This is a very simple simulation of the upcoming South Australia State election
The election is simulated 50K times adding in random polling errors
- 100 simulated elections are shown on a line graph, sorted from highest incumbant seats to least incumbant seats
- This model is limited and likely under predicting crossbench seats
- Seemingly paradoxically I think that it's also overestimating the chance of a hung parliament
- Tweaks are needed to preference flows - on a 2PP of ALP 50 - 50 Coalition it seems to unrealistically favour one side by a large margin