Delegate math
Category politics
Looks like Barack has won Vermont, and Hillary has won Rhode Island and Ohio. Texas is still "too close to call", and looks to be a 50-50 split. Earlier this evening I stumbled upon an interesting page: the Slate Delegate Calculator. It allows you to plug in hypothetical poll results for each of the remaining contests (including today's). Based on the numbers reported so far today, if Hillary were to win Pennsylvania (the "largest" remaining contest) 65% - 35% and all other contests 60% - 40%, the two would arrive at the convention precisely tied in pledged delegates.
Trouble is, the last-5 poll average in PA gives her a 10.4% advantage - not 30% - and the last-5 poll average here in NC gives Obama an 8.2% advantage. Should those numbers hold, a 60-40 win by Clinton everywhere else leaves her behind by 62 delegates. In fact, 69-31 would still leave her behind by 2... so she'd actually have to win 70% - 30% everywhere else to make up for a weaker performance in PA/NC and arrive at the convention with a mere 4 delegate lead. Just to belabor the obvious, that's a 40% spread. Here's the advantage she's enjoyed in the contests she's won so far (assuming that CNN's numbers for today hold solid):
Conveniently enough, if we ignore the two she won when Edwards was still in the running, that leaves 12 results... and there are 12 left to take place. So... just for fun, let's ignore the polling in PA & NC and map out a scenario in which she wins all 12 of the remaining contests by the margins above... but map it to where she wins the states with the most delegates by the highest margins (i.e. wins PA by 42 and NC by 18, but only wins Wyoming and Guam by 2%). In the unlikely scenario where she repeats her historical margins in a way that nets her the most possible delegates, she ends the campaign down by 10 delegates. The reality is that the furthest ahead polling has ever shown her in PA was 33% all the way back in November; 24% in NC the same week. Those numbers have shifted roughly 22% in Obama's favor in PA, 32% in NC. In fact, the 5-poll average in PA gives her a 10.4% advantage, but here's her advantage in each of the last eight polls:
Any given poll can later prove to be an outlier, but taken together, those 8 polls outline a pretty clear trend. So if she does win any remaining contests by 42%, it's not going to be PA. Given that there's still 7 weeks left, were that trend to continue unabated, Obama would win by a sizable margin. While Hillary's 42 point spread in Arkansas was the only time she won by more than 20%, Obama's won by more than 20% in 16 different contests. Where he wins, he tends to win big.
So... I agree with Chuck Todd that MS, MT, NC, OR, SD, WY, Guam, and Puerto Rico feel like Barack territory, whereas IN, KY, PA and WV feel like Hillary territory (although the most recent polling available actually shows the reverse in OR, MT, and IN). Assuming that's precisely how it ends up, and it's a 10% spread each time, Barack ends up ahead by 131 delegates. It's still down to the "superdelegates" to choose a nominee, but if she stays in for three more months and has barely eroded the current delegate lead, an argument that the superdelegates should overturn the verdict of the pledged delegates would sound pretty hollow. If the PA trend continues and he actually wins that by 10% as well, then by June his delegate lead has actually grown by 23, so by staying in Hillary will have bought McCain 3 months to start his general election campaign unopposed, kept the Democrats divided, and still accomplished nothing. Even worse, she's now resorting to "vote for me or your children will die" advertisements in a desperate attempt to scare voters into supporting her (for the record, I was disappointed in how Barack chose to respond to that ad). I sincerely hope that she puts party and nation above ambition and bows out before it's too late.
Looks like Barack has won Vermont, and Hillary has won Rhode Island and Ohio. Texas is still "too close to call", and looks to be a 50-50 split. Earlier this evening I stumbled upon an interesting page: the Slate Delegate Calculator. It allows you to plug in hypothetical poll results for each of the remaining contests (including today's). Based on the numbers reported so far today, if Hillary were to win Pennsylvania (the "largest" remaining contest) 65% - 35% and all other contests 60% - 40%, the two would arrive at the convention precisely tied in pledged delegates.
Trouble is, the last-5 poll average in PA gives her a 10.4% advantage - not 30% - and the last-5 poll average here in NC gives Obama an 8.2% advantage. Should those numbers hold, a 60-40 win by Clinton everywhere else leaves her behind by 62 delegates. In fact, 69-31 would still leave her behind by 2... so she'd actually have to win 70% - 30% everywhere else to make up for a weaker performance in PA/NC and arrive at the convention with a mere 4 delegate lead. Just to belabor the obvious, that's a 40% spread. Here's the advantage she's enjoyed in the contests she's won so far (assuming that CNN's numbers for today hold solid):
- New Mexico +2
- Texas +2
- New Hampshire +3 (John Edwards @ 17%)
- Massachusetts +5
- Nevada +5 (John Edwards @ 4%)
- Arizona +9
- California +10
- New Jersey +10
- Tennessee +13
- Ohio +14
- Oklahoma +14
- New York +17
- Rhode Island +18
- Arkansas +42
Conveniently enough, if we ignore the two she won when Edwards was still in the running, that leaves 12 results... and there are 12 left to take place. So... just for fun, let's ignore the polling in PA & NC and map out a scenario in which she wins all 12 of the remaining contests by the margins above... but map it to where she wins the states with the most delegates by the highest margins (i.e. wins PA by 42 and NC by 18, but only wins Wyoming and Guam by 2%). In the unlikely scenario where she repeats her historical margins in a way that nets her the most possible delegates, she ends the campaign down by 10 delegates. The reality is that the furthest ahead polling has ever shown her in PA was 33% all the way back in November; 24% in NC the same week. Those numbers have shifted roughly 22% in Obama's favor in PA, 32% in NC. In fact, the 5-poll average in PA gives her a 10.4% advantage, but here's her advantage in each of the last eight polls:
- 10/31-11/5: +33
- 11/26-12/3: +28
- 1/8-14: +20
- 2/6-12: +16
- 2/9-17: +14
- 2/13-18: +8
- 2/21-25: +6
- 2/26: +4
Any given poll can later prove to be an outlier, but taken together, those 8 polls outline a pretty clear trend. So if she does win any remaining contests by 42%, it's not going to be PA. Given that there's still 7 weeks left, were that trend to continue unabated, Obama would win by a sizable margin. While Hillary's 42 point spread in Arkansas was the only time she won by more than 20%, Obama's won by more than 20% in 16 different contests. Where he wins, he tends to win big.
So... I agree with Chuck Todd that MS, MT, NC, OR, SD, WY, Guam, and Puerto Rico feel like Barack territory, whereas IN, KY, PA and WV feel like Hillary territory (although the most recent polling available actually shows the reverse in OR, MT, and IN). Assuming that's precisely how it ends up, and it's a 10% spread each time, Barack ends up ahead by 131 delegates. It's still down to the "superdelegates" to choose a nominee, but if she stays in for three more months and has barely eroded the current delegate lead, an argument that the superdelegates should overturn the verdict of the pledged delegates would sound pretty hollow. If the PA trend continues and he actually wins that by 10% as well, then by June his delegate lead has actually grown by 23, so by staying in Hillary will have bought McCain 3 months to start his general election campaign unopposed, kept the Democrats divided, and still accomplished nothing. Even worse, she's now resorting to "vote for me or your children will die" advertisements in a desperate attempt to scare voters into supporting her (for the record, I was disappointed in how Barack chose to respond to that ad). I sincerely hope that she puts party and nation above ambition and bows out before it's too late.
