Hacker News

points by NewJazz 4 days ago | 0 comments

* Lula did something unusual, which is win in the second round after trailing in the first. This is a first in Brazilian politics, so simple pattern matchers are surprised.

* Turnout. Made all the difference.

[–] ArcHound 4 days ago

I'm not surprised. If you have one terrible candidate, in the first round you can vote for whomever you like. In the second round, however, you need to ensure they lose, so you vote for the opposite party, whomever they might be.

[–] Twisell 4 days ago

France here might disagree. If you have an overly scattered left wing by voting whomever in the first round... you end up with a macron flavored moron as the sole non far-right option in the second round and have to deal with this shit for the next five years.

[–] ArcHound 4 days ago

Yes, for sure, it's not a serious model. But I've observed exactly this scenario a couple of times, so it tracks well for me.

[–] nkmnz 4 days ago

The only way for the far right to win the second round is to end up against Mélanchon, so what you write is totally in line with what Le Pen would want you to propagate.

[–] chernobogdan 4 days ago

There were 10 presidential elections in Brazil after the post-1988 constitutional system, with 2 of those decided in the first round.

8 is a really small sample to consider it as unusual, plus the reversal was not that huge 1.8%

[–] NewJazz 3 days ago

Right, but simpler pattern matchers don't always consider that.q