eight-race winning streak, as Hamilton took six of those checkered flags himself, Vettel had also outdriven Leclerc 12 times in 14 chances (including both qualifying and races) over the first seven events on the schedule. (The only exception was the Bahrain Grand Prix, where Leclerc won the pole and finished third to Vettel’s fifth.) Vettel even technically crossed the finish line first in Canada, only to see Hamilton be awarded the win because of a controversial penalty assessed when Vettel swerved back onto the track after a missed turn.

Ever since that moment, however, Leclerc has surged past his older, more decorated teammate at a breathtaking pace. If we build an Elo rating using the same head-to-head approach that I used in this story about Fernando Alonso from last summer — which just compared teammate performances with each other (to control for differences in constructor quality) and gave qualifiers half-weight as compared with races — Leclerc moved past Vettel for the very first time in his career when he won the Italian Grand Prix last Sunday:

This teammate-versus-teammate approach isn’t perfect, particularly when it involves drivers who haven’t changed teams much. But it is able to infer one driver’s performance from how often he beats another driver of an established ability level. And Vettel is certainly established; he went into the season (and his partnership with Leclerc) with the third-best rating of any driver in the 2019 field, trailing only Hamilton and Red Bull’s Max Verstappen — another Leclerc-esque wunderkind (also 21 years old) trying to end the Hamilton/Vettel hegemony atop the F1 standings.

For Leclerc’s part, he has risen from a relatively mediocre head-to-head Elo of 1454 after the Canadian Grand Prix on June 9 to a 1529 mark (fourth-best in the field) after back-to-back victories in Belgium and Italy these past two weeks. He has now outperformed Vettel in each of the past seven qualifying sessions and five of the past seven races (the only exceptions being Germany and Hungary). By coolly fending off the best attacks Hamilton and Bottas could throw at him, the unflappable Leclerc gave Ferrari a win at Monza — its home race — for the first time since Alonso did it in 2010. Along the way, he has given Vettel more fits than just about any teammate in his entire career:

Vettel’s teammates seldom challenge as much as Leclerc

Head-to-head comparisons between Sebastian Vettel and teammates — in qualifying and races — by season, 2007-19

Qualifying Races*
Season Team Teammate H2H Wins Win% H2H Wins Win% Champ. Rk
2007 Sauber/Toro Rosso Heidfeld/Luizzi 3 38% 3 38% 14th
2008 Toro Rosso S. Bourdais 14 78 13 72 8th
2009 Red Bull Racing Mark Webber 14 82 8 47 2nd
2010 Red Bull Racing Mark Webber 13 68 11 58 1st
2011 Red Bull Racing Mark Webber 16 84 14 74 1st
2012 Red Bull Racing Mark Webber 12 60 14 70 1st
2013 Red Bull Racing Mark Webber 17 89 17 89 1st
2014 Red Bull Racing Daniel Ricciardo 9 47 6 32 5th
2015 Scuderia Ferrari Kimi Raikkonen 15 79 16 84 3rd
2016 Scuderia Ferrari Kimi Raikkonen 9 43 15 71 4th
2017 Scuderia Ferrari Kimi Raikkonen 15 75 16 80 2nd
2018 Scuderia Ferrari Kimi Raikkonen 15 71 12 57 2nd
2019 Scuderia Ferrari Charles Leclerc 6 43 8 57 5th
Total 158 68 153 65

*Race performance is adjusted slightly for starting grid position


Despite the back-to-back losses, Mercedes isn’t exactly worried about Leclerc chasing Hamilton down for the title. The quick circuits of the past two races probably favored Ferrari, with the superior straight-line speed of its SF90 car. And even after Leclerc’s big breakthrough, he remains 102 points behind Hamilton (and 39 behind Bottas) in the standings, with Ferrari running 154 behind Mercedes in the team tally as they look ahead to the season’s final seven races.

But although Leclerc’s wins didn’t make much impact on the overall championship picture, they may have represented something of a turning point in F1 history. Going into 2019, Vettel had outqualified his teammates — a strong group that included Raikkonen, Daniel Ricciardo and Mark Webber — 69 percent of the time and outraced them 66 percent of the time in his career. Against Leclerc this season, those numbers are only 43 percent and 57 percent, respectively, and getting worse by the moment.

Vettel entered the season as a championship contender and the clear standard-bearer for the sport’s most storied team. Now his place is in doubt. With another year left on his contract, Vettel probably isn’t going anywhere (despite tabloid rumors to the contrary), but it’s hard not to juxtapose Leclerc’s recent surge against Vettel’s growing tendency toward crucial errors and on-track decisions that put other drivers at risk. So the biggest question for the season’s home stretch isn’t whether Leclerc is the future of the sport — that now seems established — but rather, it’s whether Vettel can avoid being left in the past.

Zack Greinke’s 8 Pitch Types Keep ‘Em Guessing

Just over a month ago, the Houston Astros pulled off the biggest move of the season: In a deal reported minutes after the trade deadline had passed, the Astros acquired Cy Young winner Zack Greinke from the Diamondbacks to form baseball’s best rotation alongside Justin Verlander and Gerrit Cole, the two likeliest Cy Young candidates.

Greinke, a future Hall of Famer, has been stellar this season — a 2.99 ERA and a sub-1 WHIP. According to FiveThirtyEight’s pitcher ratings, he would be the top pitcher on two-thirds of teams and the No. 3 on just three — and that’s the one he’s on. The move already has paid dividends for the ‘Stros, who are 5-2 in Greinke’s starts and have baseball’s best run differential since the trade — by nearly 50 runs.

Now six weeks into team No. 6 (remember when he was traded to the Angels for the 2012 stretch run?), Greinke continues to adapt. At 35 years old, it’s anyone’s guess how long he can keep up this performance, but he’s signed through 2021 and should contribute through then. And because of the way he’s dealt with his decreasing velocity by relying on command and movement, he should be set up well for continued long-term success.

According to Statcast, Greinke throws eight pitches: four-seam fastball, changeup, slider, curve, sinker (or two-seam fastball), split finger, cutter and eephus. (We’ll get back to that last one.) Only Yu Darvish has as many listed on his Statcast page, with the same eight (though the classifications may hide some of Darvish’s arsenal). Anibal Sanchez, Rich Hill and Odrisamer Despaigne are the only pitchers with seven.

But it’s not just the variety of pitches that makes Greinke special. It’s how he throws them.

Consider his changeup. Greinke throws his offspeed on 21.9 percent of pitches — a fairly steady increase from 7.9 percent back in 2008. Yet as his fastball has lost velocity, from once hitting more than 100 miles per hour in 2010 to averaging below 90 in 2019, his changeup has gotten faster.

Here, we point out that the goal of a changeup is usually to fool the batter by presenting a speed different from that of the fastball. Greinke does not do that.

Instead, Greinke uses a power changeup with devastating late movement. Only Edubray Ramos has a smaller average speed difference. Greinke’s pitch has surpassed his slider, which used to be considered his best pitch, as the second option. Along with this, Greinke’s cutter, a staple of his arsenal in 2012 and 2013, has all but disappeared.

Then there’s the curveball, a slow sweeping pitch. Greinke’s curveball is the second-slowest among qualified starters, behind the Nationals’ Patrick Corbin, at just over 70 miles per hour.

This is where the eephus comes in. Greinke’s curve can be thrown so slow that Statcast registers it as the arcing pitch. But it’s not clear whether it’s a different pitch or just a curveball thrown slower. Nobody is throwing a true eephus, though six pitchers are credited with the pitch this year; only Greinke has one under 60 miles per hour. But even if you consider his eephus and his curveball as the same pitch, Greinke would still be tied with Sanchez and Darvish for the lead with seven different pitches.

MLB pitchers have struck out 16 hitters on sub-67 mph pitches this year. Greinke owns eight of those (and four of the rest are from position players) with his slow curve that can make batters look silly.

The newest one is the split-finger, which he threw in April for the first time since pitch tracking began in 2008. He’s thrown five so far in 2019, including three to Jacob DeGrom in the same at bat. If he’s experimenting with it now, there’s a chance it becomes a regular part of his arsenal in the future, especially with the Astros’ penchant for getting the most out of pitch selection.

But beyond his wide repertoire of pitches, Greinke’s pitching style is one of a kind. He throws most pitches low but gets strikes. Even though he throws fewer pitches in the strike zone than average, he almost never falls behind. And his .198 wOBA allowed on pitches out of the zone is second in MLB, also behind Corbin.

Greinke has faced just 11 3-0 counts this year and had thrown a fastball every time, almost always on the edge, until he gave Christian Yelich a perfect changeup last week. None of the 20 other pitchers with as many pitches this year has seen fewer than 15 such counts. In the month of July, Greinke threw 479 pitches and none was in a 3-0 count. He threw eight pitches with a 2-0 count — seven were in the strike zone and the other was fouled off. He’ll throw in the strike zone when he falls behind; that just doesn’t happen very often. And even when he does, batters can’t take advantage — they’re just 2-16 on 2-0 counts this year despite seeing 65 percent of pitches in the strike zone.

When he’s ahead, it’s a different story. That’s when the sub-70 curveball becomes devastating. Ahead in the count, Greinke throws just 27 percent of pitches in the strike zone; the league average is 38 percent. And 76 percent of his strikeouts have been on pitches out of the zone, well higher than the league average of 56 percent. And his plan of attack is to go low. On 1-2 counts, specifically, Greinke throws in either the lower third or below the strike zone more often than any other pitcher.

Greinke is truly a unique pitcher. His fastball and offspeed have nearly the same velocity, but his curveball is one of the slowest. He throws outside of the strike zone but never falls behind, and batters can’t seem to figure out any of his pitches.

Through his impressive career, the one thing Greinke lacks is a ring. He has 11 postseason appearances, but his biggest impact was probably his lone start in the 2014 NLDS (in which he scored more runs than he allowed in seven innings). He makes the top 10 list of career games started without a World Series appearance. But if he earns a huge postseason moment, he could move from likely Hall of Famer to potentially first ballot. Perhaps he’ll have that chance in Houston this October.

Check out our latest MLB predictions.

Can You Help Dakota Jones Raid The Lost Arc?

Welcome to The Riddler. Every week, I offer up problems related to the things we hold dear around here: math, logic and probability. Two puzzles are presented each week: the Riddler Express for those of you who want something bite-size and the Riddler Classic for those of you in the slow-puzzle movement. Submit a correct answer for either,4 and you may get a shoutout in next week’s column. If you need a hint or have a favorite puzzle collecting dust in your attic, find me on Twitter.

Riddler Express

At long last, Dakota Jones is close to finding the Lost Arc, a geometric antiquity buried deep in the sands of Egypt. Along the way, she discovered what she described as a “highly symmetric crystal” that’s needed to precisely locate the Arc. Dakota measured the crystal using her laser scanner and relayed the results to you. But nefarious agents have gotten wind of her plans, and Dakota and the crystal are nowhere to be found.

Locating the Arc is now up to you. To do that, you must recreate the crystal using the data from Dakota’s laser scanner. The scanner takes a 3D object, and records 2D cross-sectional slices along the third dimension. Here’s the looping animation file the scanner produced for the crystal:

Cross-sectional areas of a mysterious three-dimensional figure. Starts as a triangle, morphs into a hexagon, and then turns back into an inverted triangle.

What sort of three-dimensional shape is the crystal? No pressure — Dakota Jones, nay, the entire world, is counting on you to locate the Lost Arc and ensure its place in a museum!

Submit your answer

Riddler Classic

Recent Riddlers have tackled Scrabble Superstrings and road trips through 48 states. For this week’s Riddler Classic, Max Maguire combines these two puzzles into one:

The challenge is to find the longest string of letters in which (1) every pair of consecutive letters is a two-letter state or territory abbreviation, and (2) no state abbreviation occurs more than once. For example, Guam, Utah and Texas can be combined into the valid four-letter string GUTX. Another valid string is ALAK (Alabama, Louisiana and Alaska), while ALAL (Alabama, Louisiana and Alabama) is invalid because it includes the same state, Alabama, twice.

For reference, the full list of abbreviations is available here, courtesy of the United States Postal Service.

Submit your answer

Solution to last week’s Riddler Express

Congratulations to 👏Dan Anderson 👏 of Encinitas, California, winner of last week’s Riddler Express.

Last week, you and your friend played a game of “Acchi, Muite, Hoi” (also known as the “lookaway challenge”). In the first round, you randomly pointed up, down, left or right, while your friend randomly looked in one of those four directions. If your friend looked in the same direction you pointed, you won. Otherwise, as long as no one wins, you kept switching off who pointed and who looked.

Your chances of winning were exactly 4/7. There are a couple ways to go about finding that number.

As solver Alison Lynch observed, you can calculate the probabilities of winning each round and add them all up. You could only win the game when you were the one doing the pointing, which happened in the first round, the third round, the fifth round and all the other odd-numbered rounds. Furthermore, within each round, the pointer had a 1/4 chance of winning, since only one out of the four directions — up, down, left and right — matched what the looker did. Your chances of winning in the first round were therefore 1/4. To win in the third round, you must have failed to win in the first round (which had a 3/4 chance of happening), your opponent must have failed to win in the second round (again, with probability 3/4), and then you won in the third round (with probability 1/4). Multiplying these together, you had a 3/4 × 3/4 × 1/4 = 9/64 chance of winning in the third round, while in the fifth round your chances were 3/4 × 3/4 × 3/4 × 3/4 × 1/4. Continuing with this logic, adding up the infinitely many probabilities from all the odd-numbered rounds gave you a geometric series whose sum was exactly 4/7.

The students of The Hewitt School in New York City, meanwhile, thought about it another way. They called your chances of winning b. What are your opponent’s chances of winning? Well, first your opponent needs the game to proceed to the second round, which happens 3/4 of the time (the other 1/4 of the time you win in the first round). At this point, their chances of winning must also be b — their situation is identical to yours back in the first round. In fact, any time one of you is the designated pointer, that person’s probability of winning is b.

Since one of you must be the winner, you and your opponent’s chances of winning have to add up to 1. That means b + 3/4b = 1, which means, sure enough, that b = 4/7, or about 57 percent.

Just to be super-duper sure, solver Angelos Tzelepis simulated 10 million games, finding that whoever goes first wins a shade over 57 percent of the time.

Why isn’t the answer just 50 percent? Because pointing in the first round and having that very first opportunity to win offers you a slight edge. So who says it’s not polite to point? (I do. Don’t point.)

Solution to last week’s Riddler Classic

Congratulations to 👏Matt Friedrichsen 👏 of Overland Park, Kansas, winner of last week’s Riddler Classic.

Last week, you competed at the U.S. Open, where you were slightly better than the competition: Your chances of winning any given point were exactly 55 percent. What were your chances of winning a three-set match and a five-set match? And what were your chances of winning the whole tournament?

As solver Dennis Okon figured out, one way to solve this problem is to work backwards. Once you find the probabilities of winning from deuce, advantage-in and advantage-out, you can find the chances of winning a single game from lower and lower scores. For example, suppose you’re up 30-0 in a game. 55 percent of the time you’ll win the point, and the score will become 40-0, while the other 45 percent of the time you’ll lose the point, and the score will become 30-15. Your probability of winning the game at 30-0 turns out to be a weighted average of those other two scores: 55 percent of the probability you’ll win at 40-0 plus 45 percent of the probability you’ll win at 30-15.

You can use this technique to work all the way back to a score of 0-0, which represents your chances of winning a game at the start. If you do the same analysis for a set, you will find the probability of winning the set from different game scores, as shown in the table below.

What are your chances of winning the set?

The probability you’ll win a set of tennis given the number of games you and your opponent have won so far, assuming you have a 55 percent chance of winning each point

Games your opponent has won
Games you’ve won 0 1 2 3 4 5 6 7
0 81.5% 70.7% 56.1% 38.5% 20.6% 6.5% 0.0%
1 88.1 79.6 66.8 49.4 29.1 10.5 0.0
2 93.2 87.3 77.3 61.7 40.4 16.8 0.0
3 96.8 93.4 86.7 74.6 54.7 27.0 0.0
4 98.9 97.4 94.0 86.6 71.5 43.4 0.0
5 99.8 99.4 98.4 95.7 88.5 69.6 40.8 0.0%
6 100.0 100.0 100.0 100.0 100.0 87.0 65.4 0.0
7 100.0 100.0

The percentage when the score is 0-0 is equivalent to your chances of winning any given set: 81.5 percent. Your chances of winning two out of three sets are then 91 percent, and they are 95.3 percent for winning three out of five sets. Raising these probabilities to the seventh power gives you your chances of winning the entire U.S. Open: 51.7 percent in the women’s tournament and 71.4 percent in the men’s. As you can see, playing more sets gives stronger players a better chance at winning. So let the women play five sets, I say.

It’s a little shocking just how high your chances of winning are when you start with what seemed like a slight advantage on every point. Scoring in tennis is highly nonlinear: A tiny advantage in winning a point snowballs into a big probability shift over the course of an entire match. Riddler Nation Hall-of-Famer Diarmuid Early elegantly captured this nonlinearity in a single graph, where you can see how quickly your chances of winning a set and a match change as your probability of winning each point crosses 50 percent:

Probability of winning a game, tiebreak, set, and match vs. probability of winning any given point. The relationship is highly nonlinear. As the probability of winning a point passes through 50%, the other probabilities jump up very quickly.

Sure enough, winning 55 percent of points does not make you “slightly better than the competition,” as the puzzle stated — instead, it puts you squarely in the pantheon of tennis’s all-time greats! Roger Federer, Rafael Nadal and Novak Djokovic have all won approximately 55 percent of their career points.

Finally, if you’d like to play around with these percentages for yourself, check out this tennis calculator by Mark Bennett. It will serve you well.

Want more riddles?

Well, aren’t you lucky? There’s a whole book full of the best puzzles from this column and some never-before-seen head-scratchers. It’s called “The Riddler,” and it’s in stores now!

Want to submit a riddle?

Email Zach Wissner-Gross at [email protected].

Is Dak Prescott Finally Ready To Be Consistently Great?

Baltimore Ravens quarterback Lamar Jackson stole the show on the NFL’s opening Sunday with a passer rating of 158.3, the maximum number possible. But Jackson wasn’t the only QB with a perfect 158.3 mark Sunday. He was joined by Dak Prescott of the Dallas Cowboys, who went 25 for 32 with 405 yards, four touchdowns and zero picks against the rival New York Giants. It was the first time ever that two passers had perfect games in the same week1 — and in fact, Prescott’s game may have been the superior perfect outing.

At least, that’s according to our new Elo QB ratings, which saw Prescott outperform both Jackson and Kansas City’s Patrick Mahomes for the best game of Week 1, after adjusting for the opposing defenses each faced. Our model considered it to be the best game of Prescott’s entire career, easily surpassing his effort against the Ravens in Week 11 of 2016 — and there’s nothing the Cowboys would welcome more than a return to form of that (mostly) storybook season, both as a team and for Prescott individually.

Dallas had high hopes Prescott would be a fixture atop these kinds of rankings ever since 2016, when he produced one of the greatest rookie QB campaigns in NFL history en route to a 13-3 record as the starter. Thanks to that immediate success, Prescott went into 2017 with the ninth-highest Elo rating of any starting QB, and he eventually rose to No. 5 in the league by Week 9, when the Cowboys had a 5-3 record and a legitimate chance of winning the Super Bowl.2 It was the best a Dallas quarterback had rated in Elo (relative to league average) since Tony Romo in the middle of the 2013 season.

From that moment onward, however, Prescott has been all over the place. Over his following 26 starts — taking him through the end of the 2018 season — Prescott’s average game-to-game performance was about 11 points of Elo value below that of an ordinary starter, while he registered exactly 13 above-average starts against 13 below-average ones.

But we’re just scratching the surface of how up-and-down Prescott has been.

Over an eight-start period from Dec. 24, 2017, to Oct. 14, 2018, Prescott alternated between an above- and below-average QB Elo performance every single game. Then he strung together two consecutive above-average starts… before embarking on a separate stretch of nine more starts in a row in which every single game alternated between an above- and below-average performance.

It was one of the most erratic runs in pro football history. Seahawks QB Russell Wilson once had an incredible 17 consecutive starts (!) waver between positive and negative during the 2014 and 2015 seasons,3 so Prescott wasn’t quite at that level of inconsistency. But he, Cam Newton, Matthew Stafford and Matt Schaub are the only QBs in our dataset (since 1950) to have two separate streaks in their careers with at least eight consecutive starts that seesawed between above- and below-average performance every game.

The kings of inconsistent quarterbacking

Quarterbacks with the longest streaks of their Elo values per game alternating between above average and below average, 1950-2019

Longest streaks of alternating above- and below-average performance
Seasons Team Quarterback Number of games
1 2014-15 SEA Russell Wilson 17
2 1994-95 ATL Jeff George 14
3 1973-74 CLE Mike Phipps 13
T-4 2003-04 ATL Michael Vick 11
T-4 2015-16 BUF Tyrod Taylor 11
T-15 2018 DAL Dak Prescott 9
T-26 2017-18 DAL Dak Prescott 8
Quarterbacks with at least two streaks of at least eight games of alternating performance
Season Team Quarterback Number of games
2012 CAR Cam Newton 10
2013 CAR Cam Newton 10
2018 DAL Dak Prescott 9
2017-18 DAL Dak Prescott 8
2009-10 HOU Matt Schaub 9
2012 HOU Matt Schaub 8
2016-17 DET Matthew Stafford 10
2014-15 DET Matthew Stafford 8


Prescott finally broke the cycle late last season, stringing together four consecutive starts with positive value to close the year (including the playoffs). That’s how Prescott already had entered this season with the ninth-best Elo rating of any starting quarterback in the league, a big improvement over his No. 22 ranking going into 2018. And after Sunday’s impressive outing, he now ranks sixth — behind only Mahomes, Ben Roethlisberger, Drew Brees, Tom Brady and Matt Ryan. According to our Elo values, Prescott’s performance against New York was the fourth-best game by a Dallas starting QB (relative to league average) in Cowboys history:

Dak’s place among the best Cowboy QB games ever

Best single-game performances for Dallas Cowboys starting quarterbacks according to QB Elo value relative to league average, 1960-2019

Quarterback Date Opponent Result QB Elo value vs. Avg
Don Meredith 10/9/1966 vs. PHI W, 56-7 481 +389
Troy Aikman 1/31/1993 vs. BUF* W, 52-17 418 +348
Don Meredith 11/13/1966 at WSH W, 31-30 439 +346
Dak Prescott 9/8/2019 vs. NYG W, 35-17 504 +340
Craig Morton 12/20/1970 vs. TEN W, 52-10 411 +331
Don Meredith 9/18/1966 vs. NYG W, 52-7 423 +330
Troy Aikman 10/27/1996 at MIA W, 29-10 421 +329
Don Meredith 11/10/1963 at SF L, 24-31 390 +304
Roger Staubach 12/12/1977 at SF W, 42-35 371 +299
Tony Romo 12/6/2009 at NYG L, 24-31 411 +296

* Super Bowl

Sources:, ESPN

The only question now is, can Prescott keep this up and put his old inconsistent ways behind him? Although our QB values are adjusted for the quality of opposing defenses, it’s still valid to wonder how much a stellar performance against the Giants — whose defense ranked last in the league according to preseason projections from ESPN’s Football Power Index — will translate against better opponents such as, say, the Vikings (who tied for ESPN’s No. 1 preseason defense) in Week 10.

But there are also reasons to think Prescott’s breakout might endure throughout the season. New offensive coordinator Kellen Moore already has added new wrinkles to the Cowboys’ scheme that were absent under Scott Linehan in previous years. Prescott also has a new(ish) set of primary targets, with Amari Cooper present from the start of the season — after averaging 80.6 yards per game upon his midseason arrival in Dallas last year — plus Michael Gallup graduating to No. 2-target status after Cole Beasley’s departure, and former Pro Bowl WR Randall Cobb coming over from the Packers. (Future Hall of Fame tight end Jason Witten is also back after a much-maligned year in the broadcast booth.)

After throwing for a below-average 7.69 air yards per attempt in 2017 and 2018, Prescott was up to 9.34 air yards per throw against the Giants, with 19 percent of his passes traveling at least 20 yards — including completions of 35 and 21 yards downfield to Cooper, 30 yards to Gallup, 22 yards to Cobb and 22 yards to tight end Blake Jarwin.

Again, those were against the Giants, so it’s important not to draw too many conclusions from Prescott’s Week 1 numbers, impressive as they were. But for a team that ranked ninth in defensive efficiency (via Football Outsiders) but only 26th in passing efficiency last season, an improvement from Prescott could vault the Cowboys to the top of the NFC East — and maybe beyond.

Looking Ahead: Week 2

Best matchup: No. 3 New Orleans at No. 5 L.A. Rams (-1.5)

Matchup quality: 97th percentile4

Matchup evenness: 76th percentile

The Rams and Saints will meet on Sunday afternoon in a rematch of that infamous NFC Championship game. Both teams won in Week 1 by slim margins, and Elo has the Saints ranked third in the NFL while the Rams are ranked fifth. A big part of that difference comes down to the quarterbacks: New Orleans’ Drew Brees ranks third in our ratings, but L.A.’s Jared Goff ranks only 26th after another subpar game (on the heels of a terrible Super Bowl and a string of mediocre outings late last season). Goff has a lot to prove, but he’ll also have a big opportunity against a Saints defense that FPI ranks just 28th in the league.

See also: Philadelphia at Atlanta (80th/81st); Minnesota at Green Bay (77th/75th).

Biggest playoff implications: No. 9 Minnesota at No. 12 Green Bay (-1.5)

Potential shift in playoff odds: 30.7 total percentage points

In terms of playoff odds, the biggest game of Week 2 squares the Vikings off against the Packers. The teams have essentially identical chances to make the postseason (52 percent and 51 percent, respectively), and the winner would be set up well in the NFC North race. If Minnesota wins, their playoff odds go up to 69 percent; if Green Bay wins, their number would be 65 percent. In either case, the loser’s playoff percentage drops into the mid-30s.

See also: Indianapolis at Tennessee (25.0); Philadelphia at Atlanta (22.3).

Best QB duel: No. 5 Matt Ryan (ATL) vs. No. 7 Carson Wentz (PHI)

See also: 2. Roethlisberger (PIT) vs. 11. Wilson (SEA); 1. Mahomes (KC) vs. 21. Carr (OAK)

FiveThirtyEight vs. the Readers

As a weekly tradition here at FiveThirtyEight, we look at how our Elo model did against everybody who made picks in our forecasting game. (If you entered, you can find yourself on our leaderboard here. I am currently in 1,995th place!) These are the games in which Elo made its best — and worst — predictions against the field last week:

Elo’s dumbest (and smartest) picks of Week 1

Average difference between points won by readers and by Elo in Week 1 matchups in FiveThirtyEight’s NFL prediction game

KC 58% KC 69% KC 40, JAX 26 +5.7
BAL 61 BAL 72 BAL 59, MIA 10 +5.0
CHI 64 CHI 58 GB 10, CHI 3 +4.8
LAR 52 LAR 59 LAR 30, CAR 27 +4.3
TB 55 TB 50 SF 31, TB 17 +3.5
SEA 75 SEA 79 SEA 21, CIN 20 +0.2
DET 51 DET 53 ARI 27, DET 27 +0.0
PHI 77 PHI 80 PHI 32, WSH 27 -0.6
NYJ 55 NYJ 54 BUF 17, NYJ 16 -0.9
LAC 72 LAC 73 LAC 30, IND 24 -1.6
NO 68 NO 68 NO 30, HOU 28 -1.7
DAL 74 DAL 73 DAL 35, NYG 17 -2.3
NE 68 NE 65 NE 33, PIT 3 -4.4
MIN 59 MIN 55 MIN 28, ATL 12 -5.1
CLE 60 CLE 64 TEN 43, CLE 13 -6.9
OAK 51 DEN 55 OAK 24, DEN 16 -7.9

Home teams are in bold.

The scoring system is nonlinear, so readers’ average points don’t necessarily match the number of points that would be given to the average reader prediction.

Readers knew better than Elo in a few notable cases — although it was all about the prognosticators’ degree of confidence in the favorite, rather than differences in opinion about who would win. (Elo was too bearish on Mahomes and the Chiefs against Jacksonville, for instance.) But Elo still won the week, beating the average reader by 7.9 points, thanks to a last-minute victory in Monday night’s Raiders-Broncos game. Perhaps because of the torrent of drama over the weekend, readers thought the Broncos would pull out the road victory; instead, Carr and the Raiders managed to win in spite of the tumult — and that was exactly the margin Elo needed.

Congratulations are in order to Joe Tito, who led all (identified) readers in Week 1 with 252.3 points. Thanks to everyone who played — and if you haven’t, be sure to get in on the action! You can make picks now and try your luck against Elo, even if you missed Week 1.

Check out our latest NFL predictions.

Buy, Sell Or Hold? A Special Democratic Debate Edition

Welcome to FiveThirtyEight’s weekly politics chat. The transcript below has been lightly edited.

sarahf (Sarah Frostenson, politics editor): Welcome, everyone, to a special debate-focused preview edition of our weekly politics chat!! In recent weeks, we’ve talked a lot about the different strategies the candidates should use on the debate stage, who the lineup is good for (and who it’s bad for) and whether the field might be consolidating around a handful of candidates.

So today, let’s have a little fun with the question of candidate debate strategy and play a game of buy/sell/hold with PredictIt prop bets (plus some I made up). We checked the prices (given in cents) of a bunch of propositions at noon Eastern on Tuesday and then translated those prices into probabilities. (We know that’s not exactly right, but it’s close enough.)

And in case you forgot how to play buy/sell/hold:

  • Buy means: “I think the chances of this happening are higher than indicated.”
  • Sell means: “I think they’re lower.”
  • Hold means: “I’m a coward and am unwilling to take a stand.”

OK, let’s start with a 🌶 spicy 🌶 proposition. Buy, sell or hold: Elizabeth Warren will win the 2020 presidential Democratic nomination? (33 percent)

nrakich (Nathaniel Rakich, elections analyst): I am bullish on Warren overall, but I still think it’s a pretty open race. I’ll sell on Warren — I think her chances are a bit lower than 33 percent.

clare.malone (Clare Malone, senior political writer): I’ll buy! If only for the argument. I think that Warren is a secure bet in this race and that she’s the only candidate who’s seen steady, significant gains. That’s gotta count for something.

I also think that she’ll pick up establishment-leaning voters as the race goes on and other people drop out — or at the very least, she’s one of a few arms that establishment-leaning types will want to fall into.

natesilver (Nate Silver, editor in chief): I’m boarding a plane, so my answers will be brief at best. Warren’s somewhere between a hold and a sell. And 33 percent is a lot, given where her position in the polls is. That price puts a lot of weight on subjective vs. objective impressions, in other words. I happen to share those subjective impressions, i.e. my “gut” says Warren is a very strong candidate. However, I’ve been doing this long enough to know that my gut is usually full of shit.

clare.malone: NATE’S ON A PLANE 🐍

geoffrey.skelley (Geoffrey Skelley, elections analyst): If Biden’s chances are, say, 40 percent based on his standing in polls from the first half of the year, that leaves like 60 percent for everyone else. Does Warren have a bit more than half of that? I’m skeptical of that so I’ll sell — for now.

ameliatd (Amelia Thomson-DeVeaux, senior writer): I think I’m somewhere between a buy and a hold? (Outing myself immediately as a quasi-coward.)

clare.malone: Lol, it’s fine Amelia — I’m a little less wedded to the strict odds of this. I’m more going on the gut that Nate speaks of and strength on the ground, etc.

ameliatd: Warren’s growth has been impressive. And she stands to gain as the field narrows. It seems like she might be starting to overcome some voters’ worries about nominating a woman, too? Which would be big, if so.

natesilver: I mean Biden is way too low in these markets, so if you’re saying Warren has a 33 percent chance or higher, you’re saying it’s basically a two-horse race. Which, maybe?

clare.malone: I don’t think that’s crazy, Nate.

Not a lot of people are super bullish on Sanders right now.

And Harris is slipping in the polls.

So … it’s not totally nuts to go with the idea that it COULD potentially be a Biden-Warren showdown.

sarahf: Bernie is in third at 14 percent, but that’s a distant third behind Warren (33 percent) and Biden (26 percent) over at PredictIt.

natesilver: If Warren got into the low 20s in national polls instead of the high teens, I might feel more comfy with that. That tends to be a big inflection point. Being in the 20s in a multi-way race is serious business.

nrakich: Harris still has a lot of untapped potential, IMO.

ameliatd: I don’t know, Nathaniel. My “buy” sentiments for Warren are probably contingent on Harris not pulling it together. But based on how her summer went, that seems increasingly plausible. Like, at some point your untapped potential needs to start turning into actual gains.

geoffrey.skelley: I know I said I sell on Warren, but I can certainly see her winning heavily-white Iowa and New Hampshire, building up — Nate’s favorite word — “momentum” and going on to win the nomination.

And that’s the tough thing about this — the sequential nature of the primary means we can look at the national polls and early-state polls, but the moment Iowa happens, that will influence what happens in New Hampshire, and so on.

sarahf: Speaking of Iowa and New Hampshire … Warren has a 35 percent chance in Iowa and a 34 percent chance in New Hampshire … does that change anyone’s wager? The markets do give Biden a 48 percent chance in South Carolina, though, and Warren only a 14 percent shot.

nrakich: I think that 35 percent in Iowa is considerably closer to reality. I think her odds of winning Iowa are higher than her odds of winning the nomination writ large.

She has an excellent ground game, which could help her in a low-turnout, activist-driven caucus state like Iowa.

And she has performed better in Iowa polls than national ones so far.

geoffrey.skelley: There are 26 days between Iowa and South Carolina and 18 days between New Hampshire and South Carolina. That’s A LOT of time for the “Biden’s a loser” theme to permeate things if he can’t win either Iowa or New Hampshire. Nevada, as it’s worth reiterating, is BEFORE South Carolina, which could maybe help him. But the Silver State is a bit of a black box.

natesilver: I’d buy on Warren in IA/NH. There should be more of a spread in these markets between her Iowa price and her nomination price. The fact that there isn’t proves these markets are kinda dumb.

ameliatd: That underscores my very definitive wager of “somewhere between a hold and a buy.” She’s doing well in Iowa and New Hampshire — which is why I would say her chances are better than you might think if you just look at the national polls.

clare.malone: Warren is going to have an easier time in those very white first two states — Biden, for a number of reasons, enjoys an advantage with black voters in South Carolina, which is pretty much how you win that state.

But Warren’s campaign was one of the two or three campaigns that people in South Carolina talked about as being strong in voter outreach. Even a lot of pro-Biden people there told me they liked her. So, I could see momentum or winning in IA and NH upping her odds in South Carolina and then perhaps in other southern states where black voters are key.

But that probably involves Biden faltering so she becomes more dominant — or doing something that finally sticks as a criticism.

natesilver: If you had a market for “will the Iowa winner win the nomination?” I’d probably be a seller of it.

Like, I think it would be priced at 70 percent when it should be priced at 50 or something.

The Democratic electorate used to be a lot whiter and have a lot more caucuses, so Iowa used to be a lot more representative.

But Iowa is no longer a good representation of the overall Democratic electorate right now.

geoffrey.skelley: Oh definitely, but if Warren wins Iowa and New Hampshire, I bet she’ll be polling better in Nevada and South Carolina than she is right now.

Also, keep in mind, Bill Clinton is the only recent nominee of a major party not to win one of the first two states, although that is complicated by the fact that no one really contested Iowa in 1992 because Iowa Sen. Tom Harkin was running.

sarahf: Interestingly though, the betting markets seem less bullish on Warren winning the presidency. Buy, sell or hold: President Trump wins the 2020 presidential election? (41 percent)

natesilver: Lol I’m not touching that question.

ameliatd: Wow, Sarah, you really went there!

clare.malone: I give a hold on that.

I truly don’t know where things stand.

There are a lot of plausible arguments to be made that Democrats come out of the nomination fight irritated at each other and unhappy with their candidate, and there’s less enthusiasm on Election Day.

nrakich: I’ll hold. Forty-one percent sounds right. Close to 50-50, as we are amid an era of close presidential elections. But every indicator right now, from his dropping approval rating to polls of the generic ballot to special elections to early general-election polls that you probably shouldn’t trust, indicates it’s a Democratic-leaning environment, so I’m willing to make a 9-point concession to that.

clare.malone: On the other hand, it could very well be a different ballgame from ’16, and people ARE really motivated to turn out against Trump, no matter who the Democratic nominee is, and the Democrats prevail against a GOP Electoral College advantage in a high-turnout election on both sides??

ameliatd: I would also say hold just because there are so many unknowns. A recent Pew poll found that most Democrats are excited by several 2020 candidates, not just their top choice, so maybe that’s good news for Democrats if it helps insulate voter enthusiasm against a potentially long and bruising primary? But I’m not sure.

geoffrey.skelley: Yeah, my feelings are similar to Nathaniel’s, though you could also make the case that 41 percent is too low for an incumbent president in a closely-divided country. So I’d say hold or maybe buy.

natesilver: He’s not much of an incumbent. He barely won last time. And the track record of incumbents who barely won is not so hot.

sarahf: And as Nathaniel mentioned, his presidential approval number seems to be 📉 (his approval rating is still above 40 percent though).

geoffrey.skelley: But his approval has about as good a chance of going up again as going down in the long run — it’s been remarkably steady overall.

natesilver: That isn’t good news for Trump, though, because he probably needs his approval rating to improve for him to win reelection.

clare.malone: Nate’s not touching this, though.

natesilver: Like it’s not good news when your weight is steady if you weigh 300 pounds.

clare.malone: It could be good news if you’re 7 feet tall.

geoffrey.skelley: A BMI discussion is what we need here.

clare.malone: Clare googles “what does LeBron weigh?”

sarahf: OK … let’s pivot back to the 2020 Democratic primary. We’ve talked about where people stand on Warren’s chances of winning the nomination, and as a result where Biden or Sanders stand — but what about some of the lower polling candidates like, say, Kamala Harris? Markets put her at 10 percent. Buy, sell, hold?

ameliatd: I am not especially bullish on Harris, but that seems a little low to me. I’d say buy?

nrakich: I’m definitely buying on Harris — I think her chances are significantly higher than 10 percent.

Harris still has the potential to appeal to multiple wings of the Democratic Party — especially two very influential wings in college-educated whites and black voters. I also think her prosecutorial background means she has several good debates left in her.

geoffrey.skelley: As for Harris, I think I’d buy there just based on potential. It may be unrealized so far, but it’s still there, whether she ends up being Marco Rubio 2.0 or not.

sarahf: What about Pete Buttigieg, at 5 percent? Buy, sell, hold?

clare.malone: What the hell, sell. I do not think he has a good chance. Maybe an Iowa win? Maybe?

nrakich: I’d buy Buttigieg at 5 percent, but I think it’s a closer call than Harris.

clare.malone: I think that the markets are incredibly sensitive to the narratives of the news cycle.

So, the fact that Harris had a purportedly shitty back half of the summer, means that her odds go waaaay down — too low.

natesilver: Y’all should actually add everyone chances up because I bet you’d be at like 130 percent.

You can’t be a buy on everyone.

clare.malone: Nate, you’re not even participating fully!

Peanut Gallery!!

natesilver: The best way to participate is half-assedly.

clare.malone: Also, close read: People have sold.

sarahf: OK, so we’ve talked about some of the leading contenders for winning the nomination — Warren, Biden and Sanders. And we’ve talked about Harris, whom the markets seem a bit bearish on. But setting those candidates aside, what about the rest of the debaters? Andrew Yang is at 11 percent (insert Yang Gang joke) while everyone is at 10 percent (Harris) or lower. Would you buy any of the other candidates on Thursday night’s stage? Or sell them all?

natesilver: I’m probably the most bullish on Yang of any of the election-analyst-types, and I think that price is kind of insane and a discredit to PredictIt tbh.

geoffrey.skelley: Definitely sell on Yang at that mark. Buttigieg is harder because he, too, has potential and a lot of resources given his fundraising. But he really needs to win Iowa or finish 2nd, maybe 3rd there to be in the mix, and I’m struggling to see how that happens with the other candidates in the field — so I’ll sell him, too.

clare.malone: I’m bullish on Cory Booker, though not incredibly so (and I do remain, sadly, sensitive to Nate’s buying spree comment).

I think that Booker has the same kind of coalition-building potential — very much unrealized yet — but I think he’s probably worth being slightly-higher-valued? I think he’s investing on the ground and could have an appeal to white voters in Iowa, and if he did decently, proving his mettle there, black voters in South Carolina or Hispanic voters in Nevada might see him as more electable?

nrakich: At 2 percent, I’d be willing to buy a few of these names just because you don’t have much to lose.

I will probably get ridiculed for this, but I think Beto O’Rourke still could have something left in the tank. There’s a reason he caught fire in Texas last year — he’s a charismatic, appealing guy — and he has been holding a ton of campaign events.

ameliatd: Here’s my problem with this lower tier. A lot of these candidates have potential upside — but mainly if one or more of the higher-polling candidates stumble. That certainly seems to be the case with someone like Booker. So how much do you bake in the possibility of another candidate falling apart?

nrakich: I am also bullish on Booker, Clare, but it looks like PredictIt is too — he’s a bit higher than all the other lower-tier candidates, at 4 percent.

natesilver: If one of the front-runners stumble, wouldn’t one of the other front-runners benefit?

geoffrey.skelley: Yes, I agree with Clare and Amelia. Booker definitely could be there to pick up the pieces if Harris slides, so I’d buy at that price.

natesilver: But Harris is in 4th place now. Which pieces is he picking up?

sarahf: Yeah, I think it’s becoming increasingly harder for some of the candidates like Booker, Amy Klobuchar or O’Rourke to see any gains.

nrakich: Sarah, I think that is obviously literally true, in that every day that passes is one fewer day until the primaries. But I still think there are several months and several opportunities to stand out (e.g., debates) left.

natesilver: I’m bullish on Booker’s odds of finishing in 3rd place. I’d pay 10 percent for that.

sarahf: I keep coming back to Julián Castro, who had a strong moment in the first debate and saw a big jump in his favorability ratings and name recognition as a result, but then nothing in the polls.

ameliatd: But I’d put Castro’s chances higher than O’Rourke’s right now (if I’m choosing between the Texans). Maybe I’m wrong and O’Rourke will wow everyone in the debate … but his performances have been super snoozy so far.

clare.malone: I think Castro suffers from not being a national figure. He hasn’t had the same kind of grist as the senators have during the Trump years, and O’Rourke, his home-state rival, really crowded him out in the roll-out department.

nrakich: Yeah, I agree with Sarah that if Castro was going to make his move, he would have done it already, as he performed quite well in the first two debates.

To me, he suffers from inconsistency. He was great in the debates, but I thought he was really flat in the latest CNN town hall.

clare.malone: In the end, no one remembers who he is. The twin thing doesn’t help either …


natesilver: I’d just say I have about 15 percent total to give out at most for everyone who’s not in the top 4. I’m not sure how I’d distribute that 15 percent, but it’s not a ton of wealth to go around.

nrakich: Yeah, I’m definitely splitting hairs — who cares if O’Rourke has a 5 percent chance instead of a 2 percent chance.

natesilver: FiveThirtyEight readers do!

ameliatd: Hair-splitting is what it’s all about, Nathaniel! Lean into it!

natesilver: And 20:1 vs 50:1 is a nontrivial difference

sarahf: OK, what about the candidates not on the stage? Lest we forget, Tom Steyer has made the October debate … but does that really change anything in your mind regarding who wins the nomination? My guess is maybe not … so here’s a buy/sell/hold I made up — Warren, Biden, Sanders, Harris don’t win the nomination (5 percent).

nrakich: I’d buy that, Sarah. As I said, there’s still a lot of time left, and several more debates where someone like O’Rourke or Booker or Castro could have their moment. I think Biden, Warren, Sanders and Harris are by far the most-likely nominees, but I wouldn’t be gobsmacked if it’s someone else.

clare.malone: I really don’t think anyone is going to have a moment who hasn’t already.

geoffrey.skelley: I think I would cautiously buy that. It’s obviously very likely that the nominee will be one of those four, but we’re still nearly five months out from Iowa so things could shift.

clare.malone: Call me a cynic, call me a stinker, but if you haven’t really started to prove you’ve got the potential for a coalition by mid-September … sorry, but it’s curtains.

nrakich: Clare, you’re the one who always says it’s early!

clare.malone: RIght, but now it’s September — people outside the top 4 or 5 have no shot.

And that’s the scope of this q.

natesilver: I mean I just said I think it’s a 15 percent chance on the high end. Maybe it’s more like 10 or 12. But definitely higher than 5. Five percent isn’t a lot!

nrakich: By this point in the 2012 Republican cycle, Herman Cain, Newt Gingrich and Rick Santorum all had yet to have their “moment.”

sarahf: President Cain was such a good president. 😉

clare.malone: heh heh

natesilver: I don’t think this race has much to do with 2012. The frontrunners are a much more robust group than Romney alone was.

And to Sarah’s snarky comment … all those candidates lost anyway!

ameliatd: I guess the way I’d think about it is — what are the odds that some series of events manages to tank the chances of all four of those candidates? That seems quite unlikely to me.

natesilver: Yeah it’s like being five games behind but in 5th place in the MLB wild card standings. #sports

Doesn’t seem like you’re that far behind, but it’s unlikely that everyone else gets cold.

geoffrey.skelley: I think everyone wants every cycle to be like 2012 because it was exciting and messy. Also because who can forget Gingrich and Cain having a “Lincoln-Douglas” debate? But yeah, I don’t think you’re going to see a collapse of all those candidates.

ameliatd: And it’s not like if Warren suddenly starts slipping, Buttigieg or one of the lower-polling candidates will grab her supporters.

I think there’s room for movement — but mostly within that upper tier.

nrakich: Nate, you had to turn this chat into a dig at the Mets.

FiveThirtyEight Politics Podcast: How to win this week’s debate