Hacker News

points by brazukadev 4 days ago | 0 comments

Really unexpected.

Polymarket was like 80% Bolsonaro Jr to 20% Lula 7pm, 7:15 it changed to 60% Lula. There were multiple reports in the news websites about it favoring Bolsonaro which itself probably made some people vote for him.

I'm quite happy with the result tho. Let's see how the far right opposition will take this loss. I'm already expecting the worse.

[–] mulderc 4 days ago

Maybe these prediction markets are less useful than has been claimed.

[–] furyofantares 4 days ago

20% happens about 1 in 5 times. I know that's snarky wording but 20% shots come though constantly in daily life.

I think as humans we are really bad at intuition about probabilities, and it being an election only makes it much worse. I don't think anyone is consciously conflating these with poll numbers but percentages of voters are the numbers we see a lot more than prediction markets, and 80% is idunno an 80% chance of getting between 50% and 54% of the vote or something.

[–] alpha_squared 4 days ago

It's not a 'prediction market', it's a gambling market. It's far more accurate to say 80% of gamblers were convinced Bolsanaro would win than to say there's an 80% likelihood he would win. It's not measuring likelihood as it's untethered from things that impact the actual results. Obscuring it basic gambling with the word 'prediction' doesn't change the fact that's naked gambling at the end of the day.

[–] furyofantares 4 days ago

It's extremely inaccurate to say 80% of the gamblers were convinced because that's not how it works at all. Each bet has someone on the other side of it - 50% of the people bet on Bolsanaro and 50% against.

If you bet on 20% chances and they happen 30% of the time you make money. The payoffs are such that you win by being a better predictor than the market. I'm fine criticizing it as a gambling blight (not to mention a corruption payoff) but it is also a prediction market.

[–] quirino 4 days ago

The 80% is not the proportion of people who think he would win, that's not how these things work.

Every single factor was pointing towards Bolsonaro, Polymarket was just consistent with that.

He unexpectedly won the first round of the election. His lead was expected to grow since he'd likely get most of the votes of the other candidates. 17 out of 18 polls favored Bolsonaro. Market movements indicated a strong expectation of him winning. The Brazilian Real futures are already down 10% in the last couple hours.

But that's how democracy works, I think this election will be studied many times in the future.

[–] matheusmoreira 4 days ago

> The Brazilian Real futures are already down 10% in the last couple hours.

Brutal.

[–] smuhakg 4 days ago

If he has an 80% chance to win that means Polymarket will be "wrong" 20% of the time. And 80% seems fair given he was leading in the polls by a razor thin but consistent margin.

[–] jeffbee 4 days ago

Polymarket isn't binary so you are understating how wrong they were. Polymaket gives outcomes as percentage odds and less than 12 hours ago they had Bolsonaro by 5-to-1.

[–] cowlevel 4 days ago

They resolve as binary. Until they resolve it's a percentage chance of the eventual resolution.

[–] augusto-moura 4 days ago

They were never useful for the results, just for market expectation, which is never the same as real chances of something happening

[–] medellin 4 days ago

They are good at some markets but they don’t understand latam and especially Brazilian politics in the least so whatever makes it into the US news drives the market although it is normally wrong

[–] axus 4 days ago

Maybe it's useful for people spending money to present the odds they wish for?

[–] verdverm 4 days ago

and more manipulated for intended public influence than realized

this kind of gambling ought be made illegal along with the explicit in-pocket gambling

want to gamble, go to a physical place,

constant, easy access apps are already showing harm

[–] bryanlarsen 4 days ago

A 20% chance of Lula winning is still a real chance of winning. A standard rule of thumb would require 2 standard deviations aka 5% aka p<0.05 for relative certainty.

[–] jakkos 4 days ago

Polymarket also just had the Serbian election at 71.5% Vučić a few hours before the results came out of him losing

[–] estebank 4 days ago

If you think this result with those odds is unexpected, you must have never played X-Com :)

[–] rsynnott 4 days ago

Gambling websites are all but useless for predicting elections.

[–] platevoltage 4 days ago

We saw how they took it last time.

[–] medellin 4 days ago

Not really. If you follow anyone who has been covering latam for a long time they had odds of this election at a coinflip.