Sunday, November 2, 2008

Election Cheat Sheet: Final version?

Ok, here's the newest version that has a bunch of improvements, and it's much easier to use.

Just type the winner's name in the 'Winner' column, and everything else happens automatically. You can use either 'Obama' or 'obama', 'McCain' or 'mccain', but anything else will not work with the calculations. After you've entered each name once or twice, the Autocomplete should take over so you will only need to enter 'o' or 'm'.

NE and ME will now work like all the others, however you can still split their votes if needed. Just enter the votes as needed. Note, this will overwrite the formula that's there now, so once you enter data in those fields, you can't go back to the autocalculations. No big deal, just something to remember.

This might be the last version, unless there is more feedback. I'll try to update the polling data at least once more, but I wouldn't expect any additional major changes.

[Excel File]

UPDATE: It looks like some of the formatting is lost if you save this a an Excel file, then open it again in OpenOffice (which I use). If you're using OpenOffice, you'll probably have better luck with this [ODS file]. The differences are minor, and don't affect the functionality, just the formatting of the charts. Excel user might want to try both versions to see which works best for them.

UPDATE 2: I've made several tweaks to the spreadsheet, so if you downloaded it earlier (say before 9:30 PST Sunday), you'll probably want to download a new copy, or just wait and get the truly final version with updated polling info probably tomorrow evening. There's also a PDF version available here if you won't be near your computer during the election.

3 comments:

Blues Tea-Cha said...

It keeps getting better! The final bar is nice.

Creating an automatically updated map is a great idea. There are some tutorials about it on the web. It looks very advanced, but at least you would be motivated to learn how. But how about just a blank state map from wikicommons that people could color in themselves on the paper version or when they print it later?

You inserted some data for demonstration purposes and to get it started. I also wanted to feed some information to it to get the pie started. But instead of putting in the states that you did, which have a fair probability of needing to be revised, how about putting in states which have a 1:100 or 1:1000 chance of changing? You could do this for the next 100 years and not get them wrong. This will also allow people to open the champagne bottles without waiting for California to be called or something like that. (They might never finish filling in the bottom cells, not legibly anyway…)

States which have both a 100% probability of voting for Obama according to 538 and a 100.0% probability according to election-projection.net:
CA NY IL MI NJ MA WA MD OR IA CT RI ME HI VT DE DC

States which have both a 100% probability of voting for McCain according to 538 and a 100.0% probability according to election-projection.net:
AK WY ID UT OK AL TN

The probabilities reported on the 2 sites were almost the same, except Wisconsin has a 0.6% probability of voting for McCain at e-p.c and 538 splits Nebraska into three probabilities.

Of course, users could turn on these no-brainer states themselves, but the "Matrix Preloaded" options could just as easily be moved back into limbo if the user doesn't believe the multiple 100% probability forecasts. (Maybe it's not worth a new version number just for preloading data, but if you make some other changes, it might be. :•)

MaxBots said...

Depending on when you downloaded the file, the data may have been just test data that was entered to confirm that it was working as expected, but that I neglected to delete. In the current, slightly tweaked, version there are a couple of states filled out to get people started, but it's a pretty safe bet that those states will end up going to McCain or Obama as the case may be. The benefit of pre-filling a few states is that it makes the autocomplete work properly right from the start. If someone does need to change one of the states, it's easy to do... Just click in the square and start typing.

Blues Tea-Cha said...

There are 17 states that have a 100.0% probability of going to Obama, and there are 7 states which have a 100.0% probability of going to McCain. (at both election-projection.net and 538)

Pre-loading your chart with the 100.0 percenters, (and excluding the mere 99.7 percenters like Wisconsin) the race begins with a big head start for Obama, 214 vs 42.

He could round up the additional 56 EVs he needs after 7:00 if he adds NH (4), VA (13), FL (27), IN (15), GA (15). If he misses a few of those, It could be over at 7:30 with Ohio or surely(?) by 8:00 when PA, NC, and MO check in.

This is very conservative in that I am considering States like MN and WI to still be up in the air although they both have a 99.7% chance of going to Obama. NH is 99.1, CO 97.9. If you assume he will get 33 EVs from those four: NH+MN+WI+CO, you just need to get 23. FL or VA+IN/GA would be enough. Heck, if you assume a 93.6% chance or better is enough, the election is already over with PA tipping the balance.

Earlier, I was just saying try pre-loading those which are set in concrete already and see if you don't think it is an improvement to just look at "swing" states, loosely defined here as any state having a less than 100.0% probability of being stable. Neurotically-worried Democrats can also rest easy when they see how close Obama already is, so it has therapeutic value, too.

Thanks for that. I hope I can add a few numbers to it before I leave for work around 7PM Eastern and know the winner, otherwise I'll catch up online around 11PM EST to see how the final map and EV count is shaping up.