Are the intergroup myself right not much as have fun hello and welcome to to define thirty politics. Podcast I'm Galen Druke Drink, I'm need silver and that is as models, Hock weed. And we're boisterous model tuck interests before. That was a subdued when we have maybe it's because we were up late last night and were already back to the grind. casting again, I have to say I didn't really get any of that rests in and relaxation none of those. Answer yoga sessions that we talked about on the past year. I take another and this morning rents less bike. what are you doing better than I am, and I

I wanted to mention that this is likely are anti penultimate model talk of the year. Eighteen can ultimately seventeen. Third to last, I what did not. I I didn't know that so we have today. We have the Friday before the election and then probably once we get results, will have a model talk round up, but yeah things are coming

down to the wire now I M excited I'm trepidations. Well, let's give folks a picture of what the forecasts looks like with eleven days to go until election day, not that much has changed. So the split for the presidential race is still eighty seven percent chance. That Biden wins twelve or thirteen percent chance that Trump wins Democrats have a seventy four percent chance of gaining control of the Senate and in ninety six percent chance of maintaining control of the house. So neat: let's focus on that steadiness to start with water. The dynamics at play here because we ve said that, as we get closer to election day, if the polling remains the same Democrats,

this would improve, and the forecasts has been relatively study for almost two weeks now. So does that mean that the racist tightening a little bit and that's why we're seeing those odds remain around the same? What's the overall seem like, like? I don't know, if I'd say it's them has been steady for a week or something but look the national poles have tightened a hare. The forecast LISA Staples, it's a bit less clear if those have tightened. However, this poles were never really consistent with Joe Biden having us like eleven point, led look more like a nine point race so basically you know the model is projecting a popular win by firearm eight points for Joe Biden that got up to eight and a half. Now eight point one. So it's tightened a tiny bit. Meanwhile time has run out the clock that does boost binds when probability Like you know, I mean things like if you're at eighty seven- that's pretty, I start,

hundred is. Very importantly, not a hundred right. Is that, like that much more room, ago right, so I unaware Biden we'd get. Maybe it said, ninety three or before something he beholds by election day. I mean it's still a pretty glacial pace right. You didn't kind of set point a day. If the poles are steady, there's been maybe a teeny tiny bit of tightening. Do you have a theory as to why the national polls, and then the national picture, as painted by the eight level. Poles are different. You have a little bit of a different mix of pollsters. The high quality poultry the drop in networks and academic pollsters, they kind of all tend to think a little bit like into the kind of pull the same states or pulled national rated at similar times right, post debate, Post covert, I notice we had like a lot of high point: pollsters pulling national pulse highquality like color poles and shall we say.

staples were a bit more mixed. There are more kind of Rasmussen and Trafalgar in public policy polling always they're, not her favorite, pollsters and when you saw the higher quality poles in the states. It did seem kind of more consistent with like a larger. Finally, It may be nine enough points or something. So it's part of the mixture pollsters right. It could be also like can be a little bit of hurting, especially the lower quality pulsars tend to want to come in pretty close to whenever the five thirty eight or milk, clear politics average was in a state of becomes probably more pronounced later in the year, so that can create some funny stuff, sometimes one of the things we dont give very many poles. Nine swing states. You and then, and I will get like a West Virginia polar Kentucky people or something that shows Joe Biden doing. surprisingly well, there are Kansas her North Dakota. Do by those poles. I'm not sure I mean it's sort of plausible right. I mean.

If you have a lot of kind of white, older rural voters, now you finally have a old white guide, a vote for it instead of Clinton or Obama right right, there might be. Our John Kerry was a white guy, but various dismisses kind of a liberal maybe by those who bear the states what game and it could be that binding is making it in these really read states that doesn't really matter a tipping point. States is another theory, but any attempts to even out you mentioned Harding. We often times end up on this topic somewhere. Round election day. Are we now on hurting alert? Are we seeing behaviour from pollsters that should make us think that some of them are hurting and I should say for listeners hurting as us, Finally, when pollsters publish results that converge on the conventional wisdom of what the outcome of the region will be seeking out a pull that represents the expected result before election days or that they were good boost their reputation but oftentimes. What actually happens is that, as pollsters heard pulls become less act

because they are not open to different twists and turns that the waste could take. So that's why we're on the look out for her, but anyone got any. I mean there are different types of hers. One paper pulled at herds as a pole, online or ivy. Are there some very good online poles nurse and must get online poles? I think some, online polls that are less good, impart, pretty strong assumptions into how they wait and massage their data, because you're not getting a random sample and semi theoretically do with a phone call. It seems like sometimes those decisions are influenced the little bit by what other pulsar saying, you're appalling averages IRA it's kind of one category undemocratic. Where is sometimes these cause. I partisan pole, there is a kind of big shadow area. Right of poles are not officially partisan. The way we define it, which is actually pretty strict definition, but in practice it Yahoo decisiveness. Poles are ready for the wind and the men

those they write or kind of framed in a particular ray and trying to influence. The averages are boost morale, those necessarily heard toward the exact average. five thirty eight, but their very aware of kind of their powers. In the landscape in a kind of seems like instead of a pole, beginning pulling inputs, I don't doubt that there actually calling people, but just extremely strong priors and your base. That is getting like somebody's opinion. The third haven't hurt. Her is something you see more in higher quality poles, which is that high prestige, poles, sometimes the last full they published the can- Nepal, they publish they know it's gonna be judged in. They may just be no more careful they haven't outlier. They would find some way not to publish it to apply it. Turn out model and get a little creator that test only at the end of the race, sober yeah, their mouth, who came to hers ice,

call Connor, who is a stat Bloomberg right on twitter quote my favorite new wild cat. Pulling theory is now that the five thirty average is seen as the Wall Street consensus. You have so much potential pollster are hurting that the real poles are the house. District poles were pulling averages, don't exist, There is a lot to unpack there, but we ve establish what the general meaning of hurting isn't and whether or not posters are doing it. Do you get the sense that the House district level polling is showing a different picture of the race than some of these? Other polls that we may be able to average is interesting, so first of all, maybe Sarah and I should assign a stewardess summit if I thirty eight about like actually rounding up all these poles and looking, which ones have pulled a presidential race but certainly anecdotally, amazement people unviable, concluded. It seems, like these polls show

very good results for Joe Biden, even once you account for the fact that a fair number of them are partisan polls. You know they might be consistent with the race. We're buying is winning by twelve or thirteen point something this house. District encounters theory. Is that because there's no five, thirty eight average, because you know frankly, no one's gonna give upholstered too much if they have trump losing a district by seven points and then four months later athletes crunch. The numbers are congressional rig. It turns out that he wanted by one point. One can remember: really right. So maybe it's a more honest, take on the race to interesting theory, I mean the national polls show a very big Biden leave right. The state policy monsieur slightly smaller by leads in the district also this convention, by its little we're working, a little bit, I think, pulsars, probably feel comfortable in sea ports with various kinds of single digit Biden leads. It seems, like you, don't actually seeks

like Erasmus. Is the world very many poles and actually show Trump ahead, even though, if thine is actually ahead by five point to the state, you are going to Bahrain chance alone get your occasional like Trump plus one returned plus to its use up. We don't see very many of those it also seems like you don't see. Very many poles was willing to publish double digit margins for by, even though, if he's plus five get occasional, plus ten plus eleven or something like a sort of thing, is like If a poster is showing Biden ahead anyway, there incentives might run pretty strongly against showing a larger margin. That makes sense, I don't know so, essentially we should be on post are hurting watch bike. We will really be able to say what is happening until we know the results, but keeping in mind that we are seeing differences between house districts, state level and national should make us

question- that there is something going on that is shaping these different results. Hurting could be one of those things is that maybe the lesson to take away here yeah. And what are the thing I think too, is like the sometimes the very last. poles because it hurting wind at being, accurate and like good poles, three or four days before? And it should be that way? But you know it's possible, I mean in twenty. Sixteen people forget it was a little bit of a tick back toward Hillary Clinton. In the final, like forty, eight hours ripe, trumpet got up to like thirty five percent and our forecasts meant shrunk to twenty nine percent I'm a forecast that different, but sometimes at last, little blip can reflect hurting may actually not directly put you in the right place, so

The other thing I've been paying attention to this week is turn out and it looks like five. Thirty eight has also updated its turn out. Projections upward revised upward our turnout projections because it looks like turnout in this election could be huge. So what led to that and what kind of turnout of we expecting at this point? I mean it seems like people on staff and external sharing, this type that number a little bit more, and I think we haven't really put frankly that much effort into it, but you're not number forecast, doesn't affect the margin or percentage. Is it just kind of side product, I'm exactly the ratio of turnout in different states effects are popular, wrote Esther. Because of the way that works for the next popular known as we have a forecast in each state Added up, you need to know what turnout isn't each state on a percentage basis by Texas Mix up, say seven per cent of turnout nationwide, nor did it awaited estimate of international popular vote is

and I should say before- we go any further- let's put some numbers to this and are Jackson is now that the total print in the presidential race is going to be a hundred. Fifty four million with an eighty percent tile range between a hundred forty four million and a hundred sixty five million in two thousand. Sixteen turnout was a hundred thirty seven million. Where was that before we made the revenge, how much did we revised upwards? So we had been interim one. Forty three: So we had another ten million. Actually, what most of that was is We have been using estimates of how much population had grown since twenty. Sixteen. That, apparently, were too conservative pressure. Michael Mcdonald as actual estimates of the voting eligible population, we kind of been indirectly trying estimate that through population growth show that there are more eligible voters in some states than we have committed by the way, the fact that in some states you have removed

Some degree of Phelan disenfranchisement can also increase the voter eligible population so using them, contemporary up to Esben how many people are actually eligible to vote, but also our model does look at poles, may I ask how enthusiastically are those are off the charge We also made an adjustment based on looking at voting laws and whether they been liberalized or made more conservative enough from your podcast with Amelia. You re in general. Actually, there are alot of laws that make voting easier this year. Some of them are permanent. Some of them are temporary because of covert, but you we expect that if there are more ways to vote than ever, which in many states there are that that would boost for now just a little bit that more people we'd like to cast about or wind up having away to cast a ballot. So Why not now in around two hundred and fifty four million, but you were saying that that doesn't actually change are projected national popular vote mark

Can you explain why that is? We were lying on pulsar, sporadic turn out and the poles have to wrestle with out they're gonna turn out. Obviously, but we don't like adjust star, margin up or down basin or turn out projection. Willowy. Which the Poles, as we average the poles just a minute, what turn out will be implicitly pulsation firstly say there are: we gonna predict seventy percent turnout of ammonium population. We have to kind of infer that, but our opinion. of what turn out will be our model of eternal, be doesn't effect. Whether Biden is up by nine points are eight point so whatever now in practice subjectively MIKE ass. Is that some pulsars you're being a little too conservative Luckily voter models, it may be that hurts

the crux of, though now over the past week, though we seem poles were like Democrats, game likely motor model. See no kind of sea may be praised him in the polls, an assumption that if there is differential turn out, it would help Democrats, it's not as the case or later in the cycle. In fact, one theory about binds growth in the polls could just be actually that, like Michelle comes from the fact that now your accounting for early voting and succumbing more apparent that Democrats will have a high turnout. Maybe Republicans will also they probably will, but you know how polls are wrestling with early voting. Is I think, one of the big important We ve actually gotten some specific questions about this. Why we're seeing significantly different results in some poles surrounding registered voters likely voters, but then also the high turnout scenarios and low turnout scenarios and one pull in particular that showed this was a Senate pull out of. I was so what is going on here. so that was a Monmouth with and the mammoth profound found. Actually, that Democrats would prefer a lower turn out because

but their voters have already voted on simplifying here and therefore you dont want more turn out cause the remaining people who have left a vote. I Republicans, but in a moment, is an interesting pulse, and I believe they use probabilistic waits for the likely voter model, which is absolutely the correct way to do it like. I don't understand why every pulsar doesn't do this. It's kind of idiotic not to have been a lot of posters. By doing that, I love you posters but uses problematic models is clearly like it's a. U can defend using his hard cut off. It doesn't make any sense at all can explain what a probabilistic turnout model is yeah that you take whenever inputs, whether it's you tell him, like a yard about some pollsters ask for how much you ve been in the past. They asked for knowledge of like where's, your voting location. I mean they call the times, have done good work on this with Sienna College, but like even if you kind of check all the boxes and here the highest repairs.

Play score you're still at a hundred percent to vote may be year. Ninety percent, or something and unity say: you're, not gonna, vote and don't at his out questions you. They might go. There's someone you twenty percent chance or something empirically. When you go back and motor via later and see whether someone bought it or not so well, for years they have like a hard cut off you're. Either hundred percent essentially or zero if you're above some pressure blossom threshold, which doesn't necessarily make a lot of sense? like what, if there is a long tail of low propensity white working class voters, each of whom have a thirty percent chance to vote individually, but collectively Three percent and will vote, and they maybe be tempt voters right earlier verse. What, if latina voters and lack voters who, having voted before your little sky, the goal right. We really have. This latino turn surgeon, maybe well won't, and maybe two thirds of people who use filter out

what we are voting by one third will and if they are really democratic vote in some states like him Arizona, then you can Lobal Democrats there. So there are questions of like if you have people who have already voted. Should you treat them as a hundred percent? and then people who is likely voter you should treat is something below. Hundred percent, I don't know if that's what mammoth is doing exactly but there's an argument for doing something like that. The flip side of emu will basically be like ok if you have democratic the voters have already voted in Luckily, borders will probably vote, but maybe well that could be a reason why why likely Poles, who underestimate how well Democrats would do given their votes, are locked in the flip side. I'd say that you're also now excluding some of these unlikely republican voters that will vote and maybe a kind of all balances out in the end. But there are various something you could make about how you tree a likely Motor Emerson already voted and how that affects year by Russia, Simples adieu, even ask whether you already voted or not at all,

It may be view and asked the question: you should keep up ulcer, be already voted, but the politics you out doesn't like the voters stream, because it even bother to ask you ve already voted already. You know so our various ways to create biases in the polls and its additional source of error. So, hopefully an hour or so fishing in the weeds of pulling methodology for our listeners, I learned something there. So let's move on and focus on some listener questions from here on out, but first, let's do into these wasn't our questions and, as we mentioned in the late night, podcast last night, getting down to the wires here, there's a week and a half left about in this campaign, and sisters, are gonna, be conducting their final pools relatively soon. So our first question comes from a visa and the pressure

How would the model react to a last minute? Surprise, story that is too late to be properly pulled. Yeah I mean that is part of the uncertainty. Come to the fact that you don't actually incorporate every last news. Development can be stopped that happens on What's your morning in theory. I keep in mind, though, that this is a case where the early voting does matter a lot, something like two slash three. The vote is expected to be in on election morning I mean they're. Already fifty seven million boats so something that happens like the people. Vote very early tend to be pretty big partisans who propaganda though the same way in a manner? Why? But you know a lot kind of more regular, even undecided voters may vote in the handful of as for the election, and so it's pretty hard for some last minute changes to have the full effect that it might have an election where eighty the people are voting on election day, and that is a pretty big change. But ya me luck

certainly, if there's some major event- and we don't know direction which affect it would have which can have to say. Okay. Well, you know the model kind of banks in some uncertainty, and this might be a case where not making enough uncertainty, because events that we know that happen, but we can't quantify the flip side of that, might be that if there is anti climactic minister, the rays then that might mean that the model has too much uncertainty current, assuming that this one added eight times right? This is crazy news development that we underprice right, but now and then You know if you have a very smooth finnish than it might make things less likely to have a polling you're. So if someone lately. We got a question that asks what does need mean when he says that the five, model is somewhat conservative. I think when we make a sum, about how much era there is an polls that we assume there's more error, then other people who build models do, for example,

they're kind of three ways in particular that we assume the final poles might be more wrong. Then other models assume some number one. Is that most the model is calibrated based on polling data from nineteen seventy two onward, because that's when we start to get somewhat robust polling, where there's both state and national polling and pulling throughout the year and not just at the end. Our estimate, though, of kind of how much the final polls might be off, also incorporate data. From back to nineteen thirty six, which is the first year Gallup, did final national polls. And galloped national poles in the thirties and fortys and a pretty checking track record, but that its accounted for in our estimate. How off the final poles might be. We calibrate that, based on data collection, dated thirty six and so like literally, the dewy versus Truman pulling error is a kind of where our model- s number one number two, because

We think that turnout is gonna, be harder to predict because this new mail and early voting, because we saw in the primary turnout was kind of all over the place relative expectations we also add some additional air for that, so we say, poles are gonna, be twenty percent more error prone than they would be, otherwise Finally, we have what we call fat tailed error, which is that, because the sample size of elections is an all that large, we use something called a student's t distribution instead of a normal distribution. Which affects the weight on the tales of the chart. It means that you know the model is reluctant to go above. Ninety eight percent, if somehow by know, if it's up a thirteen point, lead instead of ten points or something Hale's at unlike a normal distribution bell curve, Ricketts very, very thin. The tales are thicker it's hard or would you go to ninety eight? Ninety, nine. Ninety nine point, five and whatever else the go beyond that, and so those are. Writhings, that, like

you can argue, we don't have to do if the rest, the miles cover. Making semi to offer the wire weep. Sudden using nineteen. Thirty six data for this part of that right is Emily pertinent to how accurate poles are. Now we have a lot more poles. Maybe maybe not you know I mean, I think the fact helping is the correct practice and it gets kind of unambiguous and we ve talked about our fat tales before on the spot. Casa lesson or should be familiar with our fat tales, but this is a just What kind of turnout is a little bit more debatable. I think I mean I kind of hate the fact that, like I don't know what kind of how people perceive five hundred and thirty eight forecast, I kind of hate that like, if you can kind of take us to some like avatar of certainty. I mean by news ways ahead in national poles and there are eleven days to go in the campaign of the time we're taking this trumpet kind of out of money there? No debates, we have a raging endemic is getting worse. everything he tries kind to see. Mister not be very popular and

He has a twelve percent chance of that's pretty. That's pretty accurate user. Secondly, I mean, I think, that's why people are curious, how we got to that place. That seems relatively serve it. If you did mention that we have no more debates laughed, it's only been less than half a day since we published our last part cats, but did we actually at some pulling on how the candidates did in the debate. Whether or not Americans thought Trump provide more Yeah there were several pulls out that that binding better, although not by the margin from his first debate,. If he led the who wanted the debate poles. My fifteen points, then you might typically expect better translate till I go further. One point Biden gain in the top line. Poles I'm not sure you'll see that here, because by was so far ahead that like having him when the post debate who want the debate pulled with eighteen point, was ahead. The horse raised by ten point, nine, a very big difference. Our

poor with Ipsos showed. Although voters thought that Biden did better, there was essentially no change in the top line of people's voting intention there. So I dont think I expect a big change if it shifted another point toward Biden, I wouldn't be shocked if it should appoint Wardrop, I'm not. That can be because the debate, I think it might be because, like there seemed to be a team, had he of tightening going on already and maybe he's kind of show em it's hard to lose an election by ten points in the United States. Given how polar as it is today that just pretty hard to do so maybe it would be a little bit of tightening mirror, which I think is because the debate, but that we should say almost every talking about show based on pre, debate polling, so if you believe that, because the debate Biden has Now- they either remained steady. The poles or increase a point which I would probably what are the steady interpretation,

then, by the time we get posed to be pulling in five or six days from now, then probably above eighty percent? Eighty seven percent in our model right on your mind posed to make pulling as a top line pulling in the horse raised? Not right. You know days from now. People posters are gonna, be asking how the candidates did in the debate our next? Russian comes from Liane and she ass is their significantly more state pulling in the battleground states in twenty twenty than in twenty. Sixteen and, of course, We get a lot of questions along these lines. So let's answer that question, but also more broadly, want to ask what does the pulling a landscape look like right? now. Are we getting a lot of high quality poles? To what extent are they disagree and to what extent are we getting low quality poles and might be messing around with our averages. But let's answer that battleground stay question first and then brought it out think we will have enough polling- in the large, Jordi of swing states?

I'm a little worried about Nevada, which is always a state that a lot of pollsters avoid menacing, repulsive disagreed quite a bit. Actually in New Hampshire, sometimes gets a little bit ignore those who look? It's not that close this year, Nebraska second district won't get put very much, but like some of the kind of bigger higher populations way states, I would expect you gonna not be lacking for higher quality pulse in part. Because of you know a few pulling auditions pick obviously, that your time Sienna College doing all those poles in keys races, as is very able there are one of the best pulses around my met- is pretty prolific and good ABC News is doing pause in states this year, which they haven't done that much for several cycle. So our colleagues embassies, russian imposed enable poles to have so you're, not really lacking for high quality poles attempt to go in a streak of a week or two. where some states ignore irregular little light on Arizona Paul's. Recently, for example, I wouldn't mind

he's revisiting Minnesota, which got pulled a lot earlier cycle and in comparison to two thousand sixteen. This is a lot more of our right in these battleground states. Amelia some relevant states in twenty succeed where, like just no good pulses were pulling the race at the end of the race. I think we're not going to have major battlegrounds are neglected in that respect, You also do have a kind of fairly high volume of state poles. I think the hired. Come on of all my poles, like you, gotta? What not a pretty good there's a lower echelon where this two heard alight, and echelon alike, pulsars from like the UK in Canada in places like that that maybe good, maybe not, I don't- have a track record of polling here and then there are the shadow parties impulsively talk about of were right, like I think that almost seems like your father group is just trying to pull enough there, always in the river politics average. You know I mean, and they always show the same number. The

we show by plus two and Pennsylvania minus two emissions, always the same. It's very weird Why do we even include those kinds of poles in are average and in our forecast we are a rule driven society Galen and we have to have actual systematic rules that I don't have to make a whole bunch of subjective decisions. So what do we do this year, as we actually kind of the House effect adjustment in our model? which is that you look at where a pole lines up relative to other poles of theirs consistent, bypassing the model adjust work. The house, if I can just now sure, like a little bit more modest than it has been in some power here is another mostly is contained within one state of her father group as full Michigan six times. Ok, you know it's has it back? You can adjust that right if they didn't go in and pole Florida for the first time an example, Lord and what's the pole float if the first time it wouldn't carry over that much of the house of ECHO other states.

That was a change we made this year. We found that, actually it's a good change in terms the accuracy of them out, because what you don't want to do is ignore the local pulsars at only pull one state, but I do think it like. There are some pollsters. I kind of abuse it a little bit and kind of like are. I mean good faith, is a strange thing to level right I'm sure that whenever the Trafalgar group Guy's name is Robert, something I'm sure he believes trumps gonna win out doubt that but like it certainly seems like if we look their politics as like Five pulsing in state and what a mister Fogg her it just one fifth that deeds opinion. You know what I mean when they described there pulling process. Words like where we use six or seven different ways to collect data on is that I'm gonna get some were quilt, whether we re going their electorate, you know you are supposed to infer what you like. It looks like from theirs Did you get, at least in part? Show me how much of our?

average and therefore of our forecast is based on these, as you call them peoples. I looked at them, it's it's not that much goes when they don't get that much weight and there are a lot of good poles. It's probably five to ten percent and the swing state Secondly, movement more in the swing states and the national policies in Poland ass the races much if you did take the time to subjectively go through these different pause and say, like this garbage do you think it would make our average any better or just looking throughout history include everything is: what gets you the closest result. Let me put it like this. I am less concerned about our average having a pro by bias, because we do include these- the spanish pause, which this you're a pretty republican? The fact that you can predict the direction of pulling bias from year to year, meaning that way

here can favoured Democrats when you could be Republicans orkan favour what the same party two years in a row, but it's not predictable. That impart reflects like the marketplace for pollsters and the fact that Trump beat his poses one is exceeding, creates more of a market for their father groups. Mr Rasmussen reports and the insider advantages that We, the meaning results sooner, say well. What about when a sixteen other kind of mainstream? wrong as part of the market. Reacting to it, you don't, I mean wreck between sixteen. maybe my kind of refugees, that is, I think these full suck. However there If you pulls out there, door waiting. My education, that's probably a fairly big mistake and have electorates, then, are we to quantity It is probably more democratic and they should and those coalsack to see you have like balancing sulkiness on both sides. There. There I would rather kind of let our algorithm take care of that then try to manually kick out these poles.

and they could be right. I mean it could be anything the thing that keeps these error prone, both sides lie right. Is that whenever we we are for years. There is gonna, be a systematic, pulling error in a certain direction, so This was in can live off the fat of twenty sixteen or other years, where we're going to be their poles and people ignore the fact that their Generic Gallup Poll was off like ten points in twenty eighteen, so you know it's a marketplace. I put it like that and I trust that overall process four for how we wait and incorporate poles sound like a proper capitalist. Let's It is a rapid fire questions from listeners. Here's the first one from day one of fingers on the keyboard programming to create the final version of the forecasts model, how long did it take you? Let me it's hard to say has like also built like these new polling averages and come iterative process, but let me say like twelve weeks

of work where on average and you other things to right, we're on average undoing forty hours a week? So probably. hundred hours of work on the presidential model. This year's, like that probably one hundred on the mid term model. They didn't change that much. It was just kind of collecting all this data, which takes framework as a ton of data the morrow uses. but put in five hours on a motto word in simple things, are very similar to the past. We made some changes certainly in extra when I meet a lotta, it's a lot of work next question is there a reason that five thirty eight doesn't forecast gubernatorial elections, this cycle? yeah, because it's not a priority is generally not. The larger population saves that have good internal elections in presidential year. Apologies to North Carolina, Missouri in Indiana and so forth,

There is a bigger cost and you might think just to maintaining all these if remodels again, but the housemaid honest than the rational models, say where we literally made like two or three changes: it still a lot of work to update all the different sources of information that we need to debug. edged case races, echo wrong. That could take a long time a Georgia special send an election or something. So it's just a matter of prioritizing next question. Why the odds that are run off in the Georgia special election determines which party controls the Senate I haven't they specifically in there could be to run us injury to. Actually What's a car somewhere in the neighbourhood of a twelve to fourteen,

some chance that the Seneca come down to one seat if it's two seats, which it could be that it could be double that refugees as a twenty or twenty five percent, but you need to have a libertarian get at a certain share. The vote and the George Irregular election or us off or produce could get fifty percent. So it's baby. Let's call it a fifteen percent high family home This is a question that requires listeners to imagine the forecast when you go and work on the page. You see that they're all of these balls that get distributed across different probabilities, different electoral college outcomes and then they show maps of of different possible outcomes and oftentimes get screenshot and people share them with us, and we, how does this map happen? and really this is just a very random process- are not representative of what we actually simple happened, necessarily under a bite and win or a trump when. But we got a question

how are the maps in the ball swarm? I guess my colleague that distribution, a ball swarm chosen. We choose maps such that the overall number of times that a state is a certain color matches the forecast. So if binding sexes conditional winning the electoral college fiction a time there picture of a binding winning maps should have Texas, colored, blue or whatever. I think the way we implement that can sometimes lead to like making a bit more likely that some weird looking maps are chosen but the same time. You know. We deliberately think that we're looking things can happen if you cannot go back and look at how accurate are the forecast in the sample mean we're back testing, especially and say, set like aren't pull that often I mean the answer. Is you can get the occasional state where your forecast is

Ten points off so randomly by wind, South Dakota in a blow out or something where there been no poles, and I try to do the forecast rivalry stuff too much, but some people at a certain rival model have extremely range ideas about how certain you can be about. We're not weird events occurring. We have this like tiny sample size of data when you actually kind of test your model. You do come up with fairly weird errors and states now and men especially statement. very much so Oh you're, gonna have weird stuff happen. Sometimes It happens. Lesser mean, as we ve been getting closely watching than the error reduces. The kind of tat

scenarios are happy with offers the maps now more or less weird they were in August. But, like you, gonna have weird stuff happens. I so before we wrap up, I want to come back to a topic from last night. Actually we are talking about the debate. We got on the topic of the editorial section of a newspaper versus the newsroom. I asked you if you were drunk actually didn't, ask you if you were drunk because you thought that our listeners didn't know the difference. I asked you if you were drunk because of how quickly you progressed into three different topics over the course of and it says, there's something like that and then we started talking about whether or not our listeners know the difference between the Editorial Board and the newsroom. I did a clicker pull on Twitter. to that poor- and these are people who follow me and listen to this part cast presumably listen to this by cast eighty five percent of people said they know the difference. Fifteen percent said they.

You, are missing my point when you really interrupted me, but I wanted to come back and say that, regardless of that pole, we got several questions asking if we could go back to that topic and talk about the difference between an editorial board and a newsroom and why it is important in the context of that Wall Street Journal story. So I am signed for cutting you off and now has the opportunity to explain. I dont think people will. I didn't realize this before I went and work for the New York Times is that they are run by separate people thought who was messing with your time since Rather Bill Keller. I didn't know bill. Keller. Doesn't us- have domain over the Abed pay to the times so that their separate products, but like that there too bosses, basically, maybe the publisher, the Salzburg, and I guess, they're kind of overseeing things they do have some joint authority, but, in a lot of newsrooms the opinion section, and the newsroom air

firewalls and be probably compete with Another and may be kind of hate when others guts to some extent. You know. Certainly I mean you see a lot of because its republic now right, you see a lot of concerns from p. New York Times or more public now than they were before in one thousand seven hundred and eleven York Times Guild right about like stuff that appears on the OP ed page. That might seem too racist points of view or whatever that can create big conflict. I mean there are various backstory stories going around about that wall. Street story antics If I were listeners, what is the story that you're talking about in the Wall Street Journal at a trial pages? I mean it's kind of the confusing scandal, so The alleged story is that some emails based on a laptop in a Delaware computer repair store. Where Hunter Biden had been talking with various associates about deal involving China What are those email seem to be trying to get Joe Biden

involved, is it? Oh, you know the big guy, what a ruse call would get a ten percent cut didn't refer to Joe Biden by name. And then later on, when the people in that ring bubble in school now The whisky is alleging that he actually took. The additional step of meeting with Joe Biden after the election, One issue here is like, even if you gonna take everything at face value You know even Kimberly Strapless opinion writer journal. He was taking it seriously, it's reported it like Biden, says kind of capital and know Joe Biden so It's a little bit weird that it doesnt unit, page or by name that negative alike. But the Virgin seem here right is that this didn't pass muster with people who are actual reporters. You pass muster with people who published opinions both based on what we saw yesterday, where they also journal. Newsroom, published story with the headline Hunter Biden this is partner. Alleges father knew about venture

a man, the southerners former members, as you know, involvement and corporate both strudel showed no role for Joe Biden. That's banning that story down and their various backstory set you'll hear about when the Trump folks were excited and what kind of tipping authors gonna be big story, Wall Street Journal that they may be thought that would be a story covered by the newsroom and that the Austria, journals newsroom, would have a higher degree of prestige, then no offense than Europe Post, but a little higher on the totem pole as far as credibility with the mainstream media. If the Wall Street Journal had entertained this, not as a question of like coming to countries with outlook actually kind of covering the allegations apparent. Whenever they saw the newsroom was not satisfied, a kind of wound up getting shocked around to another set of where's that we're lower on the total, one of which is a Wall Street Journal OP, Ed page, which has lower editorial

standards. It's very strange to me by the way why newsrooms by newspapers, I should say not newsrooms way his papers have these two entities with totally different sets of standards. It doesn't make any sense I think that if I ran, I will never running your times to bolster unwashed imposed ray, I'm sure. There's no say when that happens. If I do ok page no more of a new page, there are many mini town and people that I love anything a section of these papers. I would make sure they get awesome jobs writing for the New York Times in the New York Times, newsroom or maybe we can catch in the autumn- the great but leaving your second atrocities unsigned opinion stories. Like endorsements like why you endorsing candidates like what are you doing and why are you kind of publishing Stuff Wall Street Journal? hey that your newsroom rejected. I mean Are you doing name? We not always agree, but this is definitely a place where we do. I feel like I've been saying this for ever. I totally agree that the opinion sections of newspapers

not exist. I think once upon a time when people who really didn't have any other platform are way to interact with the news are way to express an argument. right in newspapers and that's how they could get published and their ideas could be put out into the public square. Now we have, medium and social media and all kinds of different ways for people who want to show their opinions and arguments that are not that it and driven by evidence they can do that in other places. I think it frustrates Americans, in view of the press and the role of newspapers too much by putting it in the same place. and for me the difference here is really in a newsroom. You start with questions and you go where the evidence takes you in order to find answers, and it should be. You know, of course we all have our biases. That obviously the case, but A rigorous process of allowing evidence to help you paint a picture of what the world looks like for your audience. Oftentimes

pinion section is starting with answers and your job is more to convince the world that the way you see it is the correct way to see it and the evidence is not always. Quite as important and again, I will reiterate what I said: that there are very talented analysed to work in some of these opinion sections and if the work is rigorous enough, you should absolutely just put them in the regular news- Right and I'm not calling for soon, nineteen fifty style I mean we try to- we do have, Dylan. We dare we're talking about us and we have our own style is now a style that you know it's that dictating us for bananas. Ray is just a start all. The actors in other writers of high three have evolved. There are plenty of pieces at five. Thirty eight than are argumentative there are plenty of peace, is a kind of taking pointed of You married lots and lots of, I think, really sharp writing at five. Thirty, eight. What's that kind of the voice of from nowhere style, but they said, possible, as it can be, a blue line, sometimes you're, starting with the evidence. First,

and ass. We come to your conclusions and you have to demonstrate things improve things, and things are fact checked in error check and you not trying to just give it someone, except maybe sometime you saying. Okay here is a devil's. You know next week I'll probably write some devil's advocate case for why Trump could win. Anyway, with with his twelve percent, stripe. That will very fairly say that this is a double atomic I'm in clear, labeling, commander, alot right and were imperfect, just like all newsrooms, but we strive we can always, I mean it's a real struggle right. It's real struggle, obviously Nino, sometimes I'm too candid and social media, but is this true for it's a journalist right you're kind of expected. Do we and also there's a pay off like a dream. Twitters issued trapper journalist, I mean a lot, journalists build huge following send twitter drive traffic to their knees rooms? I have lots of other thoughts on twitter, but I think

we have already talked about the opinion section versus the newsroom as our one journalistic topic for the day, we will get to twitter some other time, maybe after the election is over, but I hope answered folks, questions about where we were going without conversation last night again, nay, I'm very sorry. I interrupted you and clear, it is clear that you would make an amazing DR time anyway. That is it for now. So thank you may think you do and I also say before we leave you that the five thirty eight store is up and running. You want your order to be received between November twenty third and November thirtieth. You should place that order by November. Fifth, it's on shipping schedule so, depending on when you order changes when your stuff comes all, that is to say that we ve got lots of new stuff on the five thirty store, totes sweatshirt, with new designs. Predicary featuring fighting the thoughts

wait. A second level are fibre flax or if I may, defects, oh, it's finally fox. I guess fighting the fox sounds more like a qualitative like the site, is a fox. You know, there's near the midcentury modern pulse, there's recalled Fox road and then is a copy number five, five bucks road, so maybe it's where he lives dies, probably ready and when we eventually create a feature film for fighting the fox and we pitcher to our Overlords ABC Abc and the Walt Disney company. That is exactly where fighting thoughts- on five thoughts road, but until then go check out all his stuff at five, thirty, eight dot com slashed store. My name is due. In truth, twenty child is in the virtual control room. Claire bitter Gary Curtis is on audio editing. You can get in touch by emailing in us at podcast, at five thirty dot com. You can also, of course, treated us with any.

Hence comments. If you're a fan of the show, leave us a rating or review in the apple pie, cast store or tells him went about us and remember to subscribe to five thirty on Youtube thanks for listening. And will cease.