(Updated 15/05/2026)
The New Zealand Herald recently released an NZ election model they commissioned from a New Zealand research firm. It produces a very different result from ours, giving the coalition roughly an 88% chance of winning. Why is the difference so profound?
New Zealand Herald Model
- 4000 simulations
- Does not model electorates
- Does not account for overhangs
- ???
- Does not have system for minor parties
- 88% chance of a coalition win
Our Model (Quasimodo)
- 50,000 simulations
- Models electorates
- Accounts for overhangs
- Uses sophisticated machine learning techniques
- Has a system to account for lack of minor polling
- 54% chance of a coalition win

When comparing Quasimodo with the New Zealand Herald model, you have got to feel sorry for the New Zealand Herald model, as ours does a lot more, but mind you, Kiwi Polls has been working on Quasimodo since 2023. What is the primary difference between the model: Seat Allocation. Quasimodo, famous for having predicted all this, predicts that the coalition will win 61-62 seats right now, where the inferior *cough* we mean Herald model predicts 63 seats for coaliton. Just a small difference like that makes a huge statistical difference in the odds. We cannot say for sure that want truly accounts for the difference, as we haven't seen the source code for Motu Research's model. The company that made New Zealand Herald's model is reputable and going forward, we assume it'll get better. But honestly? For an election polling this close? 88% seems to much to us. Time will tell. Just like Quasimodo, their model when backdated to the 2017/2020/2023, gets the result right, but they don't say by how much. In 2023, Quasimodo is just one seat off the result. We hope that Motu Research can catch up and account for them overhangs, because it really does matter.
Disclaimer: It is illegal for prediction markets such as Polymarket to allow New Zealanders to place bets on them. The Kiwi Polls model may not reflect the current feeling in the prediction markets.
THESE MARKETS DO NOT CURRENTLY HAVE A LOT OF VOLUME, MEANING THE ARE SUBJECT TO RAPID SWINGS.