Transcript for Model Talk: Trump's Electoral College advantage | FiveThirtyEight
Hello and welcome to the five to politics podcast I deal in tahrir. I'm Nate Silver. And that's and it's a game on the outs Iraq. A bit less enthusiasm than usual. And it's been a few weeks we've gotten together. Listeners also model out of questions so working. Did the riddle now's the time. As of Thursday afternoon the forecast shows why it was about a 70% chance of winning and from which he 30% chance and our national poll average shares Biden meetings from my seventh meet the read. Sunni on Monday we said that we didn't have enough recent polling to fully characterized the race since the conventions and the violence and what did we learn from the most recent slate. So there were almost. Twenty different national polls I believe that have been released since the conventions. And so there's a lot of data and they show that trump. Did not get much of a bounce. Maybe a little bounce so trump is currently behind by seven point three. He was behind by eight point four. Before the convention circuit and nets out to I said before either convention Biden. Moved up a tiny bit and it shifted back a bit toward trumpet overall. Purpose maybe a point. Closer to Biden nationally than it was before you know roughly seven and a half percent of eaten half. And to have only need a point. After your convention. When your down eaten hat is like is like not great right. We've seen past conventions where you know in 2008. John McCain and Sarah Palin. Briefly pulled into a tie with Obama after their convention right. In torn. Hillary Clinton got it pretty doubts from her dimension went from what you three point lead to like a seven point lead. So this is really on the low end as far as a commitment shaped shook up the race not a direction. And how has that affected our forecast model if federal. So not very much in part because our model. Number when it kind of builds and an adjustment for the convention bounce and also the kind of hedges a bit council explain each one. You just emerged looking a commissions historically. And trying to it trying to guess. How they mention balance will be a how many gotten smaller and smaller over time. With increased partisanship we also. Assumed for this year is one of the things you have to make an educated guess about it because of the virtual nature of the conventions that. You might not have a traditional typical convention bounce so. We just sit well it's kind of halfway between. In real dimension and not and there are for organist him that much about to be half as large as before. So so how big of a bounce did. Forecasts assume the candidate Mike. So the model assumed it would bounce and if you start at eight it would assumed to bounce up to around. Ten for Biden. And then fall like seven so it's actually very close to what's happening Biden I do not quite get. By did not quite get up to ten you have to nine point something right. But yes it is kind of just guessed that there'd be this kind of half. Convention bounce in about right the Oracle got it right but in any case so the forecasts. More lax saw this coming. So does not change as a result of basically act right so means Biden's bounce was like a little bit smaller than it bought. And so it dipped down a little bit when Biden got me beat the pointer appoint only in the polling average. But that it wound up. In the same place and expected it to wind up some kind of you know things are pretty small ships right there went from like. Biden at 72% to 67% netspective 70% or something pending. Constantly arriving new polls so. I mean even if it had gone from like 72 to 67 and steam it we're not talking about a major change. But you know. But from what I understand the forecast basically says we expect the candidates to get a convention bounce and so we're not going to take that into consideration. In terms of who will ultimately who ultimately is likely to win unless it's adorable routes or Costa week. Maybe two weeks or something like that to show that movement in the polls is real and not just a fleeting a fact. Is that. What we're seeing clearly had not change according to seven you know speaking to a seven and a half point race until another week or something like. So there are kind of two defense mechanisms of the model has right. One is actually to you but just the polls. For the anticipated convention bounce which is what we've been talking about. And two is to just hedge right we're basically says here's a snapshot that we had. Before the DNC and we're gonna use that snapshot. As part of our forecast even though we have newer data because. Action the commission sometimes. Polls. Change it and revert back to where they were before us that use some of the old. Pre to mention polling average. In this case. Again kind of old average. And the new adjusted average. Kind of winds up being the same right as if it says okay. Trump's minus seven point three we just had to eight point three in the previous average was eight point four write to Karen doesn't matter. What the formula was anyway. But it also serves as a hedge their other hedge by the way will also apply. After the debates so. If trump has a great first debate. And all of a sudden the polling average Joes from Biden plus 82 Biden plus forward. Then the model had a little bit at a time but a little bit until it knows that gained trip has made has been more sustained. So we've gotten limited. State polling at least not as much national polling as we have since Monday. But in particular are Monmouth poll from Pennsylvania Kmart stores showing items lead down to three points there are average shows him leading Arab four points in the state. And given that are forecast at this point sees Pennsylvania as being they'll likely ST to be the tipping. What does that say about the expected popular vote Electoral College GAAP you know at this point in time in the beginning of September. So I think it's always a mistake to focus on. Any one poll if you look at all the state polls sleeping. Pretty good for Biden accurate Dickey got the good for him said of C polls in North Carolina. Arizona and Wisconsin. From Fox News. An okay result from. Monmouth in North Carolina and in this frankly pretty mediocre poll from moment in Pennsylvania. Abby mammoth two different and abide would wanna be up more than two points among likely voters would of the poll hasn't pretty close right. But it is 1400 person poll the totality of data. It's a little perplexed net Biden keeps getting these kind of mediocre polls in Pennsylvania. And then pretty good polls in Wisconsin. You rethink that Wisconsin will be the harder state to you win back but you know the polling continues show was cuts and having a larger bite margin in Pennsylvania. The recent like trump has a 30% chance and not a 15% chance or 10% chance is because the Electoral College. There is still a pretty B gap between. We're the tipping point states are polling. And where the popular Broder is right CC Biden with leads. You know. It's caught four to six points the tipping point states. Verses seven and a half points in our national average. That's pretty different right if it's really like a four or five or may be six point race. Then. Trips and out of it trump could mean that from drew Willett debate and then maybe the polls are a little bit off right. And then you have a second term for president trump and so Electoral College is so. So important and I think people kind of like. I don't know and it's your like can do that mental math and accumulate OK. How titans up seven or eight it's like. Bigger than or equal to Barack Obama's margin that's like bush over Dukakis roughly pretty close to it right. Not far from Reagan over Carter was nine point seven points or something right. But still. While a eight point win would be a blow out a landslide by most people's terminology. All of a sudden at five points. It's a competitive race because Electoral College right there are scenarios we're Biden could win the popular vote but. Four and a half points. And still lose Electoral College after about five point two becomes. Mathematically more difficult although not impossible. But you know that's keeping trump. With much. Livelier odds than he would have otherwise. Yet it seems like there's a thin line between what's a blow out and why is a tight race given. What we're seeing an Electoral College I want to read. The numbers that you posted yesterday in terms. 1% chance Biden has of winning given on it is now it's given his national polling league season it. Biden leading 01 point he has just 86% chance of winning ambulance for college. One to two point 822%. Chance at winning the Electoral College two to three points 46% chance of still you know. Below majority chance three to four points as what you got a 74%. Chance of winning the elect or college. Four to five point 89% chance five to six points 90% chance and it goes up from there. And so it's really that divide UC between maybe three points to five points where you go from its a really tight race last. Biden is not favorite necessarily to win. To all of a sudden you could have something looking like a landslide in the Electoral College you know why exactly is that can we tease out that kind of thin line between the tight race and elect Paul blow out. So let's look at it this way right. Right now but Biden Biden is leading international polling average by seven point three points which. Is exactly the margin was seven point two hours and we've read and Obama beat McCain died. In 2008. Similar look is to say. Where is Biden doing better than Obama. Because by definition mathematically. The place where he's not doing better yes we doing worse to kind of counteract that right. So one group of seats for Biden is doing better. He's very Blue States. He is ahead in a California polling average by like 32 points. Instead of 25. He's ahead and test used by like 33 points a set of 25 right you know New Jersey. And Connecticut Maryland. Right. Washington State Alison some longer. Fifteen point lead but a twenty something point lead. Twenty fire restriction record earnings are these really Blue States we hear. Democrats are winning biker dance when margins that helps them not at all Electoral College right. So kind of wasted votes OK the other two states where Biden is doing better. It's kind of in the Sunbelt. So Biden will almost certainly be better than Obama in Texas he may win Texas it's pretty close right Biden will very likely be better than Obama. In Arizona for example. You know drew jobs is trending more purple but those states don't actually. Help Biden all that much I either. In less they actually become what we call the tipping point state right so you know again Texas is very competitively nappy surprise if Joseph Biden. Wins Texas of them are bottle has compasses. A slip the clear favorite there's some of their models if Biden wins Texas is probably RD one. Air zone and Florida and North Carolina. And so it's not at the tipping point right it's superfluous. Now Texas shifts further. In a couple of years then that becomes more a problem for Republicans and you start to erode. The cure the current Electoral College advantage they have. But like a vote in Texas. Is actually less important. The name though in the average state. I'm our model that yours it's more important vote New York or Washington DC. But. There are lot of people there. And it's not very good return on investment although it's better than it used to be so basically all the you know also Biden skating it's a small thing. Democrats have gained ground with mormons see you make up make upload ground in like you time Idaho. Your jacket when those states though probably. And so it's something light. Biden has to win by a say three or four points in order to hold the midwest but then all of a sudden if he's winning by six or seven points. He's past the point. Where it's like closing he's actually winning some of those Sunbelt states which put an end to something like rental our territory. Yet and yet it. Six or seven. Or eighty. That's when you start to we in. George and North Carolina air isn't currently RD one at that point bees are when Georgia. North Carolina Texas. You might also win back Ohio and Iowa. So yeah there bunch of states that are alike. You know polling currently it. With by with a seven inch point national lead us and between you know Biden plus to North Carolina added truck plus two and Iowa or whatever right. That whole group. If Biden you know beaches both likable points a whole group probably goes Biden and Elvis and you have. A really impressive looking map. But they're not really tipping point states except maybe North Carolina still on the fringe. So we focus on polling here when it comes to the fundamentals in terms the economy Maurer the trajectory of Kobe case is how is actually are forecast. I. So coated cases in sales. Only figure to remodel and a very indirect way. Having do with how the states are correlated so so leave that aside for now. The economic news is generally generally been pretty decent. Relative to this obviously. Unprecedented downturn. That we took when the economy closed under code and but yet one thing it's helping trump in our forecasts is that our model assumes that. The race will tighten. Because economic conditions will net basilica bad by November. Is it jobs report coming out on Friday morning that you may know the results about a stimulus in this. You know that a thinking where forecast. The stock market eggs into work forecast right but in general. If you we're assuming that you're gonna have some type of typical. Recession. We have a slump that lasts for word. Months or years. That's not what's happening here right you don't usually at this kind of theory share. Uptick just a couple of months later. And so you know. We think by the way if you look at approval rating pollster company economy. His numbers are pretty decent like 49% approve 40% disapproves of like that. So we think that like. The economy is. Perversely. May be more a strength rate from. Then a weakness that voters buy the argument that like. Things are going great until Covert came along. They enabling trump for not handling Coldwell. That's one reason why he stumps every points right. But they buy his excuse that. Cove it is not an indictment of his economic management experts say. Also remember people got a lot of money in their pockets. Back in the spring that is starting to Wear off that could have an affect potentially. But if you give people money actually some keys were making more income in the had been forward. And that will affect perceptions as well. We have been inundated by questions from listeners lots of really fantastic questions are gonna get here as many as we can of course will be back with more modeled off in the futures we don't get your question today. Hope we will get rich in the future. And given that world try to keep these answers relatively concise bullets are with Chris who asks. What what didn't outcasts are Saturday today. Essentially. If the now cast was are forecasting used to say basically what the model project if the election were happening written. So essentially if the only source Marin a want our forecasts. I don't know whose equipment and outcasts. In kind of a broom closet and don't talk to anymore. Act I. I would think you'd be in the vicinity. Biden being in 90% to win the Electoral College something like that I mean look the seven and a half point lead. Even given the possibility of pulling air and terms Electoral College advantage SM and a half point lead in national polls. On Election Day. Is. Recently solid. It's not rock solid though like. You know if Biden is up pipes and a half points on November 2 and were doing the last version. A vile talk right. I think we say okay they're anti pretty big Capone screw up or. The Electoral College I was gonna acted. Fall just right for trumpet it's possible that a one in ten chance wouldn't be nothing that's decently high. But that's more the realm you're getting a little bit more on the tale. Speaking. Polling errors we got a question from a reporter through. Asked that her name not be used to I'm so happy that we got a question from reporter. And the question is can you explain weeding by education I was weeding by education factored into the current polls that foresee. So essentially what does that mean. So. To back up a little bit here right. One dirty little secret about polling is that if you just like kind of randomly call people on the phone. You'll not get a truly random sample. Women are more like that it's the phone than men. Older people more than younger people white people more than in. Black and Hispanic people. So you have to wait. Your pool to populous demographics to basically say okay. We know we only got 5% of black people in a state where they're going to be crops and turn us there for let's count every black person. Two and half tax rates can obviously what happens. Now one of the variables that also kind of distinguishes him the answers of Poland who doesn't is based I kind of news consumption and education levels. And if you are more highly educated. You read more news you more likely be interest in taking a survey. Dan you're also more likely to respond to a poll. You're us a bit more likely to turn out to vote but it still. There is a bias in polls toward people who are. More bigger news consumers more educated in the sense of having it. College education right. The education of life might be different but you know have more formal education. And it used to be that education was not very predictive. Of which party you vote for word but now it is weird. The more edge to people vote democratic primarily. So if you don't. Wait recent education or some proxy of education. They and you risk. Having to me Democrats and your sample. Most pollsters. Well in unanimous. Many pollsters do wait for education now there kind of a few did before hand if you were made aware of it by 2016. Some other pollsters. Do not wait for education. But. Have other mechanisms they used to try to avoid the bias that could be inner beast by not doing that. And then a few pollsters are kind of oblivious. And they are doing what they always to aid probably at the rescue mover rating. Democrats again. Now is this reason I think if there'd be a democratic. Bias in the polls height I don't necessarily think so. When reason why is that you may have some pollsters it kind of went up. Over compensating in the direction. So they'll correctly way for education will be they'll also use like he really different. Likely voter screen. That actually isn't helpful and there's like well we got to make sure we're not an underage trap again so. So they do three things for the ailing east you one thing and wind up missing in the direction right. They're also for Felix. Automated polls. IV are polls that. Are very dot G method to logically. Some of them don't. Call cell phones there's one pollster. To photograph it tries to make you shy from voter adjustment which is varied duchy. And so you know so it may be a case of like. You have a few straggler like collar polls but don't. Think about was education thing. And you have a few. Spam me crappy. Robo polls. Are kind of Republican leaning in don't even bother to fix that because they're you know. And it kind of winds. Canceled cow and a follow up question. And here was does this happen any relationship to shy trop voters and I think the answer here is no waiting for education takes care ought. People in the population or difficult for pollsters to reach. Whereas shy trump voters usually refers to people who support trumpeter actually unwilling to tell pollsters now. Right I mean the notion that like. Like if you. Actually just randomly dialed numbers and made no demographic weighting whatsoever. You probably wind up with their really trump. Sample looks like a lot of people white people use by their fountain. Right younger people. Are hard to get on the phone even if you maker PT contact right but like all the people with landmines. Are more likely to it's just kind of been how he's been a center right he did the phone you answer stranger's phone call. So it's not a matter shy. Per say writes a matter of like. There are different by its season he got on the phone. And either once perfectly honest then. You're not gonna get a totally. Truly random sample and have to. You know. Do several things try to wait your sample to true population demographics. And just to get us out of the right because people happen curious. Is there any reason to expect that respondents born be honest what pollsters. Com. Not particularly. There is some evidence. Social desire ability bias weird. We years. If you believe something which you think that like. The stranger on the stone might be offended by that you might. Not say it. But generally speaking. People art embarrassed by the presidential choices they are happy to talk about them. And you know it's it was a big theory that like in 2008. That a lot of people would say they were voting. For Obama. But Witten is in one it you know. They are racist but that in one it. Being mean to this pollsters thought okay this person into a the F American right and Obama kind of hit his full spot on actually outperformed him a bit crisis kind of one example of social desirability bias not actually having an effect. Me part of what's weird is like. Number when the idea like. Trump voters are shy. And now I have you met some from voters I don't think they're particularly shy about. About their phantom for trump if anything there may be. More demonstrative about it so maybe they're shy Biden voters' right. Also the idea of like okay you can it go ahead and take this poll in and lie in a pollster but that you can and you know. Give your honest answer that in the voting Booth and that just now actually much evidence for this right. We've also looked at. At. Because the US it's a small sample right you have had a can't electric before. International though there have been many nationalist parties. Throughout Europe for example. And if you look at all the polls of Europe involving. Right wing Nationalists. You know racial identity politics parties. They do not do any better. Then polls show. Even those kind of constant myth that others right wing party in Germany year Damron going to be they're poles you know it's just not true over pretty large sample one other thing too is like. You know a lot of Paulson or not done. By the phone anyway they're done online. So are you also. Now acting or really real opinion on line right and there's not much of the gap between. With the online polls say. And what telephone polls say so. I mean the short term bowler theory is. Under evidence. Under theorize. Kind of paints a weird view of trump supporters I don't think really kind of matches. Any anecdotal or empirical evidence. And although we need to be the area we're. Of the possibility of systematic polling here meaning the polls are office seem direction every single swing states. The assumption that has been much more robust of retirements and acting ghetto in either direction. It's absolutely possible. It wake up on election morning. And trump beats is poles and wins again right it's ultimately possible that like would wake up and Padilla to election right and Biden is one by. Twelve points to sit at eight and one South Carolina or something and we're like what the hell was that and it turns out that impulses we're doing enough to capture. Black or Hispanic voters right or maybe they were kind of being very careful because a 2016. And actually went over compensating right. So I know I will armistice and Twitter before like. If you go out and talk to me. About Shaq dropped voters. As if being as opponents of one of many. Possible hypotheses but if you're like a welcome to it is a check from voters you know I can think that you have a sophomore view. Of elections. Where. You re just not to think that you have some proprietary now you don't actually read. The really not for people who have looked at this pretty carefully and said there's not really much evidence that it's a thing. Are so let's get on to some more listener questions the next one it. Is do you ever worry that the models tales might be too fat. And the birds that some of the edged cases are pretty far out there. We don't multiple questions about this one person noted that there's there's variation on our forecast whom page. Where New Jersey goes for trump and then every other state goes for buy it and as we show some of these examples that the forecasts in the rates and so essentially people just want to know like WT. Is op wears some of those entities. And about the New Jersey wonder. I mean in develop we actually had like a bug where like one in every 100000 simulations like. Blue state would become red. So managers US house weird all the other stuff is is very deliberate. And you know so it number one. It used to be that was pretty hard. To predict how things shake out state to state now. Since like 2000 on words twists and you had kind of the same map. Right. But. You know that's actually a bit and usual. Historically. And we want our model to be robust suited. Two different regime changes right some here there's going to be a realignment. It may may not be detected ahead of time right. And so we need our model will be robust suited. Not just assuming because if you assume if you assume OK let's calibrate or model off of 2000 toys sixteen the states are very predictable. And the polls are pretty good US emissions between Eckstein obviously right. If you cabaret it off just the recent data then maybe trump would be 85%. To lose or something instead of 70%. But we don't think it's a good idea or hate it's very small sample. Number one and number two we are living in somewhat unprecedented. Conditions and it's true that. True that's a particles have been stable and if the polls remain seem like a said. Biden's chance will come adopted. 859092%. After something right. But we think dispassionately. Before the conventions immunity listen now it's. Partly too many presumptions about kind of but things to look like in cuts goes and say okay but to all the data we have right we have. National polls elect 1936. And state polls connect 1972. Or 76 right. You know look at all the data we have then. It would compel you to be. A little bit cautious. But again like it's also like. Because you it is big national lead. That Biden had us right. Anybody is only ahead by. I three or four points. Projected. Not a plan for projected in the tipping point states right if the model expects a race to tighten. By a point or so. In all of a sudden he's up you know 45 comes three or four. Just say okay it clear four point effective. Projected and Election Day march and on September or early September. Has about a 70% chance of holding up right. That's what you get incorrectly. The way we do you are model. I think intuitively that like that kind of holds up right OK there's an. You know between the possibilities change further construction in the polling Aaron Election Day write about a city to chance it Biden wins seems like a pretty reasonable answer. So our next question comes from. Chloe who actually went to culprits shout out to club she asked. What is need to biggest concern about the model. And then we got a similar question that was. What do you think is the model's biggest shortcoming another way to put it might he if you add unlimited time and resources what would you change about them. Two and concerns about his drew actually are caught a strict. Well. I mean OK if you wanna be honest like you can kind of like. A like no election that I remember root for her. The model has we're kind of pulling back the curtain a little bit right. There's a humongous amount of running room that the model has between. On the one hand. Like prediction market to have the race a toss up which is ridiculous. And neither hand. Other models that have buying it 85% what you think is even ridiculous but like. We don't believe reflects modeling that's practices. And we have believe reflects the kind of environment. Isn't you know it's almost like conditional forecasts were if your processor is true about that things then OK can see Hank Payton get there right. But we don't know about listings in advance in an environment where you had. A small sample size seeking place in unprecedented conditions of by the way. The you know the models weren't so accurate between sixteen fighter right if you think okay now we gonna solve this riddle well okay. How can he had this mediocre performance. For the polls or projections between sixteen naturally great fact right so like. So a lot of firemen there and watch a lot of running room where like for. As an it be concerned obviously less concern right because like. Wherever the graduates and what are you most concerned not what you bees can. I'm not really a concern and am concerned that like. For the actual election itself. Right I'm you know I'm concerned for what happens give a close result. And it's disputed or a not so close result in trump refused to concede right. What if there is some based F two with. With mail balloting would you try to account for some of that we do you look at like I mean we're sitting that turnout and their forward. The margins are less predictable because a coded and mail voting. But like. But I don't know I mean I would not I don't think the model is the most important thing for the country right now in and I mean. I think it's fair enough let's move on and right so I want to get you some. Rapid fire questions so in. A word or sentence. Let's run through these corrections. Does the model accounts for combing letter like events or other unforeseeable October surprise us. Sure yeah I mean the call me later is kind of it. Normal October surprise in like what ordinary news events his use. Even somewhat extraordinary news events is is. Part of why that binds it symphony in that and higher right now. Next question comes from worry does the model Ron every X amount of hours or when there's a new set of polls how frequently does that refresh. It runs whenever. We inner new poll and it also will run. At 7 PM every night just to nature's one at least rents. Per day but now there's a ripple they so usually runs constantly. You know there also other inputs like technically speaking. Every tick of the S&P 500 technically affects a model right. That we don't update it for economic data alone with its own nature variable that has been updated. So mostly running it. You know Peter time in ten times Dave or entry in a new poll plus once. For the early evening crowd. Next question. Which state is most likely to flip from Democrats to Republicans and a scenario where I and wins the presidency I'm guessing New Hampshire minister. A cracked. Yeah advocacy can you know they were pretty close last time. And Biden can afford to lose New Hampshire Minnesota. Well actually let me take that back I think probably see ticket but the most are. New Hampshire and Nevada. Because that's kind of collective went off right as the exact math and Biden kind of winds back those in western states and he can afford to lose. I think it is I didn't Nevada or New Hampshire but not vote. There are some maps appears to CC next nine ties it comes down to like does he also win. Congressional districts in Maine and Nebraska. But yet you capturing about our weird states in that they're kind of march their own drummer a bit. Minnesota. If Biden loses Minnesota than well that he would Wisconsin ended when Michigan. A little bit dicier. Not the case and where it's like the sun belt ended up being super strong provider and the poles were pretty. Mismatched in in expecting that in the upper midwest will be closer and I'll. But it rights question and there be a pretty sharp reversal yet. Our next question. Where would you most like to see new polls from right now I mean we just got a lot of new polls but I guess is our particular ski that you want you calling from. Eyed looks is a Michael articles in Minnesota. Nevada mean a thing is we actually in the its can democratic issue now are right is democratic lack of of high quality staples periods looks like never liked you even get two point now we like oh they're five quadruples the state if you'll saturated you know I would still take. More polls of Pennsylvania and Florida just because they're so important. To our model. You know. Every we're really right. I was gonna say will be fun to see a poll of kind of Texas here venture in the year have been fewer recently but like texas' actually. Not knowing that attributed to influence its immediate matters bit less. But I don't know and we actually have a different amorous couples Pennsylvania just kind of like. If such an important statements kind of still to me hard to reconcile why Biden is like doing surprisingly well Wisconsin. And everywhere on Pennsylvania just kind of weird. And if you're one if your tree sitter crescent say. Minnesota. And Nevada. Ari record that. What. Next question is. How are you considering the impact of third party candidates in the forecast. I was just look at the forecast from two dollars and sixteen other day we actually had Gary Johnson. We've stated with his chances of winning the electoral college of course it was very low by. You know people are looking at four processor can tell that were not protecting the odds. A libertarian or Green Party candidate winning the Electoral College is it still affecting the forecast. Not. Exactly. I mean so. We have what we call a named. Third party candidate right which is at the candy is polling if their eight. Polling yet. In the mixing of its nationally. The are included in most polls. And see our on the ballot in mr. Allstate's right. Then we will model them explicitly. Gary Johnson ministry Kerio last time around nobody does this year. So motto will reserve. So vote for other candidates. But. He doesn't do anything picker early fancy with that right now to the mine if it's a libertarian candidate or Kanye West whoever's on the ballot in certain states. Many pollsters can ask about them. And so if. Kanye West in the balance of swing state hurts Biden by appointment to think about ways far from clear can be it away. Then. That would affect buying sending the polls and so therefore the model would account for that implicitly. But you know their party cancer Nana detected this year by the way there are also not like. Any really. Highly relevant third party candidates in races for congress is here either. Would you ever consider making a model open source I know the answer to us it's no because there's a need Nixon money c'mon guys. Kind of tennis I mean we so we tried very detailed methodology provide all the inputs the models fight all the output it's right. You can. You know let's be perfectly honest straight. There are people that have. Reverse engineered versions of the 530 model. Without having the code but now I mean the code is. Who is proprietary. Last question here when of the house and senate forecasts coming out. They're coming out soon I am kind of finishing up the initial version of them. Today we are gonna so. When he bought the house the senate is like various like. Adjusted ton of data. That we used to house and sent model. In some ways to much more like. Rich data big data exercise in a presidential election where you just have twelve examples right you know for race for congress you have like. 470. Races every year which are somewhat independent when other right lots information you can use and so. It takes a while to like actually Rangel the data. Where do a couple of things to and we were we really like to return modeled when he eighteen perform really well. And we've designed to pretty carefully so we're not making a lot of changes. There a couple of things with with respect like how the mountains of house effects in a polling averages. Maybe accounting for word the effects of partisanship a bit more effectively there a couple of things around the margin weird we. May introduce some changes but will be the be pretty minor will also assume. For congress as who were the presence it is a little bit more. Potential for air on Election Day because of Covert and male vote on a ton more alike makes things a little bit we're here for a. And I'll say we did get a number of questions about for example party fox the new fox scares her that helps. You have to understand the forecast we also got questions about. The presentation in general why there's no map this year things like that and it actually want to have a special models hawk episode where we bring on. Some of the factor dieters who helped design the forecast and reaching kind of talk through some of those questions that invite. Have heard your questions and doubt we will we will get to them. In due time our final comment. Comes from Kean and Keenan says people should stop cheating on its under lit mid century modern masterpiece. So there's I'm my. Advice for CNET. I think you key and I appreciate it. It's not under let it slip for people living in the house right it's for people who aren't you looking outside right it's not lit her. Propellers. And if you're over television for you to. Ferret out people should not a defiant our houses so that they're well way for you to hear us. It SI really looking forward to when you have our. Design interactive people line you know I do you think that like. I think philosophically. Right. Tony when he is. A weird election right it's a weird it's a weird. Time we pretty on election forecast right. And so maybe some visualization center Brit more kind of unexpected in and surprising. And a little bit more kind of nonlinear write like that is. Intentional to some extent you know could I mean. Based what it wanted to reflect the cast of 20/20 in our forecast. Clearly it's yet it's been simplified actually I think right. But like. But it's telling little bit more has a littler bit through line with forecasts. And tenants being a dashboard we might revert back to where desperate looking for between two it's a mid term you know frankly that viewers we have for mid term elections are our. Are different they're much interest in all the detail matures in cells are murky tale right. By the way. If you don't know if you go down. In our forecast and you can see a section called. You know look predict. Like it's as download data right. And you'll find several additional files that includes more detailed data and then we published on the interactive. And we're adding we're files that all the time so. Some info that like. That hardcore users wanted. It's still there. It's in a place where you can download nice. CST file with all that and so in effect on him for going back so we keep we're anger that still we gonna add a few more elements. Keep minuses like. The first time ever where we try to forecast the presidency. The house and the senate in the same year. Frankly looks like it may be a bit anti climactic because based on in. Initial returns the house models will everybody else's house model and an expert forecasters and whatnot. The house does not look super competitive. But we are trying to give you everything this year were skipping governors' races of story. Sorted Indiana and North Carolina in the weird states you know for Matt New Hampshire you have your. Every two years you know whatever story bright New Hampshire. No gubernatorial forecasts this year but we are doing everything else. Our law. Clearly your creditors are considered and you just mentioned so we'll leave it there thank you name. Thank you get. My name is Dillinger Tony chow is in a virtual control room you can get in touch by emailing us at pod cast at 538 dot com. You can also of course treated us with any questions or comments have future model talk episodes if your questions weren't answered. This time around. You know he Melvin again rich we can etc. If you're a fan of the show beavers are reading or review in the apple podcast or or tell someone about us also subscribe to 530 on YouTube. Dates for listening and Wilson's. And yeah.
This transcript has been automatically generated and may not be 100% accurate.