It's hard to get perspective sometimes on the differences between Clinton and Obama in the delegate counts. For me at least, numbers and text just don't do it. I'm more of a visual person. Here are some graphs visualizing the delegate lead Obama has. I made some similar graphs last night, but got suggestions on how to improve them. Thanks to all who gave input.
The first graph shows Obama's net delegate lead from day 1, not counting FL&MI (see later for those). The y-value of each actual point represents the net Obama lead after 1 full election day. For reference, a few of the points have an orange label that shows which states went that day (not all points are labeled since some are too small to distinguish). The distance along the x-axis shows what percentage of the pledged delegates had been awarded at each voting day (and also how many are left). Last night I was listing things evenly by date, but I think it gives a better picture this way of how far along we are in total delegates awarded.
The black line is what has already transpired using the data on Obama's results page. His data includes all delegates that should be awarded, as opposed to some sites that don't include delegates that are not officially awarded. Yes a few delegates are projections towards state conventions and what-not, but the margin of error is miniscule.
The blue line projects the delegates into the future using the projections on Obama's leaked spreadsheet after super Tuesday. The spreadsheet has been pretty darn good so far.
I like seeing a range of results, so there are 2 other scenarios shown. The red line shows what happens if Obama has a very bad streak and gets the following:
O-50/C-50 in MS, NC, Guam, MT, and SD
O-45/C-55 in IN and OR
O-40/C-60 in PA, WV, KY, and PR
These are very unlikely scenarios (especially for delegate splits), but it shows a worst case. All of these splits are worse than the polls would indicate for Obama (where polls exist at least).
The green line shows what happens if he keeps beating the projections on the spreadsheet like he has been. I calculated that on average he has done about 3 delegates better per contest than the spreadsheet, so I added that to each result (except Guam). Yeehaw for a great ground game!
Observations:
- We are VERY far along in the selection of delegates. There is little time and few delegates left to shift the race.
- Obama is virtually guaranteed to win the PD race not counting MI&FL.
- Although the narrative is that caucuses are responsible for his lead, the graph shows that his lead is from big days in WA, LA, MD, and VA. 3 out of 4 of those are primaries. While it might have been caucuses that kept him tied with Clinton's big state wins, it is these 4 states that have given him the lead.
Popular Vote
A lot of Clinton supporters are starting to point to the popular vote totals as being more representative. So here's the popular vote graph. Here's how I got the totals:
- Popular votes are from RealClearPolitics with the exception of IA, NV, WA, and ME. The exact caucus attendence for these states hasn't been released yet, so I shamelessly stole caucus-goer estimates from Max Fletcher at OpenLeft for these 4 states. These are estimates of actual people who attended the caucuses instead of a simplified delegate count.
- Note that ALL popular votes I used are just simple attendees or votes. I have not projected caucus results onto state populations or anything fancy like that. I just wanted to start with what the media and Clinton campaign will use.
- Popular vote totals for the projected contests were also taken from Max Fletcher's diary. He estimated future turnout of each state individually based on similar demographic contests. I can't imagine the vote totals would be so different from estimates that it would influence the results. We already know attendence is high.
- The x-axis again represents each contest as progress in terms of how many votes have been/estimated to be cast.
I did similar projections as before. The blue line is from the spreadsheet, the red line is from Clinton winning the popular votes using splits that are listed above, and the green line is from the spreadsheet plus the average amount by which the spreadsheet underestimated his wins. In the last 8 primaries the spreadsheet underestimated by an average of 8%, but last Tuesday the spreadsheet only underestimated by an average of 1% (yes, he did better on Tuesday then he predicted). I settled on a 4% undershoot margin to balance his overall trend with Clinton's latest momentum. I didn't include his winning margin in caucuses in the estimate since the remaining contests are all primaries (might as well compare apples to apples).
Observations:
- The popular vote is also nearing the end, and Obama is very likely to win it as well.
- The Potomac primary is even more prominent in the popular vote. Practically all his lead comes from that one day. Take out that one contest and it's close to even. Again... he's winnng because of the Potomac PRIMARY. ;-)
- His lead today is about 700,000. This is a little bigger than media reports due to the inclusion of actual numbers from WA, IA, ME, and NV.
Michigan and Florida
So what happens if you add MI & FL in? I decided to look at the worst case scenario for Obama: both delegations seated as is. That means in MI, Obama 0 votes (0 delegates), Clinton 328151 (73). In FL, Obama 569041 (67), Clinton 857208 (105). Now you might complain that this is harsh not giving Obama ANY votes in MI, but we might as well start at the worst. Jerome on MyDD keeps Clinton in the lead on his counter by this method, so others will as well. Some of the undeclared delegates would obviously be Obama supporters, but some might be Edwards. Clinton will argue that way.
...and the popular vote
Observations:
- Obama still likely wins the pledged delegate race, and has a better than 50/50 chance of winning the popular vote. Only the red line drops below zero, and remember that line is a really horrible set of results for him. If he does this bad, something else is wrong.
- Given that you can mentally add in some missing Michigan votes and delegates, this doesn't seem like a bad result. MI delegates would never be included this way unless Obama has already been declared the winner. A do-over would likely cause Obama to at least net the same amount of delegates from FL he already has and greatly increase his haul in MI. I'm not quite so sure why they haven't been more supportive of new races. I guess if you have the race sewn up one way, you'll not likely to want to change the rules.
- Depending on what happens in Mississippi tomorrow, he can go a long way to shifting the red projection over zero (i.e. a popular vote lead) on the popular vote count. The red projection has him winning only 50% of the vote there.
- His still has a lead today of about 100,000, even including no votes in Michigan.
As I mentioned last night, my impression of the race is one in which a football analogy is appropriate. They were close to even at halftime. Obama scored big with a couple of touchdowns to open the 3rd quarter. The rest of the time they've been trading fieldgoals. At the beginning of the 4th quarter, Clinton had a couple of spectacular long catches, but couldn't score in the red zone and only got a couple of field goals. We're down to about 10 minutes to play, and she's still way behind. You don't give the game to the last team to score or the team who had the coolest play.
None of these graphs include superdelegates, but it's hard to map supers in time like it's being done here. I'll think about it some more, but basically the super difference can just be tacked on to the end. These are the kinds of graphs Obama will be using to convince supers (and of course he will use other things to convince them as well).
I hope this helps people "see" the numbers of the race in a better way.