RCV Simulation Model
Simulation Model Overview
In addition to the basic RCV tabulation above, we ran an RCV simulation model to explore other possible outcomes and account for uncertainty in the race. Our RCV simulation model is based on 10,000 simulated RCV elections with ballots drawn from the weighted poll responses.
Our model confirms Eric Adams’ position as a leader, with Adams almost always earning a spot in the final round. Our model did not identify any situations in which a candidate other than Adams, Wiley, or Garcia wins the election.
Most likely final-round pairing: Adams vs Wiley
Our simulations indicate Adams and Wiley as the most likely pair to make it to the final round, occurring in 59% of our simulations. In those scenarios, Adams typically wins with a median margin of victory of 56%-44%, matching the scenario in the basic RCV tally above.
In an Adams-Wiley final round, Wiley will gain the most support from voters who initially supported Morales and Garcia. Adams will gain the most support from voters whose first choice was Yang or McGuire.
|
First Choice |
% of votes to Eric Adams |
% of votes to Maya Wiley |
% of ballots ranking neither |
|
Kathryn Garcia |
35% |
47% |
18% |
|
Andrew Yang |
46% |
28% |
26% |
|
Dianne Morales |
11% |
66% |
24% |
|
Ray McGuire |
50% |
35% |
15% |
|
Scott Stringer |
40% |
35% |
25% |
|
Shaun Donovan |
32% |
31% |
37% |
|
other |
17% |
11% |
72% |
Another possible final-round pairing: Adams vs Garcia
The next most likely result is for Garcia to advance to the final round instead of Wiley, which occurred 41% of the time in our simulations. When Adams faces Garcia in the final round, Adams’ median margin of victory is 54%-46%.
In an Adams-Garcia final round, Garcia will gain the most votes from voters whose first choice was Morales, Wiley, and Stringer. Adams will gain the most votes from supporters of Yang and McGuire.
|
First Choice |
% of votes to Eric Adams |
% of votes to Kathryn Garcia |
% of ballots ranking neither |
|
Maya Wiley |
26% |
61% |
14% |
|
Andrew Yang |
49% |
34% |
17% |
|
Dianne Morales |
17% |
63% |
20% |
|
Ray McGuire |
62% |
29% |
9% |
|
Scott Stringer |
39% |
54% |
8% |
|
Shaun Donovan |
42% |
30% |
28% |
|
other |
20% |
10% |
70% |
RCV simulation methodology
Our RCV Simulation is based on 10,000 bootstrapped RCV tabulations created from weighted poll responses. Each simulated RCV tabulation samples from the distribution of weighted poll responses and tabulates the ranked choice voting tally based on that sample.
