Will the scientific consensus that the universe's expansion is accelerating end before the end of 2030?
➕
Plus
91
Ṁ71k
2031
12%
chance

Specifically, on Dec 31 2030.

https://en.wikipedia.org/wiki/Accelerating_expansion_of_the_universe

Based on a discord conversation

Get
Ṁ1,000
and
S1.00
Sort by:

I'll have to look into the quality of the claim before disagreeing with the market price or anything like that, but just so that people are clear, this question is asking about the consensus that the expansion is accelerating, not the consensus that the universe is expanding at all.

That the universe is expanding at all is extremely, ridiculously well established, very easy to verify, not at all subtle, and we would be extremely confused about all astronomical data if it weren't true. Nothing would make any kind of sense. There's little chance that's getting overturned unless we're all living in a simulation and the gods running it fed us fake data.

That it's accelerating was discovered in 1995 1998 and relies on subtler data. Whilst I'll defer to cosmologists on how firm this is (if they're generally saying there's no chance that will be overturned, then they're probably right), this being wrong is a considerably weaker claim than the universe not expanding at all. If an oracle told us the acceleration was wrong, I'm not under the impression that this would irreparably break astronomy. It would rule out some cosmological models and un-rule-out others, and I believe life would go on.

I believe the most accepted models of early evolution of the universe do rely on rapid acceleration, but I'm also under the impression that this is basically a different mechanism to the current, much more modest, acceleration, which didn't come to dominate until much later in the universe's evolution.

Anyway, this is one of the situations where actually looking at the quality of the data might matter more than priors. I'm not a cosmologist, maybe there are strong priors for acceleration at this point that I'm unaware of. If so, maybe cosmologists can dismiss this on priors. Otherwise it's an empirical question about whose data was better - the data establishing the acceleration, or the data failing to replicate it. The former won a Nobel prize though, so I'm guessing it's pretty good!

@chrisjbillington I knew I'd seen this reported on somewhere, and it must've been this:

https://www.ox.ac.uk/news/science-blog/universe-expanding-accelerating-rate-%E2%80%93-or-it

This seems to be somewhat unrelated to the discussion on discord that prompted this market, but as far as I can tell these are serious people published in serious journals, working with large datasets - much larger than were used to establish the acceleration initially.

The result is still more consistent with acceleration than not. Three sigma is still a lot - and I don't really buy the comparison to particle accelerator results, since particle accelerators are the epitome of the "multiple hypothesis testing" problem and therefore need smaller p values, whereas this is just one hypothesis being tested.

But yeah, interesting, it does seem like it's not totally beyond the pale that the supernova observations could be a statistical fluke. Three sigma is a p value of .003, so still very small, but you'd like it to be more definitive. Seems like a bit of a stretch for the author to describe three sigma as "consistent with no acceleration though", like, it's a pretty small p value, it's not very consistent with no acceleration, if I'm understanding it correctly.

And yeah, this is ignoring other evidence for dark energy that I'm not very familiar with. Perhaps cosmologists don't really care much about the supernovae anymore, I don't know.

predicts NO

@chrisjbillington The problem with this is that for any real but small effect, you can do a study to look for it and get statistically insignificant results, especially when the threshold for significance is as high as it is in physics. Getting a three sigma result instead of a five sigma one is still evidence in favor of the universe's accelerating expansion and can't undermine all the other evidence.

It would be evidence against acceleration if an experiment ruled out what's currently believed to be the expansion rate with high confidence. Instead of just saying "the deviation from zero acceleration is <5 sigma", say, "This puts on upper bound on the rate of acceleration that is very close to zero". But I don't think anyone has done an experiment like that.

predicts NO

@chrisjbillington Also worth noting that this is from 2016. If the study that was reported on has merit, it's probably been followed up on already. I don't know what follow-up has been done, but obviously nothing has happened yet to break the consensus.

@PlasmaBallin Right, I totally agree and said this above - three sigma is still heaps. But if we're only talking about supernova data, then this study used loads more data than the previous ones establishing the acceleration in the first place, so I don't think ruling out the previous measurement is necessary - instead you can think of it as a fairly high powered failed replication attempt. If it had a one or two sigma result, that is. Three is enough that it is actually a replication, though it's interesting that the result isn't stronger still.

Also worth noting that this is from 2016. If the study that was reported on has merit, it's probably been followed up on already.

I'm not sure about that, are we discovering meaningfully more type Ia supernovae still? Or was there one huge survey prior to 2016 that got most of the ones available, such that anyone wanting to follow up on this result just has the same data to work with? 2016 doesn't feel like that long ago to me, it wouldn't surprise me if it just hasn't been followed up on (not with new data, at least - people could at least check the analysis).

FWIW, I'm getting the impression that this isn't thought of as a fringe paper in particular need of reproduction (not until we have more/different supernovae to analyse, anyway), and that it's just another fact added to the pile of astronomical observations cosmologists have to keep in mind - I would guess people in the field simply think "ah yes, the current evidence for acceleration in the supernova data is a little under 3 sigma". And if they believe in accelerated expansion more strongly than that, it'll be for reasons other than supernovae.

Probability should be <1%.

predicts NO

@PlasmaBallin yes¹ but there are plenty of other markets like this which are going to resolve much sooner

___

¹I mean, with the implicit "Conditional on Manifold still existing by then"

predicts NO

@ArmandodiMatteo My comment was just a statement on what the actual probability is, not on the probability that a rational agent should bet this market down to. This market should of course be expected to have a higher probability than the true probability for the reasons people have already mentioned.

predicts NO

I asked one of the guys who discovered that the universe is expanding (now a Distinguished Professor of cosmology, director of an institute, and former chair of the physics department at UC Davis) if he felt this market was accurate. He thinks the percentage is more than double what it should be, as does a colleague he discussed it with.

I don’t personally know enough about the object level to have an opinion.

predicts NO

@SoniaAlbrecht I know him pretty well, and deeply respect his epistemic humility, so I would expect him to be unusually good at not overestimating the probability that he is right that it is accelerating.

predicts NO

I agree.


There's no incentive for me to bet on that, even with loans, because it resolves in 2030.

@SoniaAlbrecht Doesn't surprise me at all, I would be shocked if any significant proportion of physicists thought there was a non-negligible chance of the consensus changing. This market might as well be free mana.

Yeah this is unsurprising, any probability on Manifold of <10% or >90% tends to be skewed towards 50%. It's a really bad look; easy to see why a physics professor would dismiss the platform if they see things like this.

predicts NO

@jacksonpolack dude you have 35k mana, why not? Looking at your graph you'd get your money back before you even decide to use it and you'd have a long term investment that will apreciate in value.

predicts NO

It'd appreciate in value by 10% over six years, which is worse than some bank accounts. There are plenty of other markets on 2-3 year terms I think are better investments

@jacksonpolack You can do both, that's what the loans are for.

predicts NO

You can do both, that's what the loans are for

No, I mean if I wanted to put all 35k into long-term investments there is a very long series of markets that I'd get better return on before this one.

And I'd like to keep quite a bit of that balance available. One reason is that if I'd been more attentive to the government shutdown markets days ago, I could've deployed much of that at the time 30k for quick profit, e.g. taking some of BTE's 20k of limit orders at 20%.

predicts NO

@SoniaAlbrecht I just put in all my mana for the day to improve the epistemic commons despite the fact that there is almost 0 incentive to do even if the real probability here is way under 1%, only to see someone else then buy it up to 10% :(

predicts NO

yeah, I created this market to trade with him at 50%

@SoniaAlbrecht who's the professor?

predicts NO

@DanielFilan A laughable (in retrospect) amount of evidence points to Andreas Albrecht

predicts NO

@DavidMathers Yeah, this is the main reason I don't bring markets like this down to the true probability. I would make a profit in 7 years, but if people bet it back up, I'd have to either constantly spend more mana to correct it, or have a big negative profit for 7 years before it finally pays off.

Also, one of the downsides of the new loan system is that when you really do have a better probability estimate than the market, you won't get all the mana you invest into a long-term market back in loans.

predicts NO

@Frogswap Yep that’s him lol. That’s why I know him well enough to expect him not to excessively overestimate the probability that he’s right.

predicts NO

@Austin is there anything you can do about this long term market problem?

predicts NO

@SoniaAlbrecht Its funny how this thread is slowly bringing down the probability. The expected behaviour

predicts NO

@ScipioFabius In this case, the probability is basically about how much I’m willing to put down versus the YES better. The thread didn’t convince me that I was more right, but it did convince me that having a correct percentage here was worth putting down more money. So far I’ve put down $200 USD worth, so I don’t want to put down much more. If someone else wants to push the probability down further though, feel free! 😃

predicts NO

The cosmologist said nothing has changed in the research that caused the probability in this market to shoot up recently. He’d still like to see the market around 2%.

@SoniaAlbrecht I think it should be even lower than that, though interest rates probably prevent it.

predicts NO

@PlasmaBallin well there is a base rate of the consensus changing

predicts NO

@MartinRandall The base rate for scientific consensus on something this firmly established changing in six years is very, very low.