预测市场即新媒体:用赔率取代民调与评论 · Prediction Markets as the New Media
主题综述
更新日志
- 2026-08-10 — 首次综述。基于 4 篇访谈(CORE 判定共 5 个 id,其中 1 个为同集归档副本,未重复计入),核心结论:两位头部交易所创始人已经把「预测市场=媒体」从比喻讲成产品事实——98-99% 的用户不交易只看赔率、CNN/CNBC 把赔率嵌进新闻、战区用户拿它当比 OSINT 更快的信息源——但语料全部来自供给侧,「比民调更准」仍停在断言;真正的分歧不在要不要取代旧媒体,而在与旧媒体是共生还是解药、真相机制是「群体参与」还是「价格形成 + 知情者进场」。
主流共识
先交代语料的形状:四场访谈、两位发言人,全部是预测市场的经营者——Kalshi 的 Tarek Mansour 占三场(20VC、Bloomberg Odd Lots 现场、Chris Power),Polymarket 的 Shayne Coplan 占一场(Axios BFD 现场)。所以本节的「共识」是两家头部交易所创始人的共识;反方(民调机构、新闻编辑部、学界)在这批材料里缺席。在这个前提下,共识有六层,而且整齐得近乎口径统一。
第一层:使命本身就是取代民调与评论员。 这不是外界给的标签,是 Mansour 自己的公司使命表述——现状被他概括为「民调 + 想说什么说什么的评论员」:
"the mission of the company is like, we want to bring more truth to a lot of these processes, whether it's the election process, right now we're using polls and pundits and people say whatever they want, or weather or climate or the economy and how we make decisions about the economy. … So imagine where election night coverage is informed by prediction markets. Which actually work better than what we're doing today."「公司的使命是,我们希望为许多流程带来更多真相,无论是选举过程,现在我们使用民意调查和专家,人们想说什么就说什么,或者天气、气候或经济以及我们如何做出有关经济的决策。……想象一下,预测市场为选举之夜的报道提供信息。实际上,这比我们今天做的更好。」
第二层:主产品是信息,不是交易。 整个主题最锋利的数字出自 Mansour——每 100 个用户只有 1-2 个在交易,其余 98-99% 把赔率当新闻读:
"the prediction market user is not just a trader. Actually, out of every 100 prediction market users, maybe there's like one or two traders. They're actually trading on the market, but the rest of the users are getting informed about the world. It's really a form of news. It's a form of media. It's a way to basically come and get truth about what's going to happen next in a world where actually it's been really hard to figure out what's happening."「预测市场的用户不仅仅是交易员。实际上,每 100 个预测市场用户中,可能只有一两个是交易员。他们真正在市场上交易,但其余的用户都在了解世界。这实际上是一种新闻形式。这是一种媒体形式。这基本上是一种来获取关于接下来会发生什么真相的方式,在一个实际上很难弄清楚发生了什么的世界里。」
Coplan 从产品端说的是同一件事:概率数字被当成大众消费级内容来设计,赔率曲线本身就是可读的信息产品:
"Not everyone can understand it, but when you distill it down to here's the percentage likelihood for this happening and here's how it trends over time, it becomes very digestible and very useful for people. … when it's distilled down from a consumer perspective, it's really informative."「不是每个人都能理解它,但是当你把它提炼成 “这是这件事发生的可能性百分比,这是它随时间变化的趋势” 时,它变得非常容易理解,对人们非常有用。…当它从消费者的角度被提炼出来时,它确实很有信息量。」Shayne Coplan · Polymarket's Shayne Coplan & Axios' Dan Primack
第三层:赔率跑在新闻前面。 速度是两人都拿得出的实证。Mansour 的版本是选举夜——传统电视网比 Kalshi 晚 5-7 个小时才敢报出谁赢:
"People flock to Kalshi because you go on the traditional news networks and you don't know who won until five, six, seven hours after Kalshi, right? And that's extremely valuable. If we can see and get a little bit smarter about the future, we can make better decisions over time, allocate resources better and manage risks better."「人们涌向 Kalshi,因为你在传统的新闻网络上,在 Kalshi 之后五六七个小时都不知道谁赢了,对吧?这非常有价值。如果我们能够看到并对未来有更多的了解,我们就能随着时间的推移做出更好的决策,更好地分配资源,更好地管理风险。」
Coplan 的版本更极端——不是电视观众,是战区居民把 Polymarket 当作事关自身安全的信息渠道:
"people who are in the Middle East, they oftentimes have heard in the WhatsApps and the group chats, it's like people are getting their information from Polymarket because it moves even quicker than the Telegram groups and the OSINT channels. And it's also obviously easier to verify. And that moves even quicker than the news."「身处中东地区的人们,经常在 WhatsApp 和群聊中听说,就像人们从 Polymarket 获取信息一样,因为它比 Telegram 群组和 OSINT 频道传播得更快。而且显然也更容易验证。并且它比新闻传播得还要快。」Shayne Coplan · Polymarket's Shayne Coplan & Axios' Dan Primack
Mansour 还把这个逻辑推到「唯一来源」的场景——TikTok 禁令这类重大新闻,民调与专家给不出概率,只有市场能("friction markets" 为转写瑕疵,原意 prediction markets,照录):
"TikTok ban is definitely my number one, because I would say two things. One, it's just such a good representation of the power of friction markets, where there's no other way to get truth about the probability of this thing happening, and it's actually important. TikTok being banned versus not is actually super important."「TikTok 禁令绝对是我的第一名,因为我想说两件事。一是,它很好地代表了摩擦市场的力量,在摩擦市场中,没有其他方法可以获得关于这件事发生的概率的真相,而且它实际上很重要。TikTok 被禁与否实际上非常重要。」Tarek Mansour · Kalshi CEO Tarek Mansour - The Future of Free Markets
第四层:真金白银是信息质量的来源。 skin in the game 是两人共享的认识论底座。Mansour 把它说成一条普适原则——一切定性分歧都该有定量裁决机制:
"I think anything that has a difference of opinion, difference of opinion, of qualitative opinion should basically have a mechanism to resolve that quantitatively. So like instead of debating subjectively or something, like let's debate about it objectively and quantitatively, which is trading, you know, what's the point."「我认为任何存在意见分歧的事情,意见分歧,定性意见,基本上应该有一种机制来定量地解决这个问题。所以,不要主观地争论,让我们客观地、定量地争论,也就是交易,你知道,有什么意义呢。」
Coplan 的机制表述与此完全兼容:yes 份额兑付 1 美元,交易价 20 美分即市场认为 20% 概率——赔率即概率(他对这套机制的进一步阐释见分歧线二,那里他与 Mansour 的词汇分道扬镳)。
第五层:媒体只是第一站,终局是决策基础设施。 Mansour 在 Odd Lots 把创业原点讲成「给未来问题定价的市场会是最大也最重要的市场」,长期路径是从价格基准走向机构风险转移——他给的门槛数字是至少八位数、约 5000 万至 1 亿美元才值得机构配置,深度机构用例预计未来 2-3 年启动。在 Chris Power 场他把层级说得最清楚:
"Prediction markets are extremely cool and very useful society in terms of being able to engage with the news more meaningfully and see the future better. But it's not just that. The core unlock that could happen with prediction markets is if you forecast the future better, you are pricing risk. … If you can price risk, then you can make better decisions about the risk. If you understand the risk better, you can make better decisions about how to manage it. Also, you can transfer it. You can hedge."「预测市场非常酷,并且在能够更有意义地参与新闻和更好地预测未来方面对社会非常有帮助。但不仅仅是这样。预测市场可能发生的关键解锁是,如果你能更好地预测未来,你就是在对风险进行定价。……如果你能对风险进行定价,那么你就可以对风险做出更好的决策。如果你更好地理解风险,你就可以对如何管理风险做出更好的决策。此外,你还可以转移它。你可以对冲。」Tarek Mansour · Kalshi CEO Tarek Mansour - The Future of Free Markets
Coplan 给出的是同一愿景的历史谱系——DARPA 政策分析市场的商业化复活:
"there was this thing called the policy analysis markets that DARPA had funded that never really came to life that Effectively forecasted the promise of this idea that you could use prediction markets, in that case combinatorial markets, to go and get a peer into the future of what was likely to happen and make better decisions. And to see that coming to life, albeit very early, It's the coolest thing to me."「以前有个东西叫政策分析市场,是 DARPA 资助的,但实际上并没有实现。它有效地预测了这种想法的前景,即你可以使用预测市场,在那种情况下,可以使用组合市场来了解未来可能发生的事情,并做出更好的决策。看到它变成现实,即使还很早期,这对我来说是最酷的事情。」Shayne Coplan · Polymarket's Shayne Coplan & Axios' Dan Primack
第六层:2024 大选是分水岭。 数字都堆在这一年:2024 年 10 月 7 日 Kalshi 诉政府上诉胜诉(Mansour 称之为行业的「ChatGPT 时刻」),一夜增长 100 倍;2024 年秋大选季 App 新增 200 万用户;此后 Kalshi 完成 10 亿美元融资、估值 110 亿美元(访谈当周宣布),团队仍约 100 人。Mansour 对转折的描述很简洁:
"Obviously the vibe has shifted pretty dramatically with the election because there is some power to these markets."「显然,由于选举,气氛已经发生了巨大的变化,因为这些市场确实存在一些力量。」
主流化的现场注脚由主持人给出:
"You're definitely having a moment. The South Park episode last night was all about prediction markets."「你们绝对是迎来了高光时刻。昨晚的《South Park》剧集全是关于预测市场的。」Joe Weisenthal(主持人) · Tarek Mansour on Kalshi's Plan to Create Markets in Everything(Odd Lots 现场)
共识到此为止。一旦追问「与旧媒体到底是什么关系、赔率凭什么可信、怎么回应赌博与操纵指控」,两位创始人——甚至 Mansour 自己在不同场合——开始分叉。
分歧在哪
先老实交代:这份语料里没有预测市场的反对者,所以下面不是「正方 vs 反方」,而是布道者阵营内部的三条裂缝。裂缝是真实的——第一条甚至从 Mansour 一个人身上穿过。
分歧线一:旧媒体是要嵌入的管道,还是要替代的坏过滤器
Mansour 在 20VC(宣布 CNN/CNBC 合作当周)讲的是共生。 CNN 独家合作加次日的 CNBC,是预测市场史上第一、第二个大型新闻网络合作,理论包装是「新闻报道现在,预测市场报道接下来」:
"this idea of just like embedding prediction markets in the news, you know, there's like so much symbiosis here, which is like the news can get better from prediction markets because prediction markets are like this sort of Think of it as like the news covers what's happening now, prediction markets can extend the news to cover what's about to happen next. Which is kind of the next sort of next iteration of news."「将预测市场嵌入新闻中的想法,你知道,这里有很多共生关系,就像新闻可以从预测市场中变得更好,因为预测市场就像这样一种东西——可以把它看作是新闻报道现在发生的事情,而预测市场可以将新闻延伸到接下来即将发生的事情。这有点像是新闻的下一次迭代。」
但共生不等于安分。被 Harry Stebbings 直接问「你不认为你可以创建下一个新闻网络吗」,他的回答是两条路并行:
"I think we can. They're not mutually exclusive. I think we can do both. Even today, there's not a single news network that owns all the audience. There's some news networks like CNBC for finance and business. CNN is for politics. They're not mutually exclusive."「我认为可以。它们并不相互排斥。我认为两者都可以做。即使在今天,也没有一家新闻网络拥有所有的受众。有些新闻网络,比如 CNBC,关注金融和商业。CNN 关注政治。它们并不相互排斥。」
扩张目标也直接对标新闻编辑部的选题面——运营上有支撑:上架一个新市场的时间已从最初的 18 个月压缩到 30 分钟,市场数从 2024 年 9 月的 100 个增至 2025 年初的 1,000 个:
"The last part is like more broad, like we want to have a market for anything that's in the news. We have some room to grow there. I think we're at maybe 15, 20% of the relevant news on any given day."「最后一部分更广泛,比如我们希望为新闻中的任何内容都设立一个市场。我们在这方面还有增长空间。我认为我们可能达到了任何一天相关新闻的 15%、20%。」Tarek Mansour · Kalshi CEO Tarek Mansour - The Future of Free Markets
但同一个 Mansour 在 Chris Power 场讲的是解药。 媒体是坏过滤器,预测市场之于民调与评论,等于 X/Twitter 之于新闻业:
"In some ways, actually, the Trumps were forced into talking directly to the American people because the media was actually a bad filter, a bad intermediary for their message. … Twitter is the best way to aggregate, use the wisdom of the crowds for where the news, the past, for what just happened or what happened. Prediction markets are the way to use the wisdom of the crowds for what is going to happen. Applied to elections, people realize, oh shit. Everything that was going on with the news actually was going on with respect to punditry, how we view the future, polling, how we think about the election, and prediction markets were the antidote, the same way that X and Twitter were the antidote for the news."「在某种程度上,实际上,特朗普家族是被迫直接与美国人民对话的,因为媒体实际上是一个糟糕的过滤器,是他们信息的糟糕中介。……Twitter 是整合的最佳方式,利用大众的智慧来了解新闻、过去,了解刚刚发生的事情或已经发生的事情。预测市场是利用大众的智慧来预测将会发生的事情的方式。应用于选举,人们意识到,糟糕。新闻中发生的一切实际上都与评论、我们如何看待未来、民意调查、我们如何看待选举有关,而预测市场是解药,就像 X 和 Twitter 是新闻的解药一样。」Tarek Mansour · Kalshi CEO Tarek Mansour - The Future of Free Markets
解药论不是抽象站队,是打过仗的。大选期间特朗普赔率领先民调叙事时,主流媒体对预测市场的定性是「完全是假的,别信」:
"So that was a time where Trump odds were up and then we were getting slandered in the press. I don't know if you remember that time. There was the Times article. There's a bunch of articles that people came out which is like, these prediction markets are totally false, don't trust them. It's normal. Not to make this partisan, I think it is either side. Whenever the odds were up for either side, that side was excited about the markets and whenever they were down for that side, that side was way less excited about the markets."「当时特朗普的赔率很高,然后我们在媒体上受到了诽谤。我不知道你是否还记得那段时间。时代杂志上有一篇文章。还有很多文章说,这些预测市场完全是假的,不要相信它们。这很正常。我不想把它变成党派之争,我认为双方都是如此。无论哪一方的赔率上升,那一方都会对市场感到兴奋;而当那一方的赔率下降时,那一方对市场的兴趣就会大大降低。」Tarek Mansour · Kalshi CEO Tarek Mansour - The Future of Free Markets
Coplan 从头到尾只有一副面孔:绕过。 他的创业动机就是媒体失灵——疫情期的社交媒体噪音里提不出信号:
"And I was on X the whole time, Twitter. It was just this like endless supply of misinformation and pontification and different people claiming different opinions. And it was just impossible to distill the signal from the noise. So I was like, fuck it. I'm going to go build this."「我一直都在 X 上,也就是 Twitter。那里充斥着无休止的虚假信息、夸夸其谈和不同的人声称不同的观点。根本不可能从噪音中提取信号。所以我想,去他的。我要自己构建它。」Shayne Coplan · Polymarket's Shayne Coplan & Axios' Dan Primack
把三段并排看,不对称很清楚:Mansour 的媒体化要借旧媒体加冕——CNN/CNBC 官宣、在 Odd Lots 上预告「很快将与一些大型新闻机构宣布一些事情」;Coplan 拿出的证据全是用户自发绕过旧媒体(中东群聊、比 Telegram/OSINT 快)。注意他在这场访谈里没有提出任何与新闻机构合作的路线。最有讽刺意味的注脚是:给 Coplan 定调的恰恰是主流媒体自己——Axios BFD 的开场旁白把「用赔率取代民调」直接写进了介绍语:
"He's turning public opinion into tradable insights, blurring the line between data, markets and transparency."「他正在将公众舆论转化为可交易的洞察,模糊了数据、市场和透明度之间的界限。」开场旁白(Axios BFD) · Polymarket's Shayne Coplan & Axios' Dan Primack
分歧线二:赔率凭什么可信——「群体智慧的参与回路」 vs 「价格形成 + 知情者进场」
这是两人之间最实质的理论分歧,虽然他们大概率不承认这是分歧。
Mansour 的真相观是参与式的。 他的词汇是「大众的智慧」(见分歧线一解药引文原话)、辩论、行为迁移——受众从被动看新闻变成主动下注的参与者:
"People are kind of turning from like passive watchers of events or news or sports or other things that they engage with in real life to active participants. That natural behavior of wanting to predict the future or debating about something is turning into activity on Kalshi. That is rare."「人们正在从被动地观看事件、新闻、体育或其他他们在现实生活中参与的事情,转变为积极的参与者。想要预测未来或辩论某些事情的自然行为正在转化为 Kalshi 上的活动。这很罕见。」
增长飞轮在他嘴里就是一个媒体行为回路——Twitter 上有人放话,人们去查赔率,然后下单:
"There is this kind of habit sometimes like someone makes a crazy prediction or says something dumb on Twitter and then people go and check the odds and then they place a trade. There's this loop that happens and I think we need to keep working on basically strengthening that loop. … I think we can do the same thing but for all the news, like all the categories, everything. And I think that could be very, very large."「有时会出现这种情况:有人在 Twitter 上做出疯狂的预测或说了些蠢话,然后人们会去查看赔率,然后进行交易。存在这样一个循环,我认为我们需要继续努力加强这个循环。……我认为我们可以对所有新闻做同样的事情,比如所有类别、所有内容。而且我认为这可能会非常非常大。」Tarek Mansour · Kalshi CEO Tarek Mansour - The Future of Free Markets
华尔街买单的也是这套「下场取代空谈」的逻辑——顶级做市商 Susquehanna 的 Jeff Yass(转写作 Sasquander/Jeff Yates,原样保留):
"so Sasquander really believes, I mean, Jeff Yates really believes in the vision of sort of like this idea of like to have some skin in the game to forecast any question about the future instead of debating about it. And he's always loved prediction markets."「Sasquander 非常相信,Jeff Yates 也非常相信这个愿景,即为了预测未来,而不是争论它,应该参与其中。而且他一直很喜欢预测市场。」
Coplan 当面拒绝了这套词汇。 在他看来「群体智慧」是被用滥的玄学,真相机制其实是价格形成:
"So the net present value of that asset is equal to the likelihood of it happening. … because you get a dollar if it occurs, and if it's trading at 20 cents, people think it's 20% likely to happen. And the term wisdom of the crowds gets thrown around a lot. But really what it is is it's price formation."「因此,该资产的净现值等于它发生的可能性。…因为如果它发生,你会得到一美元,如果它的交易价格为 20 美分,人们认为它有 20% 的可能性发生。“群体智慧” 这个词经常被提及。但实际上,它是价格形成。」Shayne Coplan · Polymarket's Shayne Coplan & Axios' Dan Primack
他比 Mansour 多走的一步,是把「知情者进场」当成功能而不是丑闻。面对内幕交易质疑,他的回答是:真金白银的激励让知道答案的人把私有信息注入价格,这正是信息市场的意义——Gemini 3 发布日期市场被推到 95 美分(Google CEO Sundar 曾转发该市场):
"It goes back to this idea of information markets, right? Like, you think about when will Gemini 3 launch? Nobody is under the impression that nobody knows the answer, right? … And I think what's cool about Polymarket is that it creates this financial incentive for people to go and divulge the information to the market and the market to change and all of a sudden it's trading at 95 cents and people are like, oh."「这又回到了信息市场的概念,对吧?比如,你觉得 Gemini 3 什么时候发布?没有人会觉得没人知道答案,对吧?…我认为 Polymarket 很酷的地方在于,它为人们创造了这种经济激励,让他们去向市场透露信息,市场发生变化,突然交易价格达到 95 美分,人们会说,哦。」Shayne Coplan · Polymarket's Shayne Coplan & Axios' Dan Primack
在这批语料里,Mansour 从不碰内幕信息这个话题;他对「凭什么信」的回答是结构中立——20VC 快问快答里他补充的说法(转述):无庄家、中立撮合、赢输同费,结构上更像纽交所而非赌场。一个把可信度押在「谁都可以来定价」(参与),一个押在「知道的人会来定价」(信息)。同一个真相层,两种认识论——而这两种认识论对应着完全不同的媒体想象:前者像互动新闻,后者像彭博终端。
分歧线三:合法性从哪来——制度加冕 vs 采用即证明
两人都被扣过「赌博/操纵」的帽子,摘法完全不同。
Mansour 的路线是把每一步都做成制度证据。 行业第一家预测市场 Hedge Street 早在 2002 年创立、后被监管封杀,他吸取的教训是先赢制度:创立后前三年半全部投入监管工作,整场合法化斗争约五年半,牌照三连(交易所 2020、清算所 2024、经纪商 2025),直到 2024 年 10 月 7 日诉政府胜诉。最能说明「真相层」自觉的是选举市场干脆免手续费——KPI 不是收入,是证明机制安全有效:
"And we didn't charge fees on the election market. I don't know if you knew that. It was an intentional decision because the KPI here is we've got to show the world that these markets work. And it's done safely and nothing bad is going to happen. And democracy is going to be totally fine. And you'll see, we're announcing a few things with some big news organizations pretty soon. You'll see the shift. It's happened in culture where the level of embracing that's happening with prediction markets now, that's pretty nice to see."「我们没有对选举市场收取费用。我不知道你是否知道。这是一个有意的决定,因为这里的关键绩效指标是,我们必须向世界展示这些市场是有效的。而且它是安全进行的,不会发生任何不好的事情。民主将会完全没问题。你会看到,我们很快将与一些大型新闻机构宣布一些事情。你会看到这种转变。这种转变发生在文化中,现在对预测市场的接受程度正在提高,这真是太好了。」
对「赌博」标签的回应是定性切割:
"One of the misconceptions is people sometimes think that Kalshi is gambling and it's not."「其中一个误解是人们有时认为 Kalshi 是赌博,但它不是。」
这个定性已经被访谈双方共享——主持人 Harry Stebbings 直接替他接完了话头(引自同集逐字稿):
"It's information and news."「它是信息和新闻。」Harry Stebbings(主持人) · 20VC: Kalshi's $1BN Raise & the Battle to Replace Traditional Media
而对「赔率被操纵」的媒体指控,他给出的是 skin in the game 式的挑战——真信就下场,当时能赚 2.5 倍:
"I was like, if you really believe that this whole thing is false, like, You know, you should pony up or shut the fuck up. Like basically it was like, you know, you should put money where your mouth is. If you're so like, if you're so convinced that, you know, then you can make like a 2.5 X right now on your money. Like, what are you doing?"「我当时想,如果你真的相信整件事都是假的,你应该拿出真金白银,否则就闭嘴。基本上就是说,你应该言行一致。如果你如此确信,那么你现在可以赚取 2.5 倍的钱。你还在等什么?」Tarek Mansour · Kalshi CEO Tarek Mansour - The Future of Free Markets
Coplan 的姿态完全不同:不做制度化辩护,直接出示采用。 面对哥伦比亚大学教授关于 Polymarket 交易量中刷量交易占比 14%、5%、近期高点 20% 的报告,他的回应(据本场访谈转述)是数字很小、系竞争对手向记者兜料、不损害准确性可信度——没有 Mansour 式的免费市场实验或牌照叙事。他最强的证据是用户行为本身(战区信息源,见共识第三层),本场甚至出现了自指案例:主持人 Dan Primack 前一晚查的「Polymarket 2025 年在美国上线」市场定价 85%——记者当场用赔率当事实来源。
这条裂缝下面还压着一个两人都没正面处理的数字张力:Mansour 一边把 Kalshi 讲成新闻基础设施,一边在 Odd Lots 承认 NFL 周日体育约占总交易量 90%、选举日 99.9% 来自选举市场、周日对周一交易量差约 3 倍,且大多数市场(含谷物期货、股市)95% 以上的活动本来就是投机(他还引了零 DTE 期权散户成交量约万亿美元级作基准)。他的辩护不是否认,而是把波动本身论证成媒体属性——交易量是新闻议程的镜像(2024 年 10 月 Twitter 上 10 条推文约 9.5 条关于大选,与 Kalshi 交易量高度相关):
"So yes, NFL Sunday is massive in America and that's on top of people's minds. But when there's a Fed decision, that's on top of people's minds and that's where the trading volume is. When there's a Mamdani election, that's when the trading volume is. So the volumes will basically ebb and flow with whatever is in the news, whatever is trending on X, whatever is on top of people's minds."「所以是的,NFL 星期天在美国非常重要,那是人们最关心的事情。但是当美联储做出决定时,那是人们最关心的事情,交易量就在那里。当有 Mamdani 选举时,交易量就在那里。所以交易量基本上会随着新闻、X 上的趋势以及人们最关心的事情而起伏。」
都没说透的
1. 反方从未出场,「更准」停在断言。 四场访谈里,民调与新闻编辑部只以被转述的形象出现(「想说什么说什么」、Times 文章即「诽谤」)。最硬的证据——选举夜快 5-7 小时——证明的是结算速度,不是预测准确性;没有任何一位主持人追问系统性的校准数据(预测市场 vs 民调聚合的逐事件对照)。 2. 98-99% 的口径没人追问。 这是媒体论的承重墙数字,但「用户」如何定义、免费围观层与 1-2% 交易者如何在商业上自洽(选举市场 0 费率、体育占周日交易量约 90%),主持人都没有深挖。如果 98% 的用户是读者,Kalshi 到底是媒体(卖注意力)还是交易所(抽费)?两位创始人在不同段落里各取所需。 3. 内幕与操纵的边界被双双绕开。 Coplan 把知情者进场当功能(95 美分的 Gemini 3 市场),但事件合约没有上市公司内幕交易法的等价物,哪类私有信息可以合法变现无人追问;Mansour 三场访谈完全不触碰这个话题,「2.5 倍下场」的挑战也没有回答薄流动性市场里资金短期压价的可能。 4. 交易所自办新闻的利益冲突。 Mansour 明说「I think we can」自建新闻网络,但撮合方同时做报道者(影响叙事)与收交易流量(从叙事获利)的角色冲突,以及 CNN/CNBC 引用赔率时的编辑独立性与标注规范,四场访谈无人提起。
我的看法
以下是判断,不是事实。中等偏高把握:「赔率作为新闻源」已经从愿景变成事实——CNN/CNBC 官宣、Axios 开场加冕、战区用户自发采用、South Park 整集戏仿,这些是性质不同的独立信号,单靠供给侧吹嘘解释不了。中等把握:本主题标题里的「取代民调与评论」在这批材料里是纲领而非结论——被证明的是速度与参与感,不是准确性,在反方证词入库之前应当保持这个区分。较低把握的直觉:分歧线二比分歧线一重要——如果 Coplan 的「价格形成 + 知情者进场」才是机制真相,那预测市场的媒体化上限不取决于观众多少,而取决于多少新闻能被定义成可结算合约(Mansour 自己说目前只覆盖每天相关新闻的 15-20%),它的终态更像彭博终端而不是 CNN。
还想知道什么
- 民调方与新闻编辑部的第一手反驳:对「5-7 小时」与「想说什么说什么」的回应,以及预测市场 vs 民调聚合的系统校准对比(分事件类别的命中率/校准曲线)。
- 98-99% 非交易用户的测量口径与第三方验证:注册口径还是活跃口径、围观者留存如何——这个数字若坐实,「预测市场=媒体」就从比喻变成用户结构事实。
- CNN/CNBC 合作的合同形态:谁付谁钱、编辑控制权归谁、赔率上屏时对流动性深度与可操纵性的标注规范。
- 操纵与对冲的实证细节:哥大刷量报告的方法论与两家的市场监察机制;大额单边持仓对薄市场赔率的实际影响;以及 Mansour 说的 5000 万-1 亿美元门槛、2-3 年启动的机构对冲,有没有第一个可核实的落地案例。
取材
- 20VC: Kalshi's $1BN Raise, the Polymarket Feud, and the Battle to Replace Traditional Media · 入库 2025-12-09 ·
2c4ea6160e718109bbeadd5e57a370a5— Tarek Mansour × Harry Stebbings - Tarek Mansour on Kalshi's Plan to Create Markets in Everything · 入库 2026-01-20 ·
2eeea6160e7181c49d0ec4a6d58b7ccc— Bloomberg Odd Lots 现场(Joe Weisenthal & Tracy Alloway) - Kalshi CEO Tarek Mansour - The Future of Free Markets · 入库 2026-01-20 ·
2eeea6160e718199a391ef8e61a0d6af— 主持 Chris Power - Polymarket's Shayne Coplan & Axios' Dan Primack · 入库 2026-01-20 ·
2eeea6160e7181008c3afe23f99051dd— Axios BFD 现场(Dan Primack) 2eeea6160e7181cb8e78ce33de465682— CORE 判定,但系 20VC $1BN 同集的去重归档副本(interviews/archive/),未重复计入取材数;其中 Harry Stebbings「It's information and news」一条引文经该副本核对,出处为同集逐字稿- Inside Sequoia's Strategy: Alfred Lin on Company Building & Partnering Long-term · 入库 2025-10-31 ·
29dea6160e718133a464c2ee79535e5a— 擦边未纳入 - 20VC: Lessons from Alfred Lin and Ron Conway(Tarek Mansour 谈护城河与在位者) · 入库 2026-01-20 ·
2eeea6160e7181219afdf78fcb4bbcc3— 擦边未纳入 - 20VC: The Best AI Companies Have Unique Data Acquisition Strategies(Joon Sung Park, Simile) · 入库 2026-08-04 ·
3b2ea6160e7181798ff6dd1b24bed215— 擦边未纳入 - Google's AI Brain Drain, SpaceX's Huge Quarter, Airtable's 90% Collapse, US Data Fuels China AI · 入库 2026-08-10 ·
3b8ea6160e71816d82dded05af31e88e— 擦边未纳入