LOESS curve of opinion polls of voters for the 45th Canadian federal election:
Dark mode for viewing pleasure:
Up-to-date as of September 12, 2024. 30-poll smoothing factor. Square root applied to sample sizes when weighting. Highlight ribbon is of 95% confidence interval of local regression standard error (not polling margin of error).
As explained earlier in the comments, this market was poorly made and will simply become less accurate if one of the candidates leaves. The JT option is for JT only, not a hypothetical Liberal candidate. As of the time of writing, this market would resolve N/A if none of the listed candidates win. However, there's some talk about whether Manifold will add an auction system, where I could close the market, change the JT option to a new Liberal candidate, and then hold an auction to make it fair when reentering the market with this new option. But unless a fair method of closing and reopening the market is added, I won't be doing that
New record for the Conservatives. New Mainstreet poll shows their highest ever lead for the Conservatives at 22pp, the second highest lead ever recorded, and the highest recorded lead for an A- or higher rating pollster. It also shows Conservatives with 45% support, the second highest support level ever, the highest being 45.6% back in early March. Liberals and NDP combined only make up 38%, a 7pp lead for the Conservatives over the failing duo
Canada and America are acting in parallel. Both incumbent left-wing leaders are being asked to not seek reelection: https://www.bbc.com/news/articles/c9781gr02p5o
New record for the Liberals, lowest ever support ever recorded for this election cycle at 21% support, NDP in a statistical tie at 20% support. Conservatives dominating with 42% support. Liberals also got another record, two consecutive worst results ever back-to-back, at 22% just 4 days prior. Last time Liberal support was this near this level was November 2023.
Abacus shows its second largest lead for the Conservatives, and largest lead of any pollster for the Conservatives since May 10th
@XVII Just advice, always buy Pierre YES, never Trudeau NO, as that pushes up the probabilities of every other candidate, not just Pierre
I’m sorry but you simply don’t understand how probability works. saying there was a 1% chance of something happening isn’t the same as thinking it will happen. I don’t even think there necessarily was a 1% chance the S&P dropped I was also hedging against my financial interest
You get 0 mana if you bet the wrong way. You didn't bet that there was a 1% chance of it happening, you simply bet that it would indeed happen, and lost mana. The hedging argument doesn't make sense as it only works if there's a belief that it could happen, which is an embarassing belief to hold. Again, loss of social capital for betting YES on that market
Not sure what relevance this paper has to our discussion. He's claiming he said there's a 1% that it'll happen, which is false, as he bet YES. Your paper simply adds theory to back up the fact that the market is the mean belief. If the average belief is 1%, and he's the only person who bet YES, it means he thought there was a much higher than 1% chance of it happening, as my interrogation skills and your paper confirm. He tries to damage control by claiming that he secretly believed it to be less than a 1% chance, which in a normal-functioning brain would create a NO bet to bring the 1% even lower, but maybe he misclicked and is trying to explain it away.
Regardless, my claim continues to remain undefeated, that it is utterly embarrassing that someone would ever think there was a chance of the S&P 500 dropping 15% or more in one day in January 2024. It's a dramatic loss of social capital
@DylanSlagh You bet YES that an event that has happened only once in history (with no close second), on Black Monday no less, would happen again, in the tiny time frame of one month. The fact your brain produced an above-zero probability for this event is, as oft-repeated, embarrassing
Not to mention, this was during a period of stable growth:
@nikki Here's Sydney (Microsoft Copilot) agreeing with me and not you that it's an irrelevant paper:
Assigning a zero probability to any event, however unlikely, is what’s embarrassing. 0 and 1 are not probability and can not be treated as such. Assuming that I must have thought there is a “much higher than 1% chance” in order to bet at 1% is idiotic. I’ve made thousands of bets on manifold many of which are silly. I don’t take it that seriously and it’s you who look idiotic for zeroing in and obsessing over a single bet.
You claimed it was embarrassing to believe in absolute certainty and linked to a post that quite literally is self-aware that it goes against probability theorists and is just a personal opinion. The fact you claimed it was a fact that they weren't probabilities just shows you don't even read what you cite, just further evidence that your brain is embarassing.
The fact you think you betting "at 1%" means something is hilarious. You bet YES, that's what you did, it was cheap, sure, but it was a binary option and you chose YES. It resolved NO, you lost your mana. At least you're damage controlling in a new way, claiming you were simply being silly. That's an infinitely better damage control than "aCtUaLlY I WaS hEdGiNg!!111"
I'm merely zeroing in on a perfect example of 0% probability bets costing you social capital. You seem pretty defensive rn, lobbing insults left and right, calling people obsessive for staying on topic. Embarrassing ngl
@Bayesian I hope you’re not actually using P(omega) to justify that far-fetched bet on the S&P 500. Assigning any non-zero probability to such an outlandish event is like saying there’s a chance I might win the lottery without buying a ticket. In the real world, we base our predictions on data, not daydreams. Betting on a 15% drop was a leap into the realm of fantasy, and let’s be honest, we’re not here to play make-believe. So, while you’re busy contemplating the probabilities of the improbable, I’ll be over here, firmly grounded in reality, where such drops are as likely as me becoming the next monarch of Canada
i think this convo is unnecessarily hostile and hedging past your beliefs is fine if it reduces risk even if it’s negative ev, and I agree 14% drop was very unlikely but it doesn’t seem to be a big deal to be making bad bets sometimes, I don’t think that’s too embarrassing. The prob of that wasn’t 0% obviously, despite the fact that non-epistemic events can have 0 or 1 probability. Taking a YES bet at 0% actually can’t possibly worsen your calibration bc if it doesn’t happen that’s the % at which u predicted it (so neutral bet) and if it happens u make infinite ROI.