Delegate Math

The primaries have now entered the math phase. Romney is chasing the magic number of 1144, competing hard in every state to maximize his delegates. Gingrich now acknowledges that it is almost mathematically impossible for him to reach 1144, and admits his strategy has switched to denying Romney the magic number.  Santorum, at least publicly, says his goal is still to win the delegate race, but he has to nearly run the table to do so.

It has become a two man race.  Romney needs 649 to reach the line.  Santorum needs 892.  Whose strategy is still viable, according to the numbers?  Is Santorum playing to win, or is he secretly playing the spoiler role too?

The number of delegates still up for grabs is 1289.   However, just looking at the national number is useless, because of the mix of delegate allocation methods, so you have to drill down to the individual states.  So far, this race has hewed to demographic preferences fairly closely.  Romney wins urban/suburban, northeastern, and western voters.  Santorum wins rural, southern, and midwestern voters.  Because of that, it is possible to make reasonable predictions for the remaining allocation of delegates, and thus to assess whether each strategy is viable.

Below is my rough guess on delegate allocation for the remaining primaries.  I did not want to bias the outcome with too much subjective judgement, so I kept the methodology simple, as follows:

  • If a state is WTA and leans fairly strongly towards one candidate, I gave that candidate all the delegates
  • If a state is proportional, mixed, or direct, and leans strongly towards one candidate, I gave that candidate 2/3 of the delegates and the opponent got 1/3
  • If a state is a toss-up, I gave each of the candidates half of the delegates
State Delegates Type Favored Santorum Romney
MO 52 nonbinding Santorum
PR 23 WTA ? 12 11
IL 69 Direct Romney 23 46
LA 46 Prop Santorum 31 15
WI 42 WTA* ? 21 21
MD 37 WTA* Romney 13 24
DC 19 WTA Romney 19
NY 95 Prop Romney 32 63
PA 72 Direct Santorum 48 24
CT 28 Mixed Romney 10 18
RI 19 Prop Romney 7 12
DE 17 WTA Romney 17
NC 55 Prop Santorum 37 18
IN 46 Mixed ? 23 23
WV 31 Direct Santorum 21 10
NE 35 nonbinding Santorum
OR 28 Prop ? 14 14
KY 45 Prop Santorum 30 15
AR 36 Prop Santorum 24 12
TX 155 Prop Santorum 100 50
CA 172 WTA* Romney 51 121
NJ 50 WTA Romney 50
SD 28 Prop Santorum 18 9
MT 26 nonbinding Romney
NM 23 Prop ? 12 11
UT 40 WTA Romney 40
Total 1289 527 649

Based on this data, I think it will be a squeaker. (Edit: data changed to reflect NJ as WTA, making it even more of a squeaker)  Romney will finish with 1120 1138 delegates, just 22 6 (!) votes short of the magic number.  However, don’t forget there are 126 “super-delegates” who can vote for anyone, and those people are overwhelmingly from The Establishment.

My estimate is preliminary and could be improved in several ways. First, perhaps I did not make enough distinction between the true proportional and the WTA/mixed primaries. Second, I did not allocate the delegates from nonbinding state caucuses because I don’t really understand the factors at play in who they eventually go to. Third, I did not look closely at polling in each state, just a general preference based on demographics, so I probably got some of the the favored states wrong. Fourth, some states lean stronger than others, so using a 1/3-2/3 allocation for each state may have been overly crude.  Frankly, I just didn’t have the time to spend doing an exhaustive analysis of each state.

So if anyone out there wants to improve on this estimate, please give it a shot.

Also, just so there are no suspicions of hidden agendas, I am a Romney supporter.  Where I was unsure on a state, I tried to lean towards Santorum, to cancel out my own bias.  But certainly my conclusions should be closely scrutinized by those favoring Santorum.