Unlike computing systems, the physical world is a shared mutable environment. It is effectively impossible to lock people out so that you can get a reliable, clean measurement even when people are not intentionally sabotaging the data model.
There are ways to robustly clean this up analytically but it is largely beyond the capabilities of current tech stacks.
You're right that this isn't a technical problem, it's a social one. This is why most sports leagues simply banned betting... until recently, now that gambling companies have bought their way in. The players always end up manipulating games.
Can you expand on that? Even just conceptually that sounds really hard, how would you know whether you're measuring genuine (unexpected) changes in the environment rather than the result of (possibly sophisticated and coordinated) deliberate manipulation?
Stitching unrelated proxies and sensing modalities into a coherent data model is a spatiotemporal graph reconstruction problem. The join predicates require non-trivial inference algorithms if you want to avoid being buried in false positives. From this you can derive an estimate of ground truth and a model of uncertainty at a point in space and time.
The model of uncertainty is dynamic and unpredictable. It is difficult to manipulate the measurement without producing data that falls outside the uncertainty model across every proxy by which someone might construct that uncertainty model. This is similar to how e.g. GPS spoofing is detected in military systems. All GPS updates must fit within a (classified) dynamic uncertainty model relative to INS; if an update falls outside the model then the GPS signal is presumed compromised and updates ignored.
At the limit, this restricts manipulation to values within the uncertainty model. If you have a lot of unrelated proxies, you can make the window of uncertainty tight enough that manipulation becomes effectively impossible. At a minimum, the adversary would need to be able to manipulate every proxy and modality feeding your uncertainty model simultaneously.
These graph, spatial, and spatiotemporal algorithms scale very poorly on traditional data infrastructure and these data models easily run into petabytes if you are stacking multiple independent data sources.
I was relaying the technical details of working in these data environments based on deep, real-world operational experience in the domain. There is no "moral dimension" to it, I was describing the world as it exists.
Does Feyerabend also have an opinion on compiler flags and sorting algorithms?
Feyerabend speaks to things that add context and nuance to the effects, consequences, and provisions of the things you've felt comfortable discussing so far.
Hence the recommendation. Your awareness could use some expanding, if I may be so blunt.
Example of this?
What kinds of purposes?
A lot of it is transnational and many countries are complicit so there is relatively little you can do to stop it. If you use data at risk of this then systems have to be designed under the assumption that it is going to happen. It has spread and become worse over time.
It is a bit like FinCEN data in that if you work with it long enough you'll see a lot more going on in it than it is your job to notice.
Now imagine we're can't have this basic resource anymore because some nerd fuckr needs to win a bet on the internet
Makes me want to bet on creative uses of a baseball bat
Ultimately they don't care too much about this kind of tampering, except when it has the potential to damage their reputation. They otherwise don't care about the specific outcome of a bet; they make their money regardless.
We are only seeing the very dumb here.
Good luck proving it.
> IIRC, someone bet on a streaker at the super bowl, did the streaking themselves, made 6-figures on the bet, bragged about it publicly, and the market negated the bet and recovered the money.
Yeah, exactly. Far easier to prove and to even get the idea to investigate that person in the first place.
So as long as you are subtle about it they won't even investigate, as a public investigation would make it seem like tampering was common.
The issue here is you'd need a lot more land because any asphalt near a temp sensor is going to give you bad data.
Ah yes. The way to measure temperature in the city is to move the thermometers outside of the city. Makes perfect sense.
Trying to eliminate gambling has vexed many an emperor and cleric.
In a bit of expensive equipment I own it happens to have 4 temperature and pressure sensors, located in two different locations on the unit. All of them generally disagree with each other on the exact temperature by around 0.1 C, which is fine for my use of it.
Building Byzantine fault tolerance into absolutely everything is expensive, and makes everything we do and buy more expensive. It would be better and cheaper to rely on social trust, if that's possible-- and it was possible before these gambling sites. Prediction markets are burning social trust as fuel to make profit, they should be heavily taxed, the way polluters are taxed, as destroyers of the commons.
I can reasonably test that the bolt is in a locked position every time I close the door and turn handle on the lock. But a single remote sensor could have malfunctioned or simply be out of calibration a few degrees.
If the actual temperature at the airport is important to any set of users enough so that the difference between it being 18 or 22 deg C is relevant, one should expect that there be at least 3 sensors (much like clocks) and assuming variances between the 3 sensors are within tolerance an average of the 3 temperatures is taken.
Three sensors doesn't solve the problem. Manipulating becomes marginally more difficult with three sensors, but it's still very possible, and with enough monetary incentive it's still even likely that it happens again.
So why not 5 sensors? How about 10?
And what about more consequential issues like the toppling of governments or military blockades where true redundancy is impossible and people are actually harmed?
Is there not a point where you start to blame the incentives that are being chased?
In my city, the airport has wildly different weather than the city center, and the data I get has no relation to what I experience, for example.
Being held hostage by bad actors until you fortify your defenses seems to be a very unnecessary technical solution to an easily predicted social one.
Normalizing this only means the subterfuge becomes more subtle, not that you remove it entirely. But you preserve the incentives by not changing this system. So you're spending a lot of resources wastefully when you could just... not.
A few years years ago, there was no conceivable profit motive for interfering with weather sensors on public property. Now there is one.
Can this really be said with a straight face?
Suppose you're an oil company, or a trader with a large position, and "hottest year on record ten years in a row" is bad PR that will make bills you don't like more likely to get passed. Or for that matter a company selling carbon capture stuff who wants to make sure it goes the other way. How about tobacco companies?
This has been a huge problem for as long as public data has been used to make decisions affecting profits.
No, that's way too broad. I don't believe that, and I didn't write that.
Oil companies indeed had enormous incentives to obscure and confuse the scientific record of climate change. But hiring thousands of Taskrabbits to go around the world blowing hot and/or cold air on weather sensors would not have helped them: some of them would have been caught, and it would've been exposed as a hamfisted and shambolic scandal.
Oil companies had to do larger-scale, longer-term stuff like funding think tanks, lobbying politicians, and writing op-eds.
Now, because of prediction markets, the "attack surface" for interfering with scientific data collection is much larger and more fine-grained. The incentives for big oil companies probably haven't changed much… but now any random amoral idiot or degenerate gambler has an incentive to go find unguarded weather stations or water quality monitors, etc, and mess with them.
Isn't that the same thing that happened here? It's a news story because they got caught. There is apparently already a law against this sort of tampering and it makes sense to have that law.
> Now, because of prediction markets, the "attack surface" for interfering with scientific data collection is much larger and more fine-grained.
Oil and tobacco companies are the canonical examples because they're huge and contemptible, but the incentives scale all the way down. A small industrial plant is only allowed to use river water for cooling as long as the temperature isn't too high, they now have the incentive to cause the reading to be lower. The local school doesn't have A/C and therefore cancels class whenever the outside temperature is above some threshold, what happens when the kids figure out that the temperature sensor they use is in a public place?
You can make messing with the sensor illegal to apply a disincentive to counter the incentive but that doesn't mean the incentive wasn't always there to begin with.
Also, the oil and tobacco companies have certainly been caught manipulating scientific results on multiple occasions.
I have bad news for you. The manipulation of weather sensors for profit has a history that long pre-dates prediction markets. You just weren't paying attention until now.
Okay, I'm intrigued about what I've missed. Give me more information about this history, please.
It's really bad, but it's not quite the same society-pervading incentive for data manipulation.
It's going to happen.
It'll happy by proxy too. When they figure 90% of Tesla's or SpaceX's valuation is tied to Elon, suddenly he becomes your mark.
There are roughly 2,500 individual companies listed in the US alone with a market cap that exceeds 1 billion dollars. Tesla's market cap is 1.4 trillion - about 1,000 times the size of all prediction markets put together.
In other words, to the extent this incentive is of societal concern, it is not because of prediction markets - that is the tail wagging the dog.
Edit: to be clear I think these markets are a scourge as they currently operate, but not for this reason.
(1) https://predik.io/en/blog/open-interest-mercados-prediccion-...
Notice also that the murder (or tampering with sensors etc.) is the thing you don't want happening in your jurisdiction, so that's the thing you have to prohibit anyway, because even if you ban short selling or prediction markets in your jurisdiction, having them in any other jurisdiction in the world would allow anyone to just place their bet there and you're right back to needing to deter that by punishing the murder rather than the bet.
> Later, a single user would make over $550,000 after betting that Ayatollah Ali Khamenei would topple, just moments before his assassination by Israeli forces.
It is entirely possible that there was causality in the forward direction if the bettor had inside information about the assassination operation beforehand.
I knew my previous employer was getting acquired before the markets did -- I had inside information -- but I had no way to make the deal go through.
Revealing inside information through prediction markets is mostly fine. You may think people shouldn't need a financial incentive to do so but clearly they do, and clearly other people are willing to put up with that money for it.
Probably automatable, and simply counting on automated systems and panic selling. Might be possible just with the AI package these days (not even requiring a traditional media article).
This is a ridiculously over-broad claim.
First of all, prediction markets shift incentives in a pernicious way; they don't magically erase or override all pre-existing incentive structures.
And secondly, professional athletes of all stripes appear to be receiving a lot more threats of violence from gamblers (including "prediction market traders") than they were a few years ago.
'Decapitated bodies, death threats, that's our daily life': Tennis players, helpless targets of angry gamblers [ https://www.lemonde.fr/en/france/article/2025/12/08/decapita... ]
https://www.wtatennis.com/news/4288181/wta-and-itf-release-a...
When the vice industry kept gamblers mainly confined to their own community, it was ugly enough (see the “people with pipes” debt collection trope you refer to). I don’t care too much about how people get their jollies. But I don’t see much good in letting vice behaviors and motivations slop out onto the wider world uninvited.
To your question, then: I’d draw the moral line at “don’t base forecasting gambles on any indicator produced by someone who didn’t agree to be used that way.” If I—as the entity or constituency who incurs the cost to produce an indicator—want to hire the prediction market to run bets on it at arms’ length from me, fine. If not, no.
The burden should be on the gambling interests to prove they’re contributing value to a specific domain, not on everyone else in every domain to prove gambling’s specific negative externalities.
Quite likely it has already happened.
There's been an enormous amount of betting on various military strikes by the Trump administration, much of which looks like insider trading.
Almost certainly there have been insiders with money on the line who were financially motivated to launch lethal attacks, which they might have otherwise paused or delayed.
It's much more effective to make the decision first and the bet second.
Your decision maker is now going to make decisions differently than they would if they hadn’t made the bet, due to the financial incentive. It hopefully won’t be blatant, but they’re going to be more committed to the original decision.
What did I say that contradicts this?
> Why almost certainly?
Because I haven't seen any clear reporting with evidence documented down to the level of individual decision-making.
I'm pointing out that the version where the market causes the decision is the much less effective way to do things. So I don't think it's so certain to have happened.
When the decision comes first it's still a problem but it's a very different problem.
I should make this explicit, I guess: If you make the decisions before betting, you don't really care what the market says beforehand, you'll make plenty of money either way.
> the version where the market influences the decision is the much less effective way to do things.
Okay, so two possible scenarios: "market influences decision" and "decision influences market."
1. Polymarket offers a contract called something like: "Will the US launch any lethal boat strikes in the Caribbean on April 30, 2026?" A commander of a drone unit for the US military bets $500 on "yes". At 11:51pm on April 30, he's trying to decide whether or not to issue an order to fire on a boat, or to wait 15 minutes for better visual identification of the target. As a result of his bet on Polymarket, he's strongly incentivized to issue the order right now.
2. It's 11:51pm on April 30. A commander of a drone unit for the US military is reviewing images of a potential target boat. He idly opens a browser tab to Polymarket. Lo and behold, he sees that the odds of "Will the US launch any lethal boat strikes in the Caribbean on April 30, 2026?" are at only 0.3%. He's strongly incentivized to issue an order to strike right now and to quickly buy that Polymarket contract before news of the strike breaks.
Does it matter which of these two scenarios might actually be taking place? I don't think so.
Also there's plenty of things that are unlikely because they're specific, not because they're irrational. Things happening in specific date ranges are an easy one.
A fortified compound would affect the temperature reading.
Here's the negative externality that no one will care about. There's no reason gamblers won't repeat this stunt, until us poor schmucks who just want an accurate temperature reading have to build a fortified compound in order to do so.