Okay, now that you got that out of the way.

How about the major controversy you have not addressed at all.

I don't even need to mention it,  There are 5 threads in the last 24 hours, one started by an Admin of your site.

We are waiting for a statement or action from you.

Read this: http://www.mydd.com/story/2008/5/15/1427 30/254

I hear you Jerome, but as I mentioned in my diary to which you are responding, it would be very helpful to have more information.

I believe it would contribute to the dialogue and analysis here.  I would say more now but I have to run off to a political meeting.

The current map gives Illinois to McCain.

I suggest ignoring the whole thing as a safe methodology.

Then change AR to blue for Hillary, as even the latest Rasmussen reports poll shows her winning there.

Plenty of people are using the maps as biblical proof of various candidates' support.

But, if the latest poll is tied, then the result remains the same as the previous latest poll.

Why not use the "uncalled" option if the current poll is tied or within the margin of error?

Using the previous poll gives the state to someone who has declined (from leading to tied) since the previous polling point.

How dare you use polling that is not favorable to The Brokered One!

Coloring a state red or blue based on a 1 point lead for a candidate in one poll is patently ridiculous.

it seems to miss polls, like today there was a poll done by Ras. where Clinton is beating McCain her Iowa results got put up but not this one...

offtopic but - Anybody else see Kevin James on Hardball making a complete fool of himself? High comedy.

for clearing this up.

1) When you first log onto mydd, it populates the two maps with the most recent single poll for each state.

2) If you then click on the map and change it (for example, you don't believe that Hillary would lose WA to McCain), the numbers update to your settings - now it becomes like an EV calculator

3) The next time you log on, or refresh the page even, the counters go back to their poll-generated state.

looks like the supers are all moving to Obama

Maybe They know something you don't?

Well, one change must be made and that is Arkansas.

Rasmussen has a new poll out today showing McCain beating Obama and Clinton beating McCain.

http://www.rasmussenreports.com/public_c ontent/politics/election_20082/2008_pres idential_election/arkansas/election_2008 _arkansas_presidential_election

go look and see:

http://www.fivethirtyeight.com/

One poll came out showing Obama and Clinton winning Indiana, and that instantly discounts every other poll showing them way behind.  But because hardly anyone bothers to poll Indiana, since it is safely red, that outlier remains the most "recent" poll to date.  You got to admit the fact that Indiana being blue, while states like Wisconsin and Michigan are red, just makes the whole counter inaccurate, misleading, and ultimately useless.

Remember when we had Obama losing New York, but winning North Dakota?

I think the moral of this is, like the delegate counter, either do it properly or not at all.

I'm assuming good intentions here (probably naively) but many of the more acrimonious discussions here revolve around these tendentious metrics on the sidebars.

I'm sure it helps to generate posts and site traffic. But in a purportedly reality based community (and unlike recent front pages I don't make this comparison likely) it reminds me of the ethos of the current US administration.

Thanks for the explanation Jerome.

I don't think there is any good way to show where the head-to-head race is now.  Anyone who spends time with these polls, for instance, will notice that the key reason for Obama's weakness in states like OH, PA, and FL, to the extent it exists, is this: fully half of all Clinton supporters are telling pollsters right now they'll vote for McCain, stay home, or don't know.  

Who thinks that will be true in November?  And, who knows, maybe that's how this will pan out.  But it's impossible to predict whether these numbers will hold up (and there's probably an asymetry here, approx. one-quarter of Obama supporters are telling pollsters they wouldn't vote for Clinton, but they're doing this in the context of expecting that Obama will win the race, eg. if Clinton actually won we don't know whether these numbers would go up, down, or stay the same).

If Clinton and Obama didn't have to contend with this problem of party unity, btw, what the polls actually suggest is that neither would have any problem winning Ohio, Pennsylvania, and Florida.

Another reason I suspect Obama's numbers are a bit low: pollsters generally norm their results to what occurred during the last election.  If Obama is the nominee, I suspect there will be greater turnout among two groups: African-Americans and voters under the age of 40.  Not expecting that older voters, white voters, etc., will turn out in greater numbers, because, well, their rate of voter participation in previous years was already pretty high.

And in fairness, if Clinton is the nominee another big question would be whether she could capture the votes of women who are indies and even Republicans (or, conversely, whether men who are currently slotted as "don't know" would break for McCain in a big way).

There are plenty of "x factors" here (unlike in 2004, when Bush and Kerry fought for a small percentage of undecided voters throughout most of the year).

My hunch, btw, is that the lay of the land is so favorable to Dems this year the only way they'll lose is if they effectively beat themselves (ie. a large fraction of core Dems stay home).  

And what exactly is the point of making a low number electoral state so dang teeny-tiny, and why do you need to graph each state with the same number of little squares as electoral votes? Makes it doubly hard to view.

Can't you just have a "standard" looking map without all the graph lines?

My two cents.

I'm back from my political meeting now and have to respond on the map.

So we're told that this is how they did it in 2004.  Well, a whole lot has changed in internet land since 2004. So what was groundbreaking then is absolutely behind the times, like a crank auto would be at a motor show.

The problem is that when you look at the map, no one knows what polls were used, if they were ones that have a good track record, what the sample size was, when the polling was done -- basically NOTHING.

Over 90% wants this change.  Some who have spoken out against it imply that I've asked for a biased presentation. I have not.  

And if some can't distinguish between their own political views and a hard-headed, clear-eyed analysis of data, then, well, I guess that explains the very strange arguments that get made sometimes.