Get down on Gore now

[Mikawl]

Hello All,

According to some crazy stat on MNF, Bush is supposed to win b/c Redskins lost on MNF.

Anyone see this crazy stat? I can't remember the specifics but something to do with Redskins on MNF.

If Redskins won then incumbent party wins presidential election.

If Redskins lost then incumbent party loses presidential election.

I think this stat was consistent for the last 15 elections! Talk about SPOOKY, could this be the reason why the line seemed to jump on Bush after Redskins lost? And can someone else please confirm that they saw this stat too?

[Jeff K]

Chilly... Nice analysis. You may have given us a great bet, because it looks like Gore will win Florida.

[Chilly]

Chesterrockwells,

Depends on the poll, the method of assigning likely voters versus registered voters, etc. Bush leads in most polls that I know of between 2-4 percent, and yes, the margin of error in the later polls is usually reduced, often 2%. But please do not let you own preferences interfere with the analysis, if this is what is going on. The popular vote in a national sample is irrelevent. Look at the states, and Gore is even or ahead in all of the ones that really matter, with his worst state being Ten, then PA and Wash.

Take a look at the electoral college map and tell me what states I incorrectly assigned. I was generous to both camps in some assumptions, which might prove to be false in the end, but you give me no information to suggest why my electoral college breakdown is incorrect. In fact, Gore likely wins FLA, which I did not yet assign to anyone, and I think he will also win PA, but I consider that fairly to be a tossup. Without PA or FLA or both, Bush can not win. Its that simple, unless Gore winds up losing states he is likely to win, including mICH, Min, and wisconsin.

Keep in mind that Bush's lead, let's say 3% in the national vote, will likely be cut in half if we assign the undecidedes remaining 50/50 to each side. This is fair, and what usually happens. I think Gore is likely to benefit more than Bush, however, because I expect some Naderites to defect, and theyfavor Gore 7-3. So, that's where I get my 1% bush pop vote from. Buyt the college is a done deal for Gore unless Bush wins FLA and/or PA, and if just PA, then Bush must add a state like Ten or Wash.

best of luck in whatever you do,

Chilly

[Rich]

Couple of things to think about

2 STATES that cancel in electoral votesw

Texas Bush up 35

New York Gore 15

mountain and western states Bush up 15-30 points in most California Gore up 5-10 because THe Republicans are spending money there and the Democrats are not.

The national polls in any statistical anylaisis are therefore skewed to Bush. turnout --- Republicans always turn out whether their targeted or not- Dems only in high numbers with a hughe get out the vote effort. Mighigan Auto workers for the first time have election day off- Biggest NAACP and Union effort ever. Bush will win popular vote but Gore is a slight favorite to win this election

[chestrockwells]

Chilly, first let me say that none of this is political for me. Secondly, I'm having a hard time understanding how you come up with the statement that assuming Bush is ahead by 3%, as you stated (which is looking like a very conservative estimate of his lead).........that "it is likely to be cut in half if the undecideds split 50/50" In half? How? What percentage are undecided according to your estimates?
Also if you read the other thread on this, I stated that I felt Bush would lose if he lost Florida. So I don't have my head in the sand, I just don't expect him to lose Florida. The reverse of that is that if he wins Florida it would take a minor miracle for Gore to win. I didn't comment on your state allocation simply because I've already gone through all that twice in this forum. Besides I'm not interested in going back and forth with you about specific states, we will know soon enough. To put in plainly, I strongly feel that this whole electoral thing will be a non-issue. This campaign has been devoid of excitement and all the press and newspapers continually talk about the one thing that seems to be in dispute.......the electoral college, almost conceding the fact that Bush will win the popular vote. Why? Cause there's nothing else to talk about and oh wouldn't it be so dramatic if all this happens, blah blah blah. No offense, but your scenarios remind me of a football handicapper who predicts a final score of say 24-21 in a given game. Now there's say 3 minutes left and the score is 24-0. Conceivably it could still end up 24-21.......but a 24-0 or 24-7 or 31-0 score is much more likely to happen. So you construct a logical and elaborate scenario to justify the possibility of 24-21 when in 19 out of 20 times it ain't gonna happen. I think we'll see the same thing happen on Tuesday.
Good luck to you as well. No hard feelings here.

[Chilly]

Chestrockwells,

(1) Bush has a lead of 2-4 % in all of the national polls that i rely on. I have no idea why this is difficult to accept.

(2) If he leads by the mean, 3%, the reason this will decline is as follows. Assume 50/50 split in vote choice among independents. Whenever you allocate additional numbers at 50/50, it numerically reduces the advantage. If 52 people liike bush and 48 like gore out of 100 people (52% to 48%), and we add 20 people to each side, its now 72 to 68, or 51.4% to 48.6%. undecidedes usually split this way, but in fact, I expect Gore to be slightly advantaged in the split, as some naderites will go his way as well.

(3) Sure, this is partly speculation by the media, but its also partly what I do. I'm telling you, based on the assumptions I make, which I don't pull out of thin air, Gore is likely to win the EC vote. Take it or leave it, I get frustrated when everyone feels they are an expert on this, when in fact, I AM an expert on this. This is my job. OF COURSE this doesn't mean I can't be wrong, but I will be right more often than not, and the way these states are more than likely to play out, I make a predication.

(4) the state polls all show gore dead even or ahead in the remaining states that matter.

(5) The reason the line on this outcome is off is twofold (1) people weigh to heavily the Bush lead in the national popular vote [it's not that big], and (2) they don't understand that the pop vote isn't what counts. That's why the media is talking about the electoral vote--people otherwise don't get it. The fact that the media in america are for profit means they entertain before they inform, and they try to be "objective" even when its ludicrous to do so. So, they don't tell you that the current bsuh advantage in EC votes doesn't mean its really 50/50: they imply bush will win, but gore might survive.

I could go on forever, which nobody wants, so this is my last post on the subject. The line on Gore is wrong.

Chilly

[chestrockwells]

Unbelievable! How did you ever get past math in college? Of course, if you have 100 people and you add a large number you change the percentages gap. But back to the real world. Let 100 people represent the entire voting pool.
Bush 47
Gore 44
Nader 5
Undecided 4
This totals 100.
Now you split the undecideds 50/50 and assume Nader doesn't get any of them and you are at Bush 47+2=49
Gore 44+2=46
Nader still at 5
The gap started at 3 and is still at 3. And please don't tell me that 20% of the people who will actually vote.....are still undecided! And this is my final post as well.

[david173]

Hey Chilly,

As a former political science grad student (and a quantitatively-oriented one at that) I may be more receptive to your approach to the election outcome than many, but I'd like to make a few of points.

First, I think your allocation of states is questionable in that you have clearly allocated some VERY marginal battleground states to Gore.

You didn't set out your allocation state-by-state, but it seems that to get Gore to 208 you must have already assigned him some hotly-contested states. My own breakdown is as follows:

Solid Gore (11): CA, CT, DC, HA, IL, MD, MA, NJ, NY RI, VT = 168

Solid Bush (24): AL, AK, AZ, CO, GA, ID, IN, KS, KY, LA, MS, MT, NE, NV, NC, ND, OH, OK, SC, SD, TX, UT, VA, WY = 209

Battlegrounds (16): AR (6), DEL (3) FL* (25), IOWA* (7), ME* (4), MI* (18), MN* (10), MO* (11), NM* (5), NH (4), OR* (7), PA* (23), WV* (5), TN (11) WASH (11) WI (11) = 161

(The 10 battleground states marked with asterisks are the ones you list as unallocated.)

You've got Bush at 215 versus my 209; if we give him Arkansas (and based on historical voting patterns he should win it) that gets Bush to 215, which is your figure.

But to get Gore to your figure of 208, we have to gve him the other five unallocated battleground states, which is a rather dubious assumption, at least at this point.

Looking just at New England, Bush has been polling quite well in New Hampshire -- it has actually been a stronger state for him in the polls than Maine. If Maine is unallocated, NH should be also. Delaware should go Democratic, but it has been quite close in the polls.

You also give Gore Wisconsin, Washington, and Tennessee. That's a remarkable assumption. For good reasons, all SHOULD be Democratic. Wisconsin and Washington, based on historical trends (i.e., Democratic in last three elections, including even Dukakis in 88), and Tennessee because it is Gore's home state. Yet, all are considered up for grabs this time. Washington seems to be leaning to Gore, but latest Zogby poll (see www.zogby.com) shows insignificant poll edge for Bush in Wisconsin.

Ditto Tennessee. Even though it's Gore's home state, Bush has consistently run ahead in the polls, and although empirically a candidate runs significantly stronger than you would otherwise expect (based on party ID, ideology, etc) in his home state, that edge for Gore is already captured in the polls. Just look at surrounding states, where Gore is running way behind Bush, such as Georgia (which is actually a more Democratic state than is Tennessee). Yet, you somehow concede the state of Tennessee to Gore, which is not justified by any of the state polls.

Turning to your unallocated states, you state that you think Gore will win Florida. That's also remarkable. Florida has only gone Democratic once since 1976. Clinton won it by about 5.5 points over Dole in '96, and lost it by about 1 point to Ford in '92. Obviously, it was a GOP blowout in the other years.

Look closer at the '92 and '96 Florida results. In '92, Clinton beat Dole by about 8 points nationally, but ran 2-3 points behind that in Florida, based on shares of the two-party vote. Ditto '92, where Clinton beat Bush by 4 points nationally, but ran 5 points behind that in Florida.

Taking a crude average, we could say that based on recent results we might expect the Democratic candidate to run slightly worse (3 points or so) in Florida vis-a-vis the GOP candidate than the two run nationally. Even you seem to concede that Bush will probably win the popular vote nationally, or at least that it will be very, very close. If so, Bush seems a favorite to win in Florida, and that's a big step toward the 270 electoral votes needed. In fact, that gets Bush to 240, 251 if you count Tennessee.

New Mexico is also a traditionally Republican state, and it borders Texas. Look through your old APSRs, JOPs, and AJPSs and you'll see that presidential candidates also get a slight boost in states that adjoin their home state. Advantage Bush for those 5 electoral votes, raising his total to 245 (256 if Tennessee).

Missouri might be a battleground, but is also a bellwether. Bush has consistently been running ahead in the polls both in Missouri and nationally. In fact, the Missouri polls have tracked the national polls remarkably. If Bush is going to win the national popular vote, and I think he is, he is going to win Missouri, too. Bush 256 (267).

Bush doesn't need much more at this point. In fact, under the "Bush wins Tennessee" scenario, he only needs one more state. Where will he get the electoral votes?

Nader has been running as well in Oregon as anywhere, and, as you say, he is drawing more heavily from Gore. Unlike its neighbor to the north, Oregon isn't a big union state, and Oregon environmentalists are as rabid as you will find. Advantage Bush??

West Virginia, traditionally Democratic, has also been an uphill battle for Gore. In the end, I think Gore may win it, but then again he must, and in any case Bush has a real shot. Ditto Iowa, a hotly contested state this time, and one that has much in common with Missouri in terms of demographics, although Iowa is generally more fertile ground for the Democrat.

Among the other battlegrounds, recent polls suggest Gore is in good shape in Michigan and Minnesota. Pennsylvania is more of a tossup, and Gore must have it to have any chance. Slight edge to Gore.

We haven't said much in this section about Maine, another of your unallocated states. Like Missouri, it is traditionally a bellwether, and in recent polls, Bush has been even here, and has actually run better in New Hampshire than in Maine. So here are two more tossup states that Gore must win, but where Bush has strong winning chances.

In short, if you are wrong about Florida going to Gore (and I would LOVE to make a little side wager with you on that), almost every state becomes a can't lose state for Gore. And there are some, such as New Mexico, that he most definitely should lose, and others, such as Tennessee and Missouri (and perhaps even Wisconsin, Maine, New Hampshire, and West Virginia), that he would lose base on recent polls.

As for the idea that the undecideds are going to break for Gore because of the booming economy, as I read Lewis-Beck, Fair, Campbell, etc., the booming economy should have long ago since decided this election for Gore, and clearly (at least to this point) that hasn't happened. To the extent you are relying on this bit of conventional political science wisdom, it is, in the end, an act of Faith on your part. Nothing wrong with Faith, but it isn't much to make a bet on.

As for the Nader supporters breaking for Gore, there was a lot of speculation in '92 at to how the Perot supporters were going to break. Perot was polling 13-15% prior to the election. When the votes were counted, Perot ended up with MORE than he was polling. His supporters didn't "break" at all.

Do you really think Nader supporters are MORE likely to break from their candidate than Perot voters? I sure don't think so.

More on the Nader effect: One of the problems with the models that politcal scientists have developed to predict shares of the two-party vote is that they have never had to deal with a situation where a third-party candidate clearly draws support away from one of the major candidate. Here, Nader voters would, overwhelmingly, be voting for the Democrat Al Gore but for Nader's candidacy. That really wasn't true with other notable third party candidates (i.e., Wallace, Anderson, Perot).

Here, Nader probably draws 3-1 -- maybe 4-1 -- from Gore. Give Gore an extra 2 or 3 percent across the board and he's a winner, both in the popular vote (probably) and in the electoral college (almost certainly). But he isn't going to have it, and there's no model out there that has or can account for this phenomenon. So I would tend to place a little less weight on the models (which I otherwise fully endorse) this time around.

Anyway, notwithstanding the above, I bow to your expertise, and I am voting Gore, but I think Bush has got to be considered the favorite. If I had to bet, I'd demand at least +230 on Gore (to win in the electoral college, that is).

I hope YOU are right, though. I really do.
Go Gore.

[Railbird]

Awesome post David.its nice to finally see someone post that doesnt eat hotdogs and chainsmoke.

[Rich]

David- your anylasis is very sound
and realisiticly Gore has to hit the mother of all 3 team parlays FloridA, Michigan and Pennsylvania, From what I hear assuming a decent Dem turnout which (WE) are really working are ass for we should win all 3 but in Gore falls short in either penn or Fla (I
think MICH is ok because of the Auto workers)
then he need to win almost all of the nader states to win and one southern state Tenn,or
Ark or NM which I also doubt.


Florida is the key and while it was a repub pres state it's been getting more democratic
every 4 years and their seniors are voting more and more Democratic in federal elections. Remember in the most republican year in the nation a Democrat managed to hold on in 1994 agaist J. Bush from what I'm
hearing Gore's got a slightly better than 50-50 shot.

[david173]

Thanks for the nice comments guys.

I agree that the key is Florida. The winner there will be our next president.

On Florida: The retirees are fired up over Social Security and ************ drugs; Lieberman is also an energizing factor for the Democrats. States do change over time, and I agree Florida seems to be becoming more Democratic. It looks like the Democrat Bill Nelson may beat a sitting GOP Congressman (Bill McCollum) for the Senate seat being vacated by Republican Connie Mack.

So it's quite possible that Gore could win in the Sunshine State. I would still bet against it, but there's ceryainly room for disagreement. Even the polls have been in conflict.

By the way, for those of you who want/need to have the first word on how the vote is actually going tomorrow, I suggest the Drudge Report (www.drudgrereport.com).

Matt Drudge vows to flout law and convention by publishing exit poll results on the web as they become available. (Remember Carter conceding the election in '80 long before the polls had closed?) Everyone in the media knows the outcome in all but the closest states long before the official election coverage begins, and now, if Drudge follows through, we will too.

Otherwise, it's an awesome site if you're a political junkie, with links to almost everything. Check it out.

And good luck in tonight's game!!

[Rich]

i know drudge but i won't totally trust
him not to lie about the exit polls to
help repub candidates. Anyway unless the
election or the key states aren't close exit polls are meaningless because they never call these states unless their blowout due to the embarassment of being wrong. Who knows thank god this is almost over already.

Just on Cnn for what it's worth Gore's internal polls show him down 1 nationally but up in the 3 team parlay or trifecta states Fla, Mich, Penn

[david173]

Thanks for the word on the Gore internal polls. But then I also the other day that Bush's internal polls show him ahead in Florida.

In the end, relying on leaked internal polls is like being at the track and getting tips from people who own or train horses: they always like their own horse to win. If you know 4 guys who have horses in a certain race, you can get "tips" on all four horses. Not much use.

It would really be news if someone said "Our internal polls show us down 5 and slipping," but of course they never say that.

I don't totally trust Drudge either, but it will be interesting to see what the exit polls say. I think what he will publish will be the actual results from the exit polls, state by state, rather than just a call for either Bush or Gore. So at least we can see the numbers, although races in certain states may indeed be too close to call even on the exit polls.

[Chilly]

David,

you brought me out of retirement. I don't have a long time to thoughtfully address the issues you raise, which are numerous, so I'll summarize my main points and hope this clarifies some things.

First, nobody but you has pointed to specific state allocations to dispute my predictions, which is welcomed. Here are my breakdowns for the states that matter:
(1) gore is leading in PA, FLA, and Mich, of which Mich is almost certain in my opinion, and fla right behind, alhough others put PA 2nd. The state polls and the likely turnout advanatge gore. Look--I'm not saying he wins these things, but at least 13 out 20 times he would, which is why I allocated to him these states. What eveidence points to Gore losing them? These are coin flips at worst, which hurts Gore, but this is worst case--a coin flip.

Lets say Bush has 235 and Gore 207, current projections that I think are credible enough to start with (reuters). Of the questionable states, this count gives Bush missouri, nevada, NH, and Ten, and gives Gore Mich, Min, and Wash. The rest are pretty solid. Assume any errors in these questionable allocations cancel out in the aggregate count (bush loses missouri at the same time Gore loses Wash). This leaves: ARK (6), DEL (3), FLA (25), Iowa (7), Maine (4), NM (5), ORE (7), Penn (23), WV (5), Wisconsin (11).

If I give EVERY one of these states to Bush EXCEPT Fla, PA and Wisc, Gore wins with 271. Nobody here has to believe that this is the actual outcome to appreciate the fact this is a freaking close contest, and $100 to win $200 on Gore is a great bet, and +160 (Wsex) is a good one. If gore does not win these states, he probably loses, however. But I have heard no evidence to date to suggest why Gore can't win them, and in fact a lot of evidence to suggest he is likely to win them (likely meaning 62 out of 100 times, which doesn't help in a single-shot contest).

(2) I rely on "conventional wisdom" as you note (a la lewis-beck) for several reasons, most of which is that it is more correct than not until it is no longer conventional wisdom, i.e. predictive. Current prediction models are breaking down this year, suggesting convention wisdom is flawed, which might make some of my assumptions more questionable. Granted. However, I don't care about its accuracy in the aggregate, only in the states, and looking at turnout and history in a partisan theory of voting I conclude that turnout will not hurt Gore (the Democrat) where it matters most because unions, "old-folks", and African-Americans are going to show up, and get out the vote in Fla, Pa, and possibly even help Gore win WV, ARK and OH. The reason the models break down is that they assume the incumbent (party's) candidate runs on the economy when it does well, but Gore has not linked his candidacy to the economic boom (fear of a clinton backlash, warranted or not) so voters don't make the connection either, explaining why Gore isn't already the winner which he should be. But Gore has strategically ran on the good times in the last weeks where he needs to win, which is why I think he is slightly advantaged among the late deciders. This assumtion is not even necessary to predicting a Gore win so long as he is not HURT by turnout and the poll leads he has in the Big 3 bear out.

(3) nader is not perot, and yes I think Nader supporters are (1) more likely to back off (2) and support Gore. I understand they are "committed", and fewer of them, but that does not mean Gore can't get another 1% in the states where Nader is a factor,a nd this might be enough. First, Perot took evenly from both candidates, despite the impression one got from the media. Second, Perot spent a lot more money and more people knew more about him than they do nader, reducing defection rates. Third, the race between DEM/REP candidates was not so close that perot backers felt they would give away the election to one side that was their 3rd choice, which wasn't really a possibility anyway given Perot's support drew evenly from clinton/bush) [in fact, Dole asked Perot in 96 to drop out, but that would not have done much]. Fourth, Nader supporters ARE gore backers 2nd. They are susceptible to seeing their worst case scenerio occur by backing their first, which wasn't the case with Perot supporters. So, I see 1% at least going to Gore at the end, which tracking polls are already identifying.

We will know soon enough, i guess. But the national polls mean nothing for the wager, and the electoral college votes can go either way, so why is Gore +160?

Chilly

[david173]

Hi Chilly,

Thanks for your fine and informative post.

I think you are dead on that the reason Gore hasn't salted this one away is that he has not been able to (or has chosen not to) run on the Clinton/Gore record of peace and record prosperity. You would hardly even know that Gore has been vice-president the last 8 years based on his campaign.

But for the Clinton factor, this election probably looks a lot like the '92 election in terms of an electoral map, I'd say the dynamics of the campaign would have been like the '88 election had Gore been able to run on the record. I hope you are correct that Gore has been able to get this message out of late without picking up Clinton baggage.

I think the polling evidence in Florida is, at best, equivocal, and in predicting that the state's 25 EVs (and with them, the election) will go to Bush, I merely rely on recent history to inform me as to which polls seem more likely to be correct. And there have indeed been recent polls showing Bush up in FL, albeit within the margin of error, although I certainly acknowledge (1) that its close; (2) that Gore has a realistic shot no matter what the "actual" underlying preferences of FL voters were as of the last poll showing Bush ahead; and (3) the polls, such as Zogby, showing Gore 3 or 4 points in FL may actually be correct.

To get down to the heart of the matter, as a betting proposition, +200 on Gore is just about reasonable even given my assessment of Florida, and certainly juicy given your scenario.

To back you up further -- and in the interest of giving potential "investors" in this election all available information --there is a new study reported at www.orvetti.com that strongly supports your view.

Apparently, two Columbia professors ran simulations of possible outcomes based on recent polls and concluded there is an 85% chance that Gore wins the electoral college. I'm not sure of the specifics of the methdology employed, which was not reported by Orvetti (and even less sure I would understand it anyway.) But, folks, check it out, because it does jibe with what Chilly is telling us. How about an 85% chance of winning a bet at +200?? Not bad.

And, keep in mind, Chilly's got the Ph.D. -- I had to drop out and go to law school to make a living. So, if you're in doubt, take Gore, Chilly, Columbia, and the +200.

Good luck to all, and may the candidate who knows that Social Security is a federal program win.