Prediction Market Sharps Outperform Wall Street on Mispriced Retail Bets

预测市场专业玩家利用散户错误定价跑赢华尔街 私密交易群体通过利用散户的党派偏见并构建专有数据集，跑赢机构预测并斩获可观利润。 主持人 Joe Weisenthal 和 Tracy Alloway 探讨了预测市场的内部运作机制，以及为何绝大多数散户参与者会在这种零和博弈中亏损。职业交易员 Brian Golden 和 Daniel Reichman 详细介绍了其私密 Discord 社群 MAGA Kiwi Club 如何通过定制选区级选举模型、重建美国劳工统计局（BLS）通胀公式以及开展一线实地调研，建立起可复制的优势。讨论还涉及了依赖大语言模型（LLM）进行交易的局限性、党派偏见对定价的影响，以及 Polymarket 和 Kalshi 等平台上面临高度关注的内幕交易案件。最后，主持人将当前预测市场的热潮与 2000 年代初的线上扑克繁荣进行了类比，强调持续的优势依赖于专有的人工数据而非公众共识。 与具有正和增长效应的传统股市不同，预测市场本质上是严格的零和博弈，每一笔收益都对应着对手方的等额亏损，绝大多数散户参与者最终均以亏损告终。 (0:00) 顶尖交易员通过 MAGA Kiwi Club 等私密 Discord 社群协作，以专人专精的“复仇者联盟模式”整合宏观经济模型与选区计票推演，系统性地从情绪化散户手中赚取收益。 (5:00) 独立交易员 Brian Golden 仅用 Excel 重构包含 200 多个细分项的劳工统计局（BLS）通胀计算公式，便在两年追踪期内取得了优于彭博华尔街机构共识预期的预测精度。 (20:00) Polymarket 数据显示平台 67% 的利润仅由 0.1% 的头部账户攫取，且经验丰富的专业交易员普遍将平台评论区的散户舆论倾向视为高度有效的反向指标。 (30:00) 资深玩家指出大语言模型（LLM）因容易盲目迎合提示词偏见而无法提供有效交易优势，真正的市场超额收益仍取决于地面一线调研与私有底层数据的挖掘能力。 (45:00)

While prediction markets are marketed to retail participants as accessible forecasting opportunities, they function as zero-sum environments where profits are heavily concentrated among a small cohort of disciplined 'sharps.' Traders Brian Golden and Daniel Reichman demonstrate that sustained edges stem from labor-intensive primary research, granular quantitative modeling, and contrarian positioning against ideological retail liquidity. Zero-Sum Structure and Profit Concentration (0:00) Unlike traditional equity markets that create aggregate value through corporate cash flows and dividends, prediction markets operate as strictly zero-sum environments where one participant's profit is another's direct loss. Co-hosts Joe Weisenthal and Tracy Alloway note that while consumer marketing portrays prediction platforms as straightforward venues to monetize common sense, unsophisticated retail traders overwhelmingly lose money over time. Citing a Wall Street Journal analysis of Polymarket, the speakers highlight that approximately 67% of all platform profits are captured by just 0.1% of accounts. To exploit this imbalance, sharp traders treat retail sentiment as a contrarian indicator. Golden and Reichman observe that prediction market comment sections reliably reflect 'dumb money,' where ideologically motivated retail bets create wide mispricings. By insulating themselves from partisan echo chambers and narrative-driven consensus, specialized traders extract steady returns directly from retail participants. Collaborative Research and Granular Modeling (5:00) Rather than operating in isolation, elite traders frequently collaborate within specialized syndicates such as the MAGA Kiwi Club, an invite-only Discord group structured around an 'Avengers model' where participants contribute distinct expertise in macroeconomics, monetary policy, or election modeling. To gain an informational advantage over public polling and cable news commentary, members commission private telephone surveys, conduct on-the-ground precinct checks, and build predictive baseline models long before public data releases occur. This methodology extends across political and macroeconomic domains. In elections, the group constructs custom precinct-by-precinct models that account for historical mail-in and late-arriving ballot distributions, allowing them to trade against panicked retail sentiment on election nights. In macroeconomics, Golden—despite having an academic background in drama rather than economics—reconstructed the Bureau of Labor Statistics' aggregation formula in Excel across more than 200 subcategories, achieving a lower average absolute error on inflation prints over a two-year period than the Bloomberg institutional consensus. Resolution Friction and Counterparty Asymmetry Operational discipline requires navigating structural platform differences and asymmetric risks. On platforms like Polymarket, settlement relies on decentralized oracle mechanisms that can trigger rule disputes, whereas Kalshi utilizes a centralized resolution authority. Traders manage capital efficiency by frequently selling winning contracts at 99 cents rather than holding to the full dollar payout, reallocating capital to higher-velocity trades when settlement horizons span weeks or months. Concurrently, market participants face counterparty hazards from informed insiders possessing non-public certainty. Golden and Reichman describe abrupt volume spikes—such as an entertainment award contract surging from $0.01 to $0.40 overnight—as clear indicators of insider presence. While federal enforcement has penalized high-stakes illicit trades, such as the arrest of a military member betting on an impending military invasion, traders must constantly evaluate who sits on the opposite side of thinly traded contracts to avoid being caught offside. The Limits of Quantitative Models and AI The guests caution against relying on Large Language Models for automated market forecasting, characterizing generic AI users as some of the most easily exploited market participants. Because LLMs exhibit severe sycophancy—shifting answers to accommodate user prompting biases—they fail as independent forecasting engines, leaving primary data collection as the only durable edge. However, purely quantitative historical models carry significant blind spots when detached from cultural realities. The Discord syndicate incurred a major loss during the 2025 Romanian presidential election after heavily backing candidate George Simion based on historical European runoff margins. Because their models relied on past statistical distributions rather than domestic sentiment, the group failed to foresee Simion's subsequent debate absences and public ridicule, allowing local Romanian bettors with ground-level context to take the other side of the trade. Drawing parallels to the 2000s online poker boom catalyzed by amateur Chris Moneymaker, the program questions whether an eventual washout of retail participants will leave prediction markets as a hyper-competitive arena dominated solely by institutional desks and sharps. The participants argue that the long-term regulatory legitimacy of the sector hinges on providing disciplined, accountable price signals on substantive political and macroeconomic events rather than facilitating speculative pop-culture novelty bets.

彭博社《Odd Lots》节目指出，与依靠底层经济增长与现金流创造净正收益的传统股票市场不同，预测市场本质上是残酷的零和博弈。尽管交易平台常通过营销将预测宣传为普通人利用日常知识即可轻松盈利的副业，但现实中极少数被称为“专业玩家（Sharps）”的交易者凭借高度细分的协作网络、自下而上的底层数据重构与一手调研，系统性地从散户资金与华尔街的分析偏差中攫取超额利润。 零和博弈与高度集中的收益分配 (0:00) 主持人 Joe Weisenthal 与 Tracy Alloway 指出，预测市场不存在股息或资本增值，每一个赢家的利润都对应着另一个参与者的等额亏损。尽管传统金融中的期权和期货也具备零和特征，但在散户高度参与的投机环境中，绝大多数非专业用户长期必然亏损。引用的《华尔街日报》数据显示，Polymarket 上高达 67% 的收益集中于仅仅 0.1% 的顶级账户，且随着萨斯奎汉纳（SIG）等华尔街量化机构设立专业做市席位，普通散户的生存空间被进一步挤压。 记者 Adam Iscoe 及其采访的专业交易员 Brian Golden 与 Daniel Reichman（网名“Carnitas Taco”）指出，平台广告宣称靠预测天气即可赚取房租，掩盖了市场残酷的专业化壁垒，使散户沦为待收割的“盲目流动性（Dumb Money）”。专业交易者甚至将平台的公开评论区作为极具价值的反向指标：越是散户群体群情激愤、大量留言争论的合约方向，往往越隐藏着可供专业资金反向套利的认知偏差与情绪溢价。 “复仇者联盟”协作与自下而上的数据解构 (5:00) 专业交易者能够持续维持认知优势，核心在于严密的去中心化协作与深度的底层微观建模。Golden 与 Reichman 所在的私密 Discord 社群“MAGA Kiwi Club”聚集了一批深耕不同领域的专业交易员。据 Golden 介绍，该团队采用类似“复仇者联盟”的运作架构，成员分别在宏观经济、货币政策、选举模型及体育等垂直领域建立独立模型并共享一手分析，同时各自执行独立的交易策略。 在宏观领域，曾主修戏剧的 Golden 花费三个月时间向美国劳工统计局（BLS）官员求教，在 Excel 中完整重构了包含 200 多个细分商品类别的消费者价格指数（CPI）核算模型。通过结合高频公开物价数据与前瞻趋势，Golden 在过去两年的平均绝对预测误差持续优于彭博一致预期（Bloomberg consensus）。他认为，这揭示了华尔街投行预测的虚浮——机构分析师无需为预测失误承担资金后果，而预测市场参与者却面临真金白银的强制纠错惩罚。在政治预测中，该群体同样避开电视民调，转而构建细化至选区（Precinct-level）的计票进度模型与自费实地抽样，在开票夜利用公众对初期选票的非理性恐慌实现精准收割。 资本流转效率、裁决风险与内幕交易博弈 在交易执行层面，专业选手的持仓逻辑高度取决于资金周转效率与平台规则机制。例如在合约胜率达到 99% 时，交易者是否选择持有至 100% 结算，完全取决于到期时间与机会成本：若合约在 48 小时内结算，持有是合理的；但若需耗费数周乃至一个月，交易者更倾向于提早以 0.99 美元清仓并投入更高周转的标的。此外，不同平台的结算机制带来不同风险，Polymarket 依赖去中心化预言机（Oracle），但实际常受制于平台官方的补充解释，在冷门标的上极易引发规则争议；而 Kalshi 则采取中心化官方裁决。 内幕交易同样是不可忽视的对手方风险。经验丰富的交易者常通过异常爆发的成交量来识别内幕线索，例如在评论家选择奖颁奖前夕，演员雅各布·艾洛蒂（Jacob Elordi）的最佳男配角合约曾在一夜之间从 0.01 美元被巨量推高至 0.40 美元。尽管监管执法正在升级——如谷歌员工利用年度搜索趋势内幕交易被解雇、美军人员因涉嫌利用军事打击行动下注近 40 万美元而被捕——但在涉及国会人事投票确认等非传统内幕灰色地带，交易员仍需时刻评估对手方背景并主动控制敞口以防被对手方彻底击穿。 统计模型的致命盲区与人工智能的无效性 尽管严谨的量化模型构筑了核心优势，但过度迷信历史统计规律往往会导致灾难性翻车。在 2025 年罗马尼亚总统大选中，候选人乔治·西米翁（George Simion）在重选的首轮中大幅领先近 20 个百分点。欧洲历史选举数据表明此类优势极难被翻盘，促使该交易团队重仓押注西米翁胜选。然而，西米翁随后缺席辩论、离开该国并在本土媒体舆论中彻底丧失支持。由于模型未能捕捉到当地真实的社会文化拐点与底层情绪，该团队遭遇了惨重亏损，反被罗马尼亚本土选民的真实认知所击溃。 此外，对于当前部分交易者试图依赖生成式大语言模型（LLM）获取交易优势的尝试，专业选手将其定性为“市场中最愚蠢的资金”。交易员虽然普遍使用 Claude Code 等工具协助编写底层代码和清洗数据集，但若直接依赖 LLM 预测事件走向则完全无效。大模型存在严重的阿谀奉承（Sycophancy）缺陷：当用户就通胀预测向 ChatGPT 施加质疑时，模型会顺从用户的引导不断推翻自身结论并给出看似合理的修饰理由。由于缺乏独立的一手私有数据支持，通用 AI 模型在预测市场中根本无法形成击败市场共识的阿尔法收益。 节目总结指出，当前的预测市场繁荣正在重演 2000 年代初的线上扑克繁荣（Poker Boom）：业余神话吸引了大量盲目散户入场，但随着散户流动性被残酷吸干，市场终将演变为专业选手与量化机构之间高烈度的存量博弈。在通用人工智能无处不在的时代，真正的交易优势依然牢牢锚定在现场深耕调研、私有数据构建以及对本土社会情绪转折的敏锐洞察之上，纯粹依赖公开数据与滞后统计模型的套利空间正在迅速收窄。

## Background 本期节目由彭博社（Bloomberg）旗下的《Odd Lots》播客于 2026 年 7 月 6 日在 RSS 及各大网络平台发布。节目简介指出，预测市场充斥着纯粹的投机与赌博，散户入场极易亏损，但本期邀请了来自私密 Discord 群组“MAGA Kiwi Club”的职业交易员 Brian Golden、Daniel Reichman 以及在《纽约时报杂志》报道该群体的记者 Adam Iscoe，共同探讨他们如何在政治与宏观经济预测市场中寻找错误定价并斩获真金白银。 ## TL;DR 节目深度拆解了预测市场的底层现实与盈利逻辑，指出其本质是残酷的零和博弈，平台宣传的“轻松致富”只是收割散户的营销幻象。来自民间顶级私密交易圈的嘉宾披露了其超额收益的来源：并非依赖玄学或黑箱 AI，而是靠自下而上复刻官方通胀公式、自费民调乃至实地地推敲门等极其繁重的基本面硬核调研。对话还客观探讨了预测市场中的内幕交易灰色地带、在罗马尼亚大选中因脱离本土现实而翻车的惨痛教训，并类比 2000 年代初的线上德扑热潮，警示随着散户“笨钱”被洗净离场，市场终将滑向高手互砍与平台抽水的终局。 ## Atmosphere & style 整场对话在严谨专业与幽默风趣之间保持了绝佳平衡，既有华尔街财经对谈的犀利审视，又带着地下极客圈子拆解复杂谜题的松弛感与草根生命力。两位交易员毫不避讳地披露底层交易细节与翻车惨案，语气坦诚且极具去神圣化意味；主持人的层层追问与即兴调侃则让讨论步步深入，不仅戳破了虚假的致富神话，更让听众切实感受到真实交易前线那种既充满硬核求真执念、又时刻面临不确定性绞杀的智力博弈张力。 ## Bulletin - `33:30` 预测市场本质为无底层现金流支撑的零和博弈，《华尔街日报》数据显示 Polymarket 上高达 67% 的利润被仅仅 0.1% 的顶级交易账户瓜分。 - `23:15` 民间高频交易者组织“MAGA Kiwi Club”通过私密协作与自下而上的数据重构击败华尔街机构，其自建通胀模型的预测精度甚至持续优于彭博宏观共识预期。 - `30:00` 政治类合约长期因散户情感投入与社交媒体信息茧房而存在严重错误定价，资深交易员将公共评论区普遍视为高胜率的反向指标。 - `48:00` 大语言模型在预测交易中极易出现迎合性与顺应幻觉，当下的核心阿尔法壁垒已转向实地敲门、专属电话民调等线下原生感知与私有数据获取。 ## Key takeaways - `00:00` 预测市场与传统股票市场具有根本差异，缺乏底层资产现金流支撑，属于绝对的零和博弈与点对点对手盘掠夺。 - `05:00` 私密 Discord 社群“MAGA Kiwi Club”采用类似多策略对冲基金的分布式架构，各细分领域交易员独立出资执行，同时共享深度阿尔法信息。 - `10:00` 预测市场的核心价值在于建立犯错必须付出真金白银代价的硬惩罚机制，从而打破了传统媒体领域专家信口开河而无须担责的文化。 - `13:00` 政治选举关乎情感与意识形态，极化的社交媒体信息茧房让散户对胜负产生严重脱离算术常识的幻觉，为职业交易员创造了巨大的逆向套利空间。 - `15:00` 顶尖选情研判团队通过自费委托专属电话民调、实地挨家挨户敲门走访以及建立选区分块实时计票动态模型，构筑了业余投机者无法企及的信息护城河。 - `20:45` 戏剧专业出身的交易员布莱恩·戈登在 Excel 中完整复刻劳工统计局包含 200 多个子类别的底层权重公式，以纯测算方式在通胀预测上持续战胜华尔街顶级机构。 - `23:15` 传统大型投行在宏观指标预测上过度依赖惯性预期与自上而下的模型，反而暴露出缺乏底层细项拆解基本功的懈怠与软弱。 - `26:15` 在胜率达到 99% 的尾部合约上，交易员必须在“锁定最后 1% 到期收益”与“规避漫长结算时间带来的资金占用及平台规则争议风险”之间权衡机会成本。 - `30:00` 由于没有持仓上限，顶级选手绝不会在公共讨论区分享底牌，充斥单边狂热情绪的评论区因此成为胜率极高的反向交易指标。 - `35:00` 2024/2025 罗马尼亚大选案例表明，即便拥有顶尖的跨国历史统计模型，缺乏对本土舆论裂变与现实情绪的现场感知，依然会遭遇灾难性惨败。 - `40:00` 预测市场存在着显著的内幕交易灰色地带，职业操盘手必须时刻警惕异常放量的盘口，并通过评估对手盘是否掌握非对称内部信息来严格控制仓位。 - `48:00` 大语言模型（LLM）具有极强的迎合性与顺从幻觉，无法生成原生阿尔法，直接询问 AI 胜率是散户在预测市场快速亏损的典型特征。 - `51:41` 预测市场正面临类似 2000 年代初线上德州扑克热潮的演变宿命：一旦散户亏光本金退场，市场将沦为高手互搏与套利空间收窄的牌局，最终仅有平台抽水稳赚。 - `53:07` 在 AI 驱动的信息过载时代，预测市场最稀缺的交易优势是无法被大模型从后视镜总结的线下真实感知、实地走访与一手人际连接。 ## Detailed walkthrough ### `00:00` 预测市场的本质：零和博弈与“遍地黄金”的营销幻象 节目开场由赞助商播报（4imprint 促销定制、The Hartford 商业风险管理、IBM 企业级 AI 应用）切入正题。主持人 Joe Weisenthal 坦言自己持有一种非常“婴儿潮一代（Boomer）”的保守观点：他打心底里厌恶任何主打“快速致富”或将股票交易高度游戏化（gamification）的套路，更反感那些暗示“人人都可能成为赢家”的博彩广告。Joe 特别提到《华尔街日报》近期关于 Polymarket 等平台广告的深度报道——这些平台试图兜售一种“遍地都是免费横财、只要押注你懂的日常知识（比如降雨量预测）就能赚大钱”的幻想，这在直觉上令他感到生理性不适。 Tracy Alloway 随即指出，预测市场与传统股票市场之间存在着根本性的结构差异：预测市场是绝对的“零和博弈”（zero-sum game）。在股票市场中，投资者可以通过底层公司的自由现金流、股息分红和整体经济增长实现共赢；但在预测市场里，任何一个人赢走的每一分钱，都必定直接对应着坐在屏幕另一端的某个人所蒙受的等额亏损。Joe 赞同这一观点并补充道，传统金融（TradFi）里的期货、期权和掉期也属于零和博弈，但那本质上是交易与投机，绝非真正的价值投资。不管是去 Robinhood 上跟风炒期权、涌入加密货币市场，还是在预测市场里下注，绝大多数散户怀揣着发财梦入局，最终的结局几乎都是“输得底裤都不剩”。 - **交易与投资的界限**：预测市场纯属点对点或与做市商博弈的零和游戏，缺乏底层实体资产的现金流支撑，投机性极高。 - **散户的错觉**：平台广告通过宣传“用你身边的认知赚钱”降低心理门槛，掩盖了大多数业余参与者沦为对手盘“流动性养料”的残酷现实。 --- ### `05:00` 揭秘“MAGA Kiwi Club”：战胜华尔街的民间“复仇者联盟” 随着华尔街专业机构（如 Susquehanna / SIG）纷纷设立预测市场交易专柜，市场上出现了一种担忧：机构的入场是否会迅速抹平普通人赖以生存的信息优势？但在《纽约时报杂志》刊登的长篇专题报道中，记者 Adam Iscoe 记录了一帮深藏不露的民间高频交易者（Sharps）。Joe 调侃道，如果自己现在充值 1000 美元进去，这笔钱很快就会被抽水和顶级高手瓜分完毕。本期节目邀请到了报道作者 Adam Iscoe，以及来自神秘私密 Discord 群组“MAGA Kiwi Club”的核心交易员：专精于宏观经济与通胀数据的 Brian Golden，以及网名为“Carnitas Taco”、擅长选举政治预测的 Daniel Reichman。 记者 Adam 坦承，自己之所以展开这项调查，是因为自己此前亲自入场下注并赔了不少钱。在沮丧之余，他极度好奇这些钱究竟流向了何方——对手盘到底是像 SIG 这样财力雄厚的机构做市商，还是潜伏在网络角落的绝顶聪明人？顺着资金与社群的线索，他最终找到了这个战绩彪炳的 Discord 小圈子。Brian 与 Daniel 随后澄清，“MAGA Kiwi Club”这个名字完全是一个反讽意味的群内梗（inside joke），与真正的政治立场无关。群组的协作模式极像一家分散式的多策略对冲基金（multistrat hedge fund），又像是一个“复仇者联盟”： ```mermaid flowchart TD accTitle: MAGA Kiwi Club 协同交易架构 accDescr: 展示 Discord 社群中不同专业领域的交易员如何交换信息并独立执行投资组合。 subgraph Experts["各领域专业交易员"] B["Brian Golden<br/>通胀与宏观经济"] D["Daniel Reichman<br/>选举与政治博弈"] O["其他核心成员<br/>体育、流行文化、科技"] end subgraph Hub["MAGA Kiwi Club (Discord 私密社群)"] H["高价值阿尔法信息交换<br/>& 逻辑交叉验证"] end subgraph Portfolio["独立决策与执行"] P1["个人独立投资组合与 P&L"] P2["预测市场合约交易<br/>(Polymarket / Kalshi / PredictIt)"] end B --> H D --> H O --> H H --> P1 P1 --> P2 ``` Brian 解释说，每个人管理着独立的资金与投资组合（P&L），但大家整天泡在群里深度交流。如果他在通胀走势上为 Daniel 提供坚实的数据分析与前瞻线索，Daniel 就会在政治合约等领域回馈同等深度的洞见。Daniel 补充表示，群内许多成员的交情可以追溯到 2016 至 2017 年 PredictIt 独大的时期；大家聚在一起的根本动力不是纯粹为了赌博，而是出于拆解选举机制、挖掘客观事实、共同破解复杂逻辑谜题的热情。即便如今市场的盘子与奖金规模呈指数级膨胀，这种彼此借力、精益求精的研究底色依然未曾改变。 - **特邀嘉宾背景**： - **Adam Iscoe**：纽约记者、播客制片人，在《纽约时报杂志》发表预测市场高手专题报道后近期入职协作软件公司 Notion。 - **Brian Golden**：资深交易者，擅长美国通胀（CPI 等）指标相关合约的定价与利差捕捉。 - **Daniel Reichman（Carnitas Taco）**：专精选情研判与政治细分事件合约的资深操盘手。 - **历史渊源**：团队核心成员自 2016-2017 年 PredictIt 时代即开始结识，历经数个美国大选周期锤炼，形成了极强的互补分工体系。 ### `10:00` 走进神秘的预测市场交易群 记者亚当·伊斯考（Adam Iscoe）分享了他最初结识布莱恩·戈尔登（Brian Golden）与丹尼尔·赖希曼（Daniel Reichman）的经历。这一切始于他结识了一位对选举极度痴迷的交易员，在长达约六周的交流与信源挖掘之后，对方才透露自己根本不是一个人在地下室孤独地建立数学模型，而是在与许多“同事”紧密协作。当伊斯考惊讶地问对方是否供职于某家对冲基金时，交易员笑称他们其实只是聚在一个 Discord 私人服务器里。播客主持人乔·韦森塔尔（Joe Weisenthal）打趣地追问：“大家私下问一句，这个群是不是 100% 都是男性？”得到了干脆的确认：“是的，毫无疑问。” ```mermaid flowchart TD accTitle: 记者挖掘预测市场交易群的过程 accDescr: 展现从与个别交易员接触到发现背后 Discord 社区的递进关系 A["结识痴迷政治选举的交易员"] --> B["长达 6 周的深度交流与信源跟进"] B --> C["交易员揭示协作内幕：并非单打独斗"] C --> D["锁定核心圈子：Maga Kiwi Club Discord 私人社区"] ``` 韦森塔尔随后从“有效市场假说（EMH）”的角度抛出核心疑问：传统金融学认为各类资产价格大多已被充分定价，即便股票市场被普遍认为有效，仍有巴菲特或顶级多空对冲基金能持续赚取超额回报；在预测市场中，媒体如今越来越频繁引用其给出的赔率和价格，这些价格究竟是高质量的情报信号，还是充斥着大片未经修剪的“低垂果实”？ 交易员们坦言，如果预测市场真的完全有效，群里的成员绝不可能取得如此惊人的超额回报。所谓“预测市场是未来、比所有传统机构都更准确”的说法，很大程度上只是狂热布道者的包装，实际上大量标的价格离谱至极。预测市场的真正核心价值在于“犯错要付出真金白银的代价”。当今社会充斥着各种冠以“专家”头衔的权威，他们在媒体上信口开河、误导公众，却完全不需要为错误承担任何经济后果；而预测市场建立了一套赏罚分明的反馈机制，输家割肉，赢家收钱。如果经济和政治媒体领域也能引入这种惩罚机制，整个舆论的专家文化将会健康得多。 ### `13:00` 政治信息茧房与离谱的错误定价 丹尼尔和布莱恩深入分析了为何政治类预测市场会长期存在巨大的错误定价。与通货膨胀等宏观经济指标不同，通胀数据只是冷冰冰的数字（比如大家不会在情感上非要执着于“3.6% 比 3.8% 好”），但政治完全关乎“心智与情感”，极易受制于信息茧房效应。政治选举天然具有零和博弈属性，两派竞选团队砸入数以亿计的资金，结果只有一人胜出、另一人淘汰；而在高度极化的政治生态中，两派选民往往生活在完全隔绝的信息世界中，对现实的理解截然相反。 ```mermaid flowchart LR accTitle: 政治预测市场失真机制 accDescr: 解释为何政治博弈容易被意识形态与信息茧房扭曲导致错误定价 subgraph S1["选民心理与环境"] A["意识形态与情感投入"] --> B["极化社交媒体信息茧房"] end subgraph S2["市场反馈"] B --> C["片面证据被误认为绝对真理"] C --> D["出现严重偏离现实的荒谬赔率"] end subgraph S3["专业交易员（Sharps）操作"] D --> E["依靠硬核历史数据与数学模型重仓狙击反向套利"] end ``` 布莱恩举了洛杉矶市长初选的经典案例：保守派媒体阵营乃至部分主流媒体不断造势，甚至开始认真探讨斯宾塞·普拉特（Spencer Pratt）不仅能挺进决选、而且真能拿下洛杉矶市长宝座的可能性。很多人沉迷在特定推特账号和偏狭媒体推送的闭环里，误以为身边所有人都支持他，营造出一种胜券在握的虚假繁荣。然而从数学常识来看，洛杉矶是一个民主党占绝对优势（D+42）的城市，共和党人根本没有赢的算术空间。当时群里几乎所有精明的交易员（Sharps）都建了这辈子最大的仓位之一，疯狂押注普拉特绝不可能当选。 - **市场零和特性**：两党投入巨资互搏，选民各持己见，极端的认知割裂促使双方愿意真金白银为偏见下注。 - **信息茧房效应**：社交网络算法不断投喂单一立场的“充沛证据”，让散户产生严重脱离现实的幻觉。 - **经济对比政治**：宏观经济指标（如通胀）缺乏强烈的情感认同，定价效率正在迅速收紧并提高，而政治类竞猜由于情感和信仰作祟，将长期是错误定价最为严重的领域。 ### `15:00` 职业团队如何建立真正的“信息护城河” 当主持人询问他们的核心交易优势（Edge）究竟是什么、是否只是一套美化了的“幸存者偏差”叙事时，丹尼尔与布莱恩披露了他们团队在幕后所付出的高强度专业工作。真正的优势并非靠直觉或运气，而是极其严苛的数据量化、对历史规律的敬畏，以及时刻准备推翻自己预设的开放心态。每当押注获胜时，他们绝不会只庆祝胜利，而是会精确复盘：自己最初评估的概率分布是多少？实际胜率和赔率偏差有多大？ 随着市场资金体量和交易规模水涨船高，团队的调研手段也全面升级： 1. **委托专属电话民调**：在得克萨斯州初选期间，群内团队自行出资委托专业的电话民意调查，获取一手原始数据。 2. **实地地推敲门调研**：部分核心成员甚至亲自奔赴选区一线，展开地毯式的挨家挨户敲门走访，获取传统民调无法捕捉的民情动态。 3. **选区分块实时动态模型**：他们针对不同选区的投票历史、邮寄选票与现场选票的到达时序建立精细模型。 布莱恩补充道，在洛杉矶市长选举开票当晚，候选人妮蒂亚·拉曼（Nithya Raman）因初步开票落后，甚至在支持者面前落泪以为大势已去；然而丹尼尔和他们的选举团队通过掌握加州后期计票偏向的历史规律与精细的分区模型，早在竞选团队自身反应过来之前，就断定她依然能成功杀入前两名，并在实时开票过程中果断下注反向重仓。这种基于进场数据实时解构选情的能力，构成了业余投机者难以企及的技术壁垒。 ### `20:45` 选举夜的噪音与通胀预言家：一位戏剧系毕业生的模型 在聊到预测市场的价格效率时，主持人指出，虽然理论上价格应具备一定程度的有效性，但在关键的选举之夜，市场往往充斥着由开票节奏引发的巨大噪音。例如在近期加拿大总理选战中，由于计票初期几个并不具备代表性的选区先开出结果，博弈盘面上保守党领袖皮埃尔·波利耶夫尔（Pierre Poilievre）的胜率瞬间出现非理性的暴涨。这种剧烈晃荡的行情甚至让主持人都怀疑，媒体到底该不该在报道中引用这些赔率数据。 然而，真正让专业圈子震惊的是亚当·艾斯科（Adam Iscoe）在《纽约时报杂志》报道中披露的一位人物——布莱恩·戈登（Brian Golden）。一些专业经济学家甚至将他的通胀预测模型惊呼为“通胀界的诺查丹玛斯”（Nostradamus of inflation）。财经名笔马特·莱文（Matt Levine）也在专栏里困惑地写道：如果一个人拥有神准的通胀预测能力，为什么他没有被华尔街挖走，去掌管一个动辄百亿美元规模的利率对冲基金？布莱恩听到后幽默地澄清说：“说明一下，根本没人给我打过电话招聘我。”随后他坦诚透露了自己的底细：他其实大学本科读的是戏剧专业，自嘲只是一个“来自中西部的戏剧少年”（theater kid from the Midwest），对宏观经济理论的了解可能比这档播客请过的任何一位嘉宾都要少。 布莱恩的“秘诀”极其质朴却耗时费力。他发现美国劳工统计局（BLS）计算通胀数字并非一个抽象黑箱，而是由一套精密的数学公式驱动。整个商品及服务篮子（basket of goods）包含 200 多个细分子类别，每个类别在总指数中所占的权重每个月都会根据美国人实际支出的变动进行调整。布莱恩花了整整三个月时间，在 Excel 中完整复刻了 BLS 的底层计算公式。在这一过程中，许多热心的 BLS 公务员不厌其烦地接听他的电话，帮他梳理清楚了所有数学细节。搭建好底层公式后，他再对各个细项价格进行拆解预测：有些类别（如汽油、天然气、二手车）能拿到高频公开数据，而其余细项则主要是基于过去六个月的趋势进行推演与推断。 ```mermaid flowchart TD accTitle: 布莱恩·戈登的通胀自下而上建模流程 accDescr: 展示布莱恩如何从拆解BLS底层公式到收集细项数据再到预测通胀的完整流程 subgraph FormulaReconstruction ["第一阶段：重构官方计算公式"] A["研读美国劳工统计局（BLS）公开方法论"] --> B["致电BLS公务员核实数学权重与计算逻辑"] B --> C["在Excel中复刻包含200多个子类别的商品篮子"] end subgraph DataGathering ["第二阶段：多渠道输入数据"] D["高频公开数据<br/>（汽油、天然气、汽车等）"] --> F["细项价格趋势外推与拟合"] E["过去6个月历史数据与趋势"] --> F end subgraph OutputAction ["第三阶段：生成预测与市场套利"] C & F --> G["输出最终CPI预测数值"] G --> H["对比彭博机构共识预期"] H --> I["在预测市场提前建仓获利"] end ``` --- ### `23:15` 华尔街机构预测的软肋与神枪手们的土办法 布莱恩直言不讳地指出，他的成功其实并不说明他有多聪明，反而暴露了传统投行与机构分析师的“软弱与懈怠”。因为预测通胀在本质上甚至算不上真正的“先知预测”，它完全是对已经发生过去的客观数据的测算。令他难以置信的是，那些为数以百亿计资产提供投资咨询建议的顶级金融机构，在这项基本功上居然表现得如此业余。 布莱恩摆出了冰冷的数据事实：过去两年里，他仅凭一个自制的 Excel 表格测算出的通胀平均绝对误差（average absolute error），竟然持续优于彭博的宏观共识预期（Bloomberg consensus）。他感叹道，机构拥有近乎无限的算力、人力和专有数据渠道，却往往不去下这种笨功夫。更致命的是，专家预测构成了整个市场的预期锚点；在官方 CPI 数据出炉前 20 分钟，彭博终端便会打出类似“受高盛、摩根大通、美银预期影响，通胀将显示为某某数值”的定调标题，仿佛一切已成定局。一旦最终数据爆冷，财经媒体又会若无其事地撤换标题，而整个传统金融市场早已被这种拙劣的预先预期裹挟着剧烈动荡。 播客主持人和记者亚当·艾斯科补充指出，像布莱恩这样从底层自下而上（bottoms-up）钻研公式的还有业内顶尖专家奥马尔·谢里夫（Omar Sharif）。更广泛地说，预测市场上那些真正能稳定盈利的“神枪手”（sharps）无论主攻哪个领域，本质上都依赖这种高密度的原始调研： - 预测纽约下雨等天气类别的交易员，会追着一线气象学家和地球物理学家频繁通话； - 搞政治与政策预测的套利者，经常潜伏在前方现场，甚至常年与彭博社前方记者保持电话热线互换信息； - 这种在 prediction markets 里大赚特赚的硬核交易员，其实践方式在很大程度上重现了顶尖基本面宏观对冲基金的调研纪律，而非单纯在网页上凭感觉点选赔率。 --- ### `26:15` 结算机制分歧、机会成本与 99% 胜率的取舍哲学 随后，对话切入到预测市场的微观交易战术与结算机制设计。针对不同平台在事件发生到最终结算（resolution）之间的流程差异，主持人抛出了一个核心问题：当一个确定性事件基本尘埃落定、赔率已经攀升至 99% 时，交易者到底应该立刻平仓套现寻找下一个机会，还是咬牙拿着头寸赚取最后 1% 的到期利润？ 布莱恩从资本的时间价值和机会成本角度做出了解释。他指出，如果一个市场在 48 小时之内就能正式结算清算，他通常会选择一直持有到终点；但若是一场选举结果虽然已经众所周知，官方结算却需要拖延整整一个月，那么把 99 美分换成 1 美元的年化回报率其实非常低，他完全可以在别的高胜率标的上用这笔流动性赚到更多钱。 ``` 【尾部利润与流动性权衡】 持有至到期（99¢ -> 100¢）： - 优点：锁定确定性最终收益。 - 缺点：占用资金仓位，承担结算延期与平台仲裁风险。 提前平仓套现： - 优点：释放资本流动性，避免规则争议与黑天鹅裁决。 - 缺点：折价让渡最后 1% 利润给承接方。 ``` 丹尼尔·赖希曼（Daniel Reichman）则从平台结构性风险的角度给出了更深一层的分析。他强调，交易员必须敏锐区分：市场盘口显示为 98% 或 99%，究竟仅仅是因为漫长的结算等待时间造成的折现，还是客观上仍存在 1% 到 2% 的翻盘概率或规则纠纷风险？丹尼尔对不同平台的结算机制表达了强烈观点： - **Polymarket 的机制漏洞**：丹尼尔批评 Polymarket 的系统存在很大问题，其去中心化预言机（Oracle，即 UMA 机制）在很多场景下被架空削弱，大量市场最终演变为取决于平台官方运营方如何“补充解释规则”的人为裁量争议，容易陷入旷日持久的规则混战。 - **Kalshi 的中心化透明度**：相比之下，他对受到监管的 Kalshi 结算更为满意，尤其是在内阁成员提名确认、大选胜负等界限分明的标的上，几乎没有扯皮空间。 - **实战持仓考量**：丹尼尔透露，自己当时正持有一笔秘鲁大选的头寸，甘愿从 99% 一直守到 100%，原因仅仅是“目前手上还没看到其他更好的资金去处”。布莱恩随即笑着插话打趣：“等我们录完这期节目赶紧私聊，我手头正好有几个绝佳的投资标的推给你。” --- ### `29:00` 预测市场与传统民调：真有先见之明，还是纯粹更新更快？ 在探讨概率变动的深层逻辑时，主持人抛出了预测市场领域经久不衰的核心争论：在预测选举结果方面，预测市场给出的赔率概率，究竟是真正意义上比传统民调更科学、更精准，还是说它们只是因为开票信息传来时能够以极快的速度刷新报价，从而在表面上制造出了“算得更准”的假象？ 丹尼尔给出的答案是：预测市场确实能产出更优的真实概率，但这高度取决于市场两端被注入的真实交易资金规模与深度。为了论证这一点，他特别提到了去年 11 月的新泽西州州长选举作为典型案例。在那场选战前夕，几乎所有传统主流民意测验均一口咬定这是一场咬得极紧、差距仅在两到三个百分点以内的拉锯战，而严肃的预测市场盘口却通过资金博弈给出了完全不同的概率分布，真实地反映出了深层选情的基本面。 ### `30:00` 散户流动性扭曲定价与“反向评论区”指标 即便专业交易员拥有更优越的模型，庞大的散户资金流依然能将市场价格带偏。在复盘某场关键竞选时，群里有位成员甚至没有建立复杂的量化模型，仅仅是在民调出炉前三个月，在餐巾纸上凭借先验条件草草算了一笔账：考虑到以往胜选优势、特朗普当时支持率下滑了八九个点、以及执政背景的变化，民主党候选人大概率会以 14 个百分点左右的绝对优势胜出。然而随后出炉的传统民调却普遍显示选情极其胶着，几乎没有机构预测会出现巨大分差。虽然群内的专业选手并没有无视民调，而是试图对数据进行标准化校准，但他们深信这场选举的实际差距在 12 到 14 个点左右。然而，由于对立面涌入了海量散户资金（square side），预测市场的价格根本无法完全纠偏——市场给出的胜幅预期虽略高于传统民调，但依旧远低于最终的实际战果。类似的情况也发生在其他冷门候选人身上，即便真实胜率不足 5%，狂热的市场情绪仍能将其推高到 27%。 这引出了预测市场交易中一个经典的反向指标：公开评论区。彭博社主播提到了关于经济学家艾伦·科尔（Alan Cole）的一篇报道，科尔押上了全家积蓄做空埃隆·马斯克（Elon Musk）所谓大幅削减联邦赤字的豪言。当时文章中最经典的一句引述来自科尔的妻子：“我翻看了预测市场的评论区，感觉里面的人全是一帮白痴——至少跟我丈夫比起来是这样。因此，我非常支持他押上全家的积蓄去赌这一单。” 针对这一现象，嘉宾们异口同声地给出了交易座右铭：“永远和评论区反着来（Always bet against the comment section）。”严肃操盘手（sharps）通常保持沉默，只在私密圈子里交换意见；当公开讨论区里充斥着单边倒的狂热言论时，往往预示着那是错误的方向。在早期平台 PredictIt 时代，由于单项合约有 850 美元的严格持仓上限，许多选手建满仓位后乐于在社区分享真知灼见；但在 Kalshi 或 Polymarket 这种没有实质性头寸限制的平台上，信息保护成了头等大事，顶级选手绝不会跑到公共留言板上去好心点醒对手。 ```mermaid flowchart TD accTitle: Prediction Market Sentiment vs Smart Money accDescr: Flow showing how retail sentiment skews market odds while sharps keep edge private. subgraph RetailFlow["散户与评论区动向"] R1["狂热情绪与直觉下注"] --> R2["公共讨论区单边造势"] R2 --> R3["散户资金推偏盘口价格"] end subgraph SharpsFlow["专业选手 (Sharps) 应对"] S1["独立模型与底牌分析"] --> S2["私密 Discord 研判"] S2 --> S3["视公开评论为反向指标"] S3 --> S4["利用无头寸限制大笔吃单"] end R3 -.->|"提供流动性/送钱"| S4 ``` ### `33:30` 德扑热潮的宿命与“鱼塘”终局思考 随着《华尔街日报》披露在 Polymarket 上高达 67% 的利润被仅仅 0.1% 的账户掠走，加之华尔街专业机构的进场，一个残酷的疑问摆在了所有人面前：当大量散户输光本金离场，这个市场是否会不可避免地“变硬”（harder）？ 这种演变路径在博彩和交易历史上屡见不鲜： - **线上德扑热潮（Online Poker Boom）**：早期到处都是容易赢钱的“软桌”（soft tables），但随着韭菜被收割殆尽，牌桌上最终只剩下职业“鲨鱼互咬”（sharks vs. sharks），利润全被平台抽水磨损。 - **每日梦幻体育（DFS）**：早期淘金者轻松斩获暴利，随后算法化与职业化迅速挤压了普通玩家的生存空间。 - **高额盲注贵宾厅逻辑**：线下赌场的高额限注区必须有挥金如土的名流或中东富豪愿意送钱，职业牌手才能获利；如果满桌全是职业高规格选手，最终谁也赚不到钱。 面对这种担忧，嘉宾们承认某些特定金融或量化细分市场的套利空间确实在收窄。但在政治选举领域，主导盘口的依然是极度情绪化的话题与个人偏见（"vibes and emotions"）。由于预测市场仍处于早期破圈阶段，源源不断的新散户仍在被引入，至少就目前而言，政治盘口依然是一个充满定价漏洞且“极易被击败（pretty beatable）”的温床。 ### `35:00` 翻车惨案：2024/2025 罗马尼亚大选中的“本土认知差” 即便是顶级团队，也无法完全摆脱认知盲区带来的惨痛代价。当主持人问及 2024/2025 年罗马尼亚总统大选时，丹尼尔·赖希曼（Daniel Reichman）苦笑着坦言那是其交易频道历史上“极其黑暗的一天（a dark day for our channel）”。 这场灾难源于模型与本地现实的致命脱节： - **历史模型的诱饵**：在罗马尼亚总统大选首轮投票因涉嫌外部干预被判定无效并重新举行后，极右翼候选人乔治·西米昂（George Simion）在第一轮中以大约 20 个百分点的巨大优势遥遥领先进入决选轮。在欧洲乃至全球所有决选制（runoff）选举的历史上，从未有人能填补如此悬殊的分差；且西米昂的选票分布与落败候选人的基本盘高度重合，数学模型显示其胜选概率极高，群内核心成员因此砸下重金。 - **致命的本土认知盲区**：首轮大胜后，西米昂做出了匪夷所思的举动——他直接离开了罗马尼亚，缺席了后续的所有公开辩论。在罗马尼亚本土媒体和社交舆论中，西米昂迅速沦为全民笑柄，形象全面崩塌。 - **认知战的结果**：远在海外的互联网政治操盘手们完全没有捕捉到当地舆情裂变的严重程度，依然死守着历史统计概率下注；而大洋彼岸的罗马尼亚本土普通人则抓住了这千载难逢的赔率机会，用真金白银反向爆锤这帮海外“专家”。最终西米昂惨败，群内损失惨重。丹尼尔心服口服地总结道：“他们是对的，我们彻彻底底错了。” ### `38:25` 预测市场的灰色地带：内幕交易是否应该存在？ 在经历对海外翻车案例的复盘后，对话直接切入了预测市场最敏感的神经末梢——内幕交易（Insider Trading）。 在传统股票和金融证券市场中，利用未公开重大信息进行内幕交易属于严重的违法犯罪；但在去中心化或新兴的事件预测合约中，界限却变得极其模糊。一部分人认为，内幕交易的存在正是预测市场相比民调或传统媒体更具效率的根本原因，它利用金钱激励迫使掌握第一手真相的人（如政府幕僚、竞选团队、企业内部人士）提前将真实信息注入价格系统；但另一部分人则质疑，如果一个市场充斥着拥有内部消息特权的知情者，普通交易者沦为纯粹的被收割对象，市场的合法性与监管红线该如何维系。对于这一行为究竟是否涉嫌违法，连市场资深玩家之间也存在着剧烈的认知分歧。 ### `40:00` 内幕交易的双刃剑：价格信号还是不公收割？ 预测市场长期面临的一个核心伦理与法律困境是“内幕信息”。当普通交易者进入市场时，极有可能是在与真正身处核心决策现场、掌握确凿内幕消息的人对赌。对于散户交易者而言，发现对手盘甚至事先知道确切答案，这种体验无疑极其糟糕。然而，预测市场最初被学界与倡导者辩护的核心论点之一，恰恰在于它能为社会提供“价格信号”——通过利益驱动，将所有分散、隐秘甚至原本无法公开的信息引入公共领域，转化为透明的价格。 尽管 Kalshi 和 Polymarket 等主流预测平台竭力向公众传达“平台零内幕、监管极其严格”的合规形象，但在实际操盘的职业交易者眼中，这完全脱离了现实。在影视颁奖等非严格合规领域，内幕交易痕迹极为明显。在某届评论家选择奖（Critics' Choice Awards）颁奖前夕，演员雅各布·艾洛蒂（Jacob Elordi）爆冷荣获最佳男配角的合约在极短时间内、伴随着反常的巨额成交量，从 0.01 美元瞬间暴涨至 0.40 美元。当时押注其绝不可能拿奖的交易员为此吞下惨重亏损，这类凭空出现、在从未成交过的价格区间迅速吃光流动性的行情，几乎可以百分之百断定是有内幕人士提前拿到了获奖名单。 职业交易者普遍认为，平台和法律监管应当重点让严重的内幕交易在财务上“得不偿失”，而这一威慑机制正在逐步显现： - 泄露“Google 年度搜寻榜单”（Google Year in Search）并参与下注的员工不仅被开除，还被移交司法指控； - 提前获知军事行动并下注约 40 万美元的一名军方人员已被正式逮捕并面临刑事起诉； - 对于内幕信息与普通见闻之间的灰色地带（例如超级碗中场秀的演出细节、偶然听到的只言片语），法律根本无法全面监管； - 在美国国会内阁提名确认投票等政治事件中，参议员幕僚往往提前知晓投票意向。交易者唯一能做的风控原则就是时刻自省：“我的对手盘是谁？他们是否可能比我掌握更多不对称信息？”，并在高风险事件中严格控制仓位规模，避免沦为内幕人士的提款机。 ```mermaid flowchart TD accTitle: 预测市场对手盘风险评估与风控流程 accDescr: 展示交易员在评估市场内幕风险时的分类处理机制 A["评估目标预测市场"] --> B{"是否存在信息泄露渠道？"} subgraph S1["宏观与高透明度市场"] B -- "否 (如全国大选/公开宏观数据)" --> C["对手盘无内幕优势<br/>正常执行模型分析与重仓"] end subgraph S2["灰色与潜在内幕市场"] B -- "是 (如内阁确认/颁奖礼/文娱表演)" --> D["警惕异常成交量与价格异动"] D --> E{"是否有绝对内幕可能？"} E -- "轻度信息优势 (偶然听闻)" --> F["严控单笔头寸规模"] E -- "绝对确定性泄露 (价格暴拉)" --> G["果断回避 / 停止接盘"] end ``` --- ### `45:00` 戳破“散户能发财”的幻象与平台广告的虚伪 《纽约时报杂志》对职业交易员小圈子的深度报道在圈内引发了热烈反响。传统商业宣传中，Kalshi 和 Polymarket 这类平台往往充斥着轻佻的误导性广告，向公众灌输一种轻而易举暴富的幻象——“我今天赚到了房租，只因为我猜对了今天不下雨”。然而，当媒体严肃呈现 Brian、Daniel 以及他们所在的 Discord 圈子如何通过工业化的情报收集、数据建模和全天候盯盘盈利时，许多普通散户才惊觉自己原来就是预测市场里的“傻钱”（dumb money）。 这种报道在长期来看反而有助于预测市场的合法化与监管推进。长期以来，外界对预测市场的猎奇报道往往集中于低俗而无关紧要的博彩噱头，比如押注政客握手持续几秒、演讲中会不会含糊吐出某个特定词汇。交易员们直言，虽然每个人都有支配自己资金去赌博的自由，但各大平台为了追求短期暴利与流量，过度上线边缘化、擦边球的荒诞赌局，不仅激化了各州监管机构的敌意，更背离了预测市场应有的社会价值。只有当市场聚焦于真正具备经济与公共意义的问题——谁将赢得大选、通胀走势如何、关键法案能否通过——并由那些比所谓公认专家更努力的研究者为其定价时，预测市场作为真实“价格信号”的社会合法性才能确立。 --- ### `48:00` 警惕“AI 幻觉”：大语言模型不是预测神话 谈及当前技术革命对交易流程的赋能，两位交易员澄清了大语言模型（LLM）的实际效用与巨大误区。AI 的核心价值主要体现在扫盲与语言壁垒跨越上，例如启动一个陌生的海外小众市场研究，或翻译和理解非母语的新闻与政策文件；团队中偏向技术背景的成员也会使用 Claude Code 快速编写脚本以辅助统计分析。 然而，那些直接在 ChatGPT 等对话框中输入提问，并以为能据此获得市场优势的人，恰恰构成了市场上最典型的“外行资金/嫩钱”（squares / square money）： - 大语言模型具有极其严重的迎合性（sycophancy）与随机顺应特征； - 如果你向 ChatGPT 询问“下个月的通胀率会是多少”，它会给出一个头头是道的数字并夸赞“这是个极具洞察力的问题”； - 只要交易员随后反驳一句“这个数字太高了”，它就会立刻认同并修正为更低的数值，并列举完全相反的理由；若继续回复“太低了”，它又会再次道谢并顺从地调高； - 盲目相信大语言模型直接给出的概率预测（如社交平台上声称“AI 计算此事件概率为万亿分之一”），是散户赔钱最快的方式之一。 ### `50:00` AI 的局限与 Daniel 绰号“Carnitas Taco”的来历 在访谈的尾声，交易员们探讨了人工智能（AI）在预测市场中的作用。交易员直言，AI 在很大程度上只是基于它所吸收的各方专家的既有知识来给出回应，“如果我正在与之对赌的那些专家的水平本身就能被我击败，那么至少在目前的形态下，我不觉得 AI 会比他们更难战胜”。它缺乏真正穿透市场噪音的原生洞察，很多时候只是给出提问者想听的答案。 在告别之前，主持人抛出了最后一个轻松却令人好奇的“重要问题”：Daniel 为什么在网络上叫“Carnitas Taco”（墨西哥炖肉塔可）？Daniel 笑着解释了这段往事：他早年经常通宵打扑克锦标赛，赛局结束后，大家往往会一起熬到大清早去吃早餐塔可（breakfast tacos）。从那以后，“Carnitas Taco”便成了他形影不离的代号——不仅用作 DJ 艺名和 Twitter 账号，也成了他在各大预测市场的交易昵称。在互相致谢之后，嘉宾 Brian Golden、Daniel Reichman 和《纽约时报杂志》记者 Adam Iscoe 正式结束了访谈。 ### `51:41` 德州扑克繁荣的宿命镜像：散户离场与平台收割 送走嘉宾后，主持人 Joe Weisenthal 与 Tracy Alloway 展开了复盘。Tracy 感慨这期节目信息量极大，并提到了预测市场与当年“线上德扑繁荣”（Poker Boom）惊人的相似之处：都是起初有大批普通人涌入线上平台试水，结果持续输钱，最终心灰意冷地离场。 Joe 随即回顾了那段历史：2003 年，一位名叫 Chris Moneymaker 的名不见经传的普通会计师，通过线上选拔赛一路逆袭夺得世界扑克大赛（WSOP）冠军。这戏剧性的一幕让全美大众产生了一种错觉，以为“任何人都能打扑克赚大钱”，直接引爆了 ESPN 的全天候转播和全民扑克狂潮。然而到了 2009 年前后，除了面临美国监管对离岸平台的铁拳打击，这场热潮本身也早已显现颓势——因为 Chris Moneymaker 的奇迹纯属侥幸，普通散户（minnows / 小鱼）在不断输钱后逐渐被清洗出局。这为预测市场的未来敲响了警钟：如果散户和“笨钱”（dumb money / squares）被收割殆尽，市场沦为顶尖高手之间的零和互砍（sharp versus sharp），最终唯一稳赚不赔的就只有抽取佣金的交易平台。 ```mermaid flowchart TD accTitle: 预测市场与线上德扑的生命周期演变对比 accDescr: 展示从偶发奇迹引爆散户热潮，到散户持续亏损退场，最终只剩高手对决和平台抽成的生态闭环。 subgraph Phase1 ["第一阶段：热潮引爆"] A["造神神话 / 财富效应<br/>（如 Moneymaker 夺冠、预测市场暴富故事）"] --> B["大批散户与'笨钱'（Minnows/Squares）<br/>涌入市场参与投机"] end subgraph Phase2 ["第二阶段：残酷分化"] B --> C["缺乏信息与分析优势的散户<br/>面临持续亏损并逐步离场"] C --> D["市场仅剩下专业交易员<br/>顶尖高手对决（Sharp vs Sharp）"] end subgraph Phase3 ["终局形态：生态收敛"] D --> E["真实信息套利空间被极度压缩<br/>交易平台成为最终抽水赢家"] end ``` ### `53:07` AI 时代的真正护城河：走访实地、敲门与感知现实 主持人进一步总结了在预测市场生存的残酷真相：如果普通听众听完节目觉得自己也能轻松靠这行赚钱，那大概率会赔个精光，除非像这些专业交易员一样投入海量的实际调研工作。 在探讨现代交易优势时，两人指出，在各类 LLM 唾手可得的时代，许多大语言模型本质上依然是极度“后视镜”式的，它们只能总结过去的数据，却极难捕捉现实中的突发拐点。因此，当下最稀缺、最具备阿尔法（Edge）的资产反而是“私有数据”和线下的真实人际连接——不是躲在屏幕后向大模型提问，而是亲自去敲门、与当地人交谈、嗅探现场真实的情绪氛围（vibes）。 - **罗马尼亚 2025 年大选案例的反思**：Joe 特别赞赏了交易员们坦承失败的坦率态度。在罗马尼亚大选中，专业交易员们彻底看走了眼，付出了真金白银的代价；而当地那些关注时局的普通居民（randos），凭借身处现场的切实感知，反而比象牙塔里的顶尖分析师掌握了更精准的真相。 - **节目尾声与常规寄语**：Joe 与 Tracy 感谢了制作团队成员（Carmen Rodriguez、Dashiell Bennett、Kale Brooks 与 Kevin Lozano），并邀请听众订阅 Bloomberg 每日通讯、加入 Odd Lots 24/7 Discord 社区交流心得，随后进入了赞助商广告。 ## Notable quotes > `00:00` 人人都可能成为赢家 — Joe Weisenthal > `10:00` 大家私下问一句，这个群是不是 100% 都是男性？ — Joe Weisenthal > `10:00` 犯错要付出真金白银的代价 — Daniel Reichman > `20:45` 说明一下，根本没人给我打过电话招聘我 — Brian Golden > `20:45` 来自中西部的戏剧少年 — Brian Golden > `30:00` 我翻看了预测市场的评论区，感觉里面的人全是一帮白痴——至少跟我丈夫比起来是这样。因此，我非常支持他押上全家的积蓄去赌这一单 — Alan Cole 的妻子 > `30:00` 永远和评论区反着来 — Daniel Reichman > `35:00` 极其黑暗的一天 — Daniel Reichman > `35:00` 他们是对的，我们彻彻底底错了 — Daniel Reichman > `40:00` 我的对手盘是谁？他们是否可能比我掌握更多不对称信息？ > `45:00` 我今天赚到了房租，只因为我猜对了今天不下雨 > `50:00` 如果我正在与之对赌的那些专家的水平本身就能被我击败，那么至少在目前的形态下，我不觉得 AI 会比他们更难战胜 — Daniel Reichman ## Mentioned Bloomberg, Odd Lots, Joe Weisenthal, Tracy Alloway, Brian Golden, Daniel Reichman, Adam Iscoe, The New York Times Magazine, Maga Kiwi Club, Polymarket, Kalshi, PredictIt, Susquehanna International Group (SIG), Robinhood, Spencer Pratt, Nithya Raman, Pierre Poilievre, Matt Levine, Omar Sharif, U.S. Bureau of Labor Statistics (BLS), Alan Cole, Elon Musk, George Simion, Jacob Elordi, Google, Claude Code, ChatGPT, Chris Moneymaker, World Series of Poker (WSOP), ESPN, Notion ## Video description <p>Here's a couple things about prediction markets. A lot of it is pure gambling and speculation, much of it on things with very little economic relevance. Another fact is that in all likelihood, if you yourself started trading right now, you'd probably lose your shirt. But there is money being made by some dedicated traders, really focused on areas like politics and economics. On this episode, we speak with Brian Golden and Daniel Reichman, who are part of a private Discord called Maga Kiwi Club, where serious prediction markets traders swap ideas and make real money. We discuss the remarkable efforts they go to in order to spot opportunities, the systematic biases among traders, how they feel about insider trading, and other major issues that surround the space. Alongside Brian and Daniel, we also speak with NYC-based journalist and producer Adam Iscoe, who <a href="https://www.nytimes.com/2026/05/26/magazine/polymarket-prediction-wall-street.html">recently profiled these traders</a> for <em>The New York Times Magazine</em>.</p> <p>Only Bloomberg.com subscribers can get the Odd Lots newsletter in their inbox, plus unlimited access to the site and app. Sign up at <a href="https://www.bloomberg.com/subscriptions/oddlots?in_source=oddlotspodcast" data-link-cookie="1" data-destination="https://www.bloomberg.com/subscriptions/oddlots?in_source=oddlotspodcast">bloomberg.com/subscriptions/oddlots</a></p><p>See <a href="https://omnystudio.com/listener">omnystudio.com/listener</a> for privacy information.</p> [Source](https://podtrac.com/pts/redirect.mp3/tracking.swap.fm/track/UVBrz8bN8aM2Xe47PEPu/traffic.omny.fm/d/clips/e73c998e-6e60-432f-8610-ae210140c5b1/8a94442e-5a74-4fa2-8b8d-ae27003a8d6b/7ed31c1e-3cf5-467d-a302-b47b014db8c5/audio.mp3?in_playlist=982f5071-765c-403d-969d-ae27003a8d83)

[0:00] Joe Weisenthal: 人人都可能成为赢家 [10:00] Joe Weisenthal: 大家私下问一句，这个群是不是 100% 都是男性？ [10:00] Daniel Reichman: 犯错要付出真金白银的代价 [20:45] Brian Golden: 说明一下，根本没人给我打过电话招聘我 [20:45] Brian Golden: 来自中西部的戏剧少年 [30:00] Alan Cole 的妻子: 我翻看了预测市场的评论区，感觉里面的人全是一帮白痴——至少跟我丈夫比起来是这样。因此，我非常支持他押上全家的积蓄去赌这一单 [30:00] Daniel Reichman: 永远和评论区反着来 [35:00] Daniel Reichman: 极其黑暗的一天 [35:00] Daniel Reichman: 他们是对的，我们彻彻底底错了 [40:00] 我的对手盘是谁？他们是否可能比我掌握更多不对称信息？ [45:00] 我今天赚到了房租，只因为我猜对了今天不下雨 [50:00] Daniel Reichman: 如果我正在与之对赌的那些专家的水平本身就能被我击败，那么至少在目前的形态下，我不觉得 AI 会比他们更难战胜

[5:00] Brian Golden: I would describe it almost as an Avengers model where there are different people in the group that have different specialties and everyone has their own portfolio. But there is kind of a spirit of community in the way that information is exchanged. [10:00] Brian Golden: The reason that I think prediction markets have value is because there are consequences when you're wrong, and there are consequences when you're right... when you're wrong, you pay a price. [25:00] I try really hard to avoid these rules disputes and these thorny markets where it ends up kind of a debate. I think Polymarket's system is very problematic. They have basically undermined their oracle, and now every market just settles on how Polymarket clarifies. [30:00] Always bet against the comment section... sharps tend to keep their mouth shut except when talking to each other, and generally the more ideas or comments there are sort of advocating for one side, I do think that tends to be the wrong side.

[0:00] Whatever your goal, trade show giveaways, client gifts, or team gear, 4imprint has the promo products to match. With thousands of options, from apparel and drinkware to tech and totes, it's easy to find the right fit for your brand and budget, with standout choices at every price point. And with their 360-degree guarantee, you can be 4imprint certain your order will show up just right, right on time. Explore more at 4imprint.com. 4imprint. 4certain. When you're running a business, the best days are the ones where priorities stay on track. For midsize and large companies, that isn't always easy. Risk can touch multiple parts of an organization at the same time, often in ways that aren't immediately obvious. It might involve property, liability, or cyber. It could stem from regulatory requirements or challenges tied to a specific industry or the scale of an operation. At that level, managing risk becomes an ongoing discipline, not a one-time decision. At the Hartford, the focus is on helping businesses manage risk before it turns into something more disruptive. That means working with companies to identify where they exposed decide what matters most and put practical standards in place so risk is managed as part of day operations And when losses do happen the Hartford can pair that risk control work with insurance coverage grounded in underwriting risk engineering and claims experience developed over time. Learn more at thehartford.com slash risk mitigation. The thing about AI for business, it may not automatically fit the way your business works. At IBM, we've seen this firsthand, But by embedding AI across HR, IT, and procurement processes, we've reduced costs by millions, slash repetitive tasks, and freed thousands of hours for strategic work. Now we're helping companies get smarter by putting AI where it actually pays off, deep in the work that moves the business. Let's create smarter business, IBM. Bloomberg Audio Studios. Podcasts, radio, news. Hello and welcome to another episode of the Odd Lots podcast. I'm Joe Weisenthal. And I'm Tracy Alloway. Tracy, can I reveal a very boomer take of mine, so to speak? Go on I actually really load anything that sort of like identified as like get rich quick or the gamification of stock trading or gamification of stock trading Gamification would be something different. That would probably be something different. I don't know why I pronounced it that way. Gamification of stock trading, gambling ads. I really don't like gambling ads that imply they're going to be a winner. And I really don't like, even though I'm very interested in prediction market, I generally don't like a lot of the marketing. There was a really good Wall Street Journal article about polymarket and these ads that they had created, this idea that everyone's a winner and there's all this free money out there. You just have to trade your knowledge of whatever rainfall or whatever. I really like it really viscerally bothers me. Yeah, I think this is the key difference with prediction markets is it's zero sum. right if you make a bet someone is on the other side taking the opposite of that bet and that feels a little different to me than like traditional stock markets where you know you would get some cash flow some dividend you sell monetize whatever you got and someone else buys like no i totally agree i mean look there are certainly zero sum there are all kinds of zero sum markets in what we call trad five right futures are zero sum options are zero sum cetera But on the other hand like but yes zero games even if we talking about like options or futures or swaps or whatever, that's not investing. That's trading. That's trading, yeah. And that can often be speculation, and a lot of people lose a lot of money. And when we know in a lot of these environments, a lot of, like most people lose. Some people are very sophisticated, but a lot of people think they're going to go into some of these new markets or whatever, whether we're talking about prediction markets, whether we're talking about options trading on Robinhood and they have dreams of making a lot of money, or just talk about crypto, which is also zero sum, I think, and they don't. Yeah, absolutely. The other interesting thing about prediction markets, just from a sort of market structure question, and we've done episodes on this with Susquehanna, is the liquidity aspect. And as more professional investors get in, does that start to maybe arbitrage some of the edge that certain traders have seen so far because you know you read these stories specific story the average guys outsmarting wall street on prediction markets like average guys i have a question over whether or not they're actually average but you can imagine a scenario where more professionals get into prediction markets and maybe it becomes harder to actually beat [5:00] Absolutely. I think everyone should actually go back and listen to our episode that we did with Jeremy Malich, who runs the prediction markets desk over at Susquehanna before listening to this. Because the other thing he talked about is like how even some of these low liquidity markets, they put a lot of stock in the price and therefore they can like feel comfortable doing OTC transactions based on what the price. So there's interesting stuff going on. Anyway, you mentioned this article came out in The New York Times, May 26th. the average guys outsmarting Wall Street on prediction markets. And it is important. There are some people who are doing very well. And there are like a handful of, they would say, sharks in various other contexts who are like sharks. Very good. Most of us are minnows and some people do really well. And the vast majority are total idiots, right? Yeah. If I went on there, I'd quickly lose $1,000. And some fraction of it would go to the house. And a bunch of it would go to some of the best traders, most likely, especially because it's peer-to-peer. Anyway, we have a very complicated special episode. I don't even know how this is going to work, but I thought it'd be fun. So this was a great piece. We have the reporter on the piece with us as well as a couple of the Sharps. They are in a Discord together a bunch of them We just going to be talking with two of them The Discord is called the MAGA Kiwi Club That sounds fun So we going to be talking with Adam the journalist as well as some of the members of the MAGA Kiwi Club We have Brian Golden who is in the article He was identified as someone who's really top-notch at trading inflation contracts. We have Daniel Reichman, who also goes by Carnitas Taco, who does a lot of election and politics. And of course, we have Adam here, who is a journalist, reporter, and producer here in NYC, recently joined the tech company Notion, which all journalists might end up working at tech companies. Well, we do. We don't get to talk to a carnitas taco that often on the podcast. I'm very excited about this episode. Thank all three of you, Brian, Daniel, and Adam, for joining us on OddLive. It's great to be here. It's a pleasure. Thanks. Let's just start, Adam. I guess actually we're sort of trusting you. We've sort of delegated some of our judgment. But you reported out this story. How did you identify that there exist people in the world who consistently win on prediction markets. Yeah, I mean, I think it started for me, I got very interested in these markets because I was playing around in them, right? Yeah. And I was losing a lot of money. You know not a lot a lot of money but for me it was a lot of money And I was wondering who the heck am I losing this money to Am I losing to market makers Am I losing to the SIGs of the world And it occurred to me maybe that maybe there are some people who are just a little bit sharper than me who are out there And then I started asking around and I sort of started looking around on – just started just asking around basically. And one person sort of leaves the next and suddenly you find yourself in a Discord with a lot of guys who are just really crushing it sharper than anybody else. And two of those folks are Brian and Daniel. Well, Brian and Daniel, tell us about the Discord group because from what I can tell from the article, a bunch of prediction market traders get together, share information. Maybe you share some actual P&L. It kind of sounds like a multistrat hedge fund where everyone's like in charge of their own portfolio. But tell us about the Discord group. Yeah, well, the first thing is the MAGA Kiwi Club name is ironic. It's not a MAGA club. But most of the names of the group end up being some form of inside joke that emerges during the conversation. because we spend almost all day every day talking to each other about markets ranging from economics to culture to sports And it is sort of like that I would describe it almost as an Avengers model where there are different people in the group that have different specialties and everyone has their own portfolio. But there is kind of a spirit of community in the way that information is exchanged in that. I know that if I am giving helpful tips to Daniel and others on inflation, that they're going to give me helpful tips back on things that I may be adequate at, but not expert at. And we sort of work together in that way. Yeah, so a lot of us actually know each other dating back all the way really to 2016, 2017, when Predict It was the dominant prediction market. And we mostly know each other through politics originally because that was the prediction market game. and pretty much all of us had some interest in politics or very large interest in politics to start with. And the prediction market space has grown very large and the money has gotten larger. But a lot of us would be doing a lot of this stuff anyway, would be unpacking elections, trying to find the truth of things, trying to solve puzzles together. And it just so happens that the space has grown large around us, but we're not. [10:00] kind of doing the same thing we've always been doing. And the way I found myself, you know, talking to Brian and Daniel was, you know, I met another trader who loved elections. You know, this is sort of where the journey began for me. And, you know, we were just talking about elections and, you know, it was after maybe six weeks of talking and source building that finally he says, oh yeah, I'm actually talking to all of these other people. I'm not just doing this alone. I'm not just building models alone in my basement. You know, I'm like, I'm talking to all of my colleagues really and i said colleagues you're working at a fund no no no no not quite we're actually all in a discord together oh yeah i have to say it does sound pretty fun wait let's just between us is it 100 male ah this one yes yes okay thanks i just sort of i sort of i sort of assumed such let me ask you guys a question like i'm like an emh bro i think that generally across all markets, most people like markets are generally well priced. And I think I think that with prediction markets, too, in the specific sense that like I think it is not it is difficult to make money. But evidently in stocks, even though I think stocks are efficiently priced, some people seem to consistently make money in stocks whether they a Warren Buffett strategy or whether they work at a long short And evidently some people consistently make money in prediction markets How do you guys think about market efficiency generally and how high quality that signal is on a price? And like, would you say like those of us in the media who sometimes increasingly quote prediction market quotes like these are you find are they decent? Or when you look at them, are they just like, oh, there's just easy pickings everywhere? Well, I think if they were all efficiently priced, you wouldn't be able to get some of the returns that some people in this group have gotten. I think it would be much harder to enter and sort of crush markets repeatedly if the price signal were better. I guess where I see it is that there is a lasso who I would call kind of prediction market evangelists would like to go out and say, oh, prediction markets are the future. They're this incredible price signal, and they're more accurate than all these other places. I don't really believe that. I've just seen too many markets that are way, way off. The reason that I think prediction markets have value is because there are consequences when you're wrong, and there are consequences when you're right. And we have so much kind of expertise in the world that says a lot of things under the expert banner but then doesn really have any bills to pay when they mislead people or when they wrong And so really the most appealing part of this ecosystem for me is that when you're wrong, you pay a price. And I think that we might have a better expert culture in this country if that were true in media that covers economics and politics. It depends on the market, certainly, in terms of how correct these prices are going to be. But But going back to what Joe was talking about in the intro about the zero-sum nature of these prediction markets, a lot of the events, most of the events that we're betting on are zero-sum. We have an election and two campaigns spend a ton of money, and then one wins and one goes home. And the nature of politics in this country, especially, but all over the world, is that people live in totally siloed environments and believe different things and often engage with reality in totally opposite ways. and having a mechanism to essentially bet your beliefs and try to find real truth is, it's not going to be a perfectly efficient way, but at the moment, it seems to be better than at least anything else we got Well I think to Daniel point unlike inflation politics brings out matters of the heart and is I think much more subjected to that siloed effect I mean, you can look back at the Los Angeles mayor's primary that just happened and Spencer Pratt's price to not just make the top two, but to actually win the mayorship of Los Angeles got so unbelievably high. And there was this right wing media ecosystem and sometimes even a mainstream media place that was like sort of flirting with this idea like, can this happen? Everyone I know is is is saying that this is live. And there wasn't a sharp that I knew that didn't have one of the biggest positions of their lives on Spencer Pratt not winning the mayorship of Los Angeles because the math was just not there to be mathing. Los Angeles is a Democrat plus 42 city. A Republican is not going to win that race. But this this sort of siloed media of people sort of only following certain accounts on Twitter and watching certain media can lead to what feels like to them an abundance of evidence involved. [15:00] a certain thing happening. And in that sense, I think that elections will always be some of the most mispriced markets because, you know, people don't really have a heart connection to, gosh, I really believe I want inflation to be 3.6 instead of 3.8. I do think that the future of economics markets is probably a tightening. But people like Daniel are fortunate because I think elections will always be a little softer. This was my problem in the last election. I spent way too much time on Reddit, and so I thought Harris was going to win, and then I was very surprised. Just to press on this further, when we talk about you guys having an edge in this market, a lot of investors will say that they have an edge, and usually it's like they got lucky a few times and then they built a whole narrative around it. But how would you describe what you are doing differently to others? So you mentioned, you know, looking rationally at the numbers, keeping feelings out of it, social media echo chambers, that sort of thing. But you're also doing some original on the ground reporting You have your own models What is the edge exactly A lot of it is really just work that we put in So it being open to changing your mind trying to quantify and test your assumptions With politics, it's a lot of history. I mean, it's just knowing this has happened before. This is the trend. We think the trend might be this big this time when it was only half as big last time. And then seeing numbers come in and trying to stay calibrated with your assumptions when you win an election or win a bet. You don't just say, yeah, I won. You say, how much did I win? You know, what was my belief in the probability? Just like really digging in and unpacking it. And then as the money has gotten bigger recently, we've started trying to learn more. We tried to do some polling in the Texas primary, commissioned our own phone polls. Some of the guys in the group have done this live door-to-door polling, which we did a bigger trip with Adam that's written about in the article. But yeah, it's mostly just work and openness to data and changing your mind. I think Daniel's even underselling how good the elections team is in our group at elections and what they do. I mean I can tell you for the Los Angeles mayor race these guys had a model built of on election night what certain areas should look like in terms of the early vote the in vote what what it would take for Pratt to have what he needed You know, what we know from California is that so much vote comes in late that really predicting what that vote is going to look like. So these guys did historical work by precinct and region. And on election night, when Nithya Raman was in tears speaking to her supporters because she thought she lost, Daniel and the crew were betting on her to make it to the top two because they had a better vibe on her chances than I think her actual campaign did. So I really I think he's it would be hard actually to oversell how good this elections crew is with data and not just needing days to solve it, but being able to really solve on the fly based on on incoming numbers. Whether you planning a big tech event launching a new campaign or just stocking up on team gear finding the right promotional products makes all the difference 4imprint offers thousands of options from on apparel and premium drinkware to tech totes and giveaways so you can find the right fit for any audience, purpose, or budget. You can customize it all, your logo, your message, your look, and many items come with no setup charge to help you save. And if you're really watching the bottom line, you'll find standout choices at every price point, so you can make a real impact while staying on budget. Plus, you'll get expert help, fast turnaround times, and their 360-degree guarantee. So you can be 4imprint certain your order will arrive on time and look exactly right. Whatever your goal, 4imprint makes it easy to find your perfect promo match. Explore the possibilities today at 4imprint.com. 4imprint. 4certain. Support for the show comes from public.com. If you're actively involved in your portfolio, you probably catch yourself repeating the same actions. Buying the dip, manually sweeping idle cash, putting on a hedge. On public, you can now create AI agents that handle all these tasks on your behalf. Just describe what you want to do in plain English. Like, if the VIX hits 25, buy a put option on the S&P 500. Or, if my cash balance goes above $20,000, move the excess into my direct [20:00] index. You approve the workflow and your agent handles the rest. Monitoring the market, watching for your conditions, and executing your strategies exactly as defined. An investing platform driven by your intent, not just your clicks. You can also get full read and write access to your account via the public API. Go to public.com slash market and fund your account in five minutes or less. That's public.com slash market. Paid for by public investing. Brokered services by open to the public investing Inc member FINRA and SIPC advisory services by public advisors, LLC, SEC registered advisor, complete disclosures available at public.com slash disclosures. The thing about AI for business, it may not automatically fit the way your business works at IBM. We've seen this firsthand, but by embedding AI across HR, IT and procurement processes, we've reduced costs by millions, slash repetitive tasks, and freed thousands of hours for strategic work. Now we're helping companies get smarter by putting AI where it actually pays off, deep in the work that moves the business Let create smarter business IBM Even though my assumption was that prices are somewhat efficient I do notice that on election nights there is a lot of noise based on timing of vote batches with multiple elections We even saw it in the recent Canadian prime minister election, for example, where there was a spike for Paulyev on election night because of some early votes that were clearly not representative. So that always does make me sort of question whether I should ever be quoting these things. But Brian, in Adam's article, and maybe Adam, you could talk about this, your inflation models are described by some economists as, quote, being like Nostradamus of inflation. Give us a general overview of what you're doing and explain to us why you are not working at a – I think Matt Levine talked about this in his newsletter. It doesn't make sense. If you're like an inflation Nostradamus, why aren't you managing a multibillion-dollar rates hedge fund or something? Well, to be clear, nobody's called. Okay, well. The first thing that I did was the BLS has a formula to the way that they take all of their price inputs to calculate the inflation number. As I know your audience knows, but just to set the table it not one big number It 200 plus subcategories of how prices have moved which we call the basket of goods And the basket of goods changes what percent each subcategory is worth every month based on how much Americans are spending on it So this formula to me, I mean, look, my degree, I have an undergraduate degree in drama. I'm just a theater kid from the Midwest. Like this is very complicated. I'm sure that I know less about the macro economy than any guest you've ever had. But I rebuilt their formula on Excel. It took me like three months to figure out exactly how they math the formula. And a lot of really helpful public servants at the BLS answered my questions about just how the math works. And then you go from there to predicting the prices. There are some price categories that have public data like gas and natural gas and sometimes cars, but most of it is just trend work and guessing and looking at the last six months and sort of where prices seem to be heading. But the real thing for me is, I just think that this says a lot more about the sort of softness of the people, the investment banks and people who predict this in the institutional end than it does about me Because when you predict inflation it not even a prediction It in the past It data that has already happened And it has been very surprising to me since I entered this space that the places who advise billions of dollars of capital with their inflation forecasts aren't better at this. Well, this is what I wanted to ask because, by the way, BLS employees. Very helpful. Very helpful. You can actually call them on the phone and they will walk you through stuff. But on that note, when you describe just like, you know, working out the BLS formula, calling some people up and asking them how it all works, why aren't more people doing this, either in the prediction market or in traditional finance? I mean, you'd have to ask them. You know, it's shocking to me when, you know, it just really shouldn't be the case that my average absolute error on predicting inflation is better over the last two years than the Bloomberg consensus. I'm just one guy with Excel, and they have pretty much unlimited resources to find this data. It's not just shocking from a kind of what are they doing and why aren't they trying harder, but I think we find in both economics and elections that expert forecasts [25:00] really shape expectations. And that can play a big role in narrative creation and how the public responds to that. I mean, I've seen many times the headline on Bloomberg 20 minutes before the inflation number will say inflation to show blank as if it's a foregone conclusion because Goldman and J.P. Morgan and Bank of America have said so. And then when the inflation number is wrong, they just change the headline like that never happened. And the market reacts to it either being above or below that expectation, which maybe wasn't that good of an expectation in the first place. So I don't know why I don't know why they're not trying harder. I mean, they certainly have more resources to chase down these numbers than I do. For what it's worth, like the best inflation people we know all did it the same way you did, Brian, in terms of like actually the bottoms up approach to learning the formula, like all the inflation and of course like someone like Omar Sharif comes to mind. Yeah, he's great. There's a handful of people who like really, yeah. And this is what, I mean, this is what not just Brian, but all of the sharps that I talked to, whether they doing inflation or they doing elections or they doing meteorology how much is it going to rain next weekend in New York City whatever They all just making calls and going on the ground and talking to people gathering so much data They calling meteorologists they calling geophysicists they're calling Bloomberg reporters in many instances. I talked to a few sharps who were like, yeah, I'm on the phone with Bloomberg reporters all the time. But I think a lot of these sharps are just, they're constantly on the phone gathering information, which makes sense, This is also what you do if you're at a fund. Yeah. Can I ask – this is more of like a sort of – I don't know of a tactical question or market structure question. Maybe it's two interrelated things. So sometimes an event happens and then there is a period of time before it resolves. And Kelsey and Polymarket have different approaches to resolution. Polymarket is based on a sort of third-party, quote, Oracle, unquote, that is also sort of like whatever. And then Kelsey is more centralized. And I'm curious, like, what your trading philosophy is, it's like, okay, you get the LA mayorship right. Do you wait until resolution or do you sell it when it hits 99% and then move on to the next big thing? And is there alpha or profits to be gained in holding from the 99 to 100 during that period of resolution And is it different on either of the two sides I think it depends how much time it is and what other plays are available I mean usually, you know, if a market is going to resolve in the next 48 hours, I mean, you're pretty much always going to hold that from 99 to 100. But there are certainly elections that the answer is clear. And then, you know, you're not getting that payout for a month. And I mean, usually I can turn 99 cents into a dollar over a month faster doing something else than waiting. But I suppose it depends on the context. I mean, I don't know, Daniel might have a different approach. It's really important to know the difference between is your market 98 or 99 just because of the time it's going to take to resolve? Or is there a one or 2% chance of it actually going the other way? I try really hard to avoid these rules disputes and these, you know, thorny markets, where it ends up kind of a debate. I think Polymarket's system is very problematic. They have basically undermined their oracle, and now every market just settles on how Polymarket clarifies. I have generally for the most part been happy with CalShield resolutions but when you trade in stuff like will a cabinet member get confirmed or who wins an election it is almost never up for debate you know 2020 notwithstanding You have clear answers. And so it just depends. I'm holding the Peruvian election right now from 99 to 100 because I don't have anywhere else to put the money. But if I had another play that I liked and I wanted the money freed up, I would sell it in a second and, you know, put it in whatever else there was. Let's talk when we get off the call. I got some places for you. Okay. Since we're talking about probabilities changing, I wanted to bring up this. There's a long running debate, I guess, about whether prediction markets are actually better than traditional polls when it comes to forecasting election results, or whether they just look better because they're able to update faster as the results come in. And Daniel, I'd be very interested in getting your take on this. Are these genuinely producing better properties? probabilities or is it just a matter of speed? Yes, but it depends on how much money is coming into each side. A really noteworthy election, I think recently, last November, was the New Jersey governor's race, where the polls really all said this was a pretty close race, you know, two, three, [30:00] eight, four, five. One of the guys on our channel had made a very simple model, not even really a model, had basically just said, you know, Kamala won the state by, I forget, five or six. Trump's approval had fallen maybe eight or nine points since then. You know, Trump was president instead of Biden. Kamala or Cheryl's probably going to win by about 14. And this was just, you know, sketched out on priors on a napkin very quickly, three months before any polls. And then all these polls came in showing this close race. Very, very few polls showed a big margin. And we actually wouldn't ignore the polls. We kind of tried to standardize them all. But we did genuinely believe in our channel that this was a 12, 13, 14 point race that the polls didn't show at all. Now, we couldn't make the prediction market prices reflect that because there was so much money on the other side. So, you know, ultimately, the prediction markets, I think, were higher than the polls, pointing to a higher margin, but still well below what it would eventually come out to. So on an election where the liquidity on the square side is large enough the prediction market prices are still going to be wrong You can look at Pratt again Pratt price was never 27 to be the mayor It was less than five always in truth But it was also still that 27 was still much more accurate than anybody bubble who would have said But actually, this brings me to a question I want to go to the bubble question again. Adam, obviously, your piece was great. There was another great piece of prediction markets journalism in the last year about Alan Cole, who had bet his life savings that Elon Musk wouldn't actually reduce the deficit very much. And he like knows the deficit very well. He's like, this is never going to happen. And the money quote in that article is from Ellen's wife, who said, I read through the comment section in the prediction markets, and they all seem like idiots, at least relative to her husband. And therefore, I was very comfortable with him risking all of our family's life savings on this one particular bet. But I'm curious, like if you like, OK, when maybe there's sort of bubble mentality emerging or it's like this is a price that reflects people not getting good information. How often in your group can you use the comment sections to gauge like, wow, there's a lot of dumb money on this contract? Always bet against the comment section. OK, for real. I mean not always but it tends to be it tends to be a guiding principle because sharps tend to keep their mouth shut except when talking to each other and generally the more ideas or comments there are sort of advocating for one side I do think that tends to be the wrong side Yeah. I was going to say it was surprising how – or is surprising how reliable the comments indicator is. Back on PredictIt, there was actually a lot of good information. People hadn't built these Discord networks. They hadn't, you know, ended up in their silos where we talk about all the useful information ourselves. And so there was a lot more sharing and helping. And, you know, part of that is also because the money's gotten larger. The importance of protecting your reliable information is so useful that you just don't see, you know, someone like me or Brian going on to Calci's, you know, message board and explaining, no, you guys have it wrong because you're not, you know, accounting for this, that and the other. I mean, predicted having an $850 limit, you know, made it very possible for, you know, the limit meant you could share like I'm done. I can't fill up anymore. But Calci not having kind of a meaningful position limit you know changes that equation dramatically since we talking about betting against the comment section you know there was the article in the journal recently talking about how on Polymarket 67 percent of profits go to 0.1 percent of accounts. And we also have more professional investors who seem to be expressing some interest in getting into this market. If a bunch of people are just losing money on this, which it seems like they are, does the dumb money eventually go away and it becomes harder for you to, you know, make these bets? It should. That's been the history of, you know, the poker boom got harder. Daily Fantasy Sports was big money for a lot of people. That got harder over time. You know, certainly it should happen that way. So far, certain markets have gotten harder. Elections appear to still be dominated by vibes and emotions. You know, we'll see. Hopefully these prediction market spaces are still new enough that there's a lot more people still to be onboarded and anything can happen. But so far they still seem to be pretty beatable. Yeah, I remember during the online poker boom, people would like talk about like soft tables and all these. And then like eventually like they all lost their money and they were like no soft tables. And then the... I don't know. [35:00] I know that about the poker. Yeah, then it was just all like sharks versus sharks. And when it's like you go to the high limit room at a casino, and there has to be some celebrity there or a shake or someone who has a bunch of money who just wants to lose to pros there that night. But if it's all like Adam Ivey and all these guys playing high limit poker against each other, the only one who's just going to grind each other down and the money is going to go to the house. How did you guys do on the 2025 Romanian election? It was a dark day for our channel. Yeah, tell us about the Romanian election and the dark day for your channel. Yeah, Daniel, what happened? So in the first round, well, the first round of the Romanian election was actually annulled due to Russian involvement. And it was a very strange event. But the next first round, this guy Simeon won it by, I believe, about 20. And it was headed to a runoff. And basically in the entire history of European runoff elections and even you know expanding beyond European nobody had ever really come back from a deficit that big And there was also a correlation where the places that the other candidates had done well this guy Simi and the leader had also done well So it really just seemed at the beginning like he would win easily. And a lot of us put a lot of money on it at various times. He then proceeded to leave the country, skip all his debates, became a laughingstock on Romanian media, which we, to be clear, did not pick up on just how much he had become a joke in Romania. And so it ended up being a case where a lot of Romanians were betting a lot of money against a lot of Internet politics sharps. Yep. I mean, they were right. We were wrong. Thank you or budget You can customize it all Your logo your message your look And many items come with no setup charge to help you save. And if you're really watching the bottom line, you'll find standout choices at every price point so you can make a real impact while staying on budget. Plus, you'll get expert help, fast turnaround times, and their 360-degree guarantee so you can be 4imprint certain your order will arrive on time and look exactly right. Whatever your goal, 4imprint makes it easy to find your perfect promo match. Explore the possibilities today at 4imprint.com. 4imprint. 4certain. Support for this show comes from public.com. If you're actively involved in your portfolio, you probably catch yourself repeating the same actions. Buying the dip, manually sweeping idle cash, putting on a hedge. On public, you can now create AI agents that handle all these tasks on your behalf. Just describe what you want to do in plain English. Like, if the VIX hits 25, buy a put option on the S&P 500. Or, if my cash balance goes above $20,000, move the excess into my direct index. You approve of the workflow and your agent handles the rest. Monitoring the market watching for your conditions and executing your strategies exactly as defined An investing platform driven by your intent not just your clicks You can also get full read and write access to your account via the public API Go to public.com slash market and fund your account in five minutes or less. That's public.com slash market. Paid for by Public Investing. Brokered services by Open to the Public Investing, Inc., member FINRA and SIPC. Advisory services by Public Advisors, LLC, SEC Registered Advisor. Complete disclosures available at public.com slash disclosures. The thing about AI for business, it may not automatically fit the way your business works. At IBM, we've seen this firsthand. But by embedding AI across HR, IT, and procurement processes, we've reduced costs by millions slash repetitive tasks and freed thousands of hours for strategic work. Now we're helping companies get smarter by putting AI where it actually pays off, deep in the work that moves the business. Let's create smarter business, IBM. How do you feel about insider trading in prediction markets? Because this is also one of the big debates now, the idea that people – I don't need – it's not illegal, right? Well, I think it is. [40:00] I can't get clarity on this, but I know there are some charges against people for using like legally protected information to make these bets. How do you feel about the idea that you might be betting against someone who is actually like fully in the room and fully informed? I feel like. there's sort of a mixed feeling here where one of them, on one side, you're like, well, yeah, that's really terrible for retail traders to know that they are, you know, going to be going against someone who literally has the answer. On the other hand, you know, I think one of the early arguments for why prediction markets should be allowable is that they can provide price signal by allowing all information that exists to come into the public sphere and not stay private. I mean, you can imagine a scenario where, I mean, you could cook up a very Hollywood scenario, which I won't, but you can imagine where someone who was aware of illegal activity and had no way to you know make that public we trading on that information and sort of creating a little bit of price signal there I don know I mean you never want to be trading against an insider And of course CalShe and Polymarket would very much like people to think that there are no insiders at all that they have policed this aggressively. I don't really find that to be the case. I could tell you specific markets that I am 100% sure that I lost to someone with inside information. I think you can pick out a market and see, is this possible for this market to be insidered? You get a little bit of a sense of what price movement looks like when someone actually knows the answer. Huge volume coming out of nowhere at a price that hadn't been traded on before. I can tell you, I would bet my life savings that someone who knew the Critics' Choice Awards winners last spring knew that Jacob Elordi was going to win Best Supporting Actor the night before because he went from $0.01 to $0.40 on massive volume, which was a trade that I lost because I really didn't think he was going to win. So we've taken some of those Ls. I think ultimately you want to not allow it but it hard because it also can be part of giving a price signal on important events So what I like to see and what I think we moving toward is the sites and the companies making the serious insider trading not make financial sense for the people that do it. So we just saw the guy who insider the Google year in search lost his Google job. And I believe charges were referred. The military guy who bet on the invasion was arrested. I mean, he made something like 400K and now he's facing charges. I mean, there's no way that math works out for those people, right? So it's really the small insider ones. But then you also have these situations where it's debatable what's an insider. Like there were insiders on the Super Bowl performance market where, you know, somebody who has no actual attachment to the gig or the performance, you know, happens to hear something that, you know, makes them know more than the market, but they really have no technical relationship of being an insider. I don't think you can police that. So I like to stick to stuff, you know, again, like in elections, you can't have insiders. We bet a lot on the cabinet confirmations. And that was interesting because you could have insiders on those. You know some staffer to a senator knows how their person is going to vote You know you just kind of have to ask in every event you know who my counterparty Could they know more than me And in situations where someone could know more than you you really got to size appropriately and make sure that you not just blowing a ton of money to somebody who only is in it because they know more than you. Just for what it's worth, not that my opinion matters at all, but I don't think the government should be expending resources to police insider trading on the length of the Super Bowl halftime show because it's like it doesn't matter if you guys lose a bunch of money to an insider on that i don't care because like i think like regulated while markets are a good thing i do not want like you know public resources protecting people who are like um gambling on it but just you know i'm curious and maybe all three of you like this story and people discovering that you guys have this discord etc like from the perspective of the companies like i could see the cal she's and polymarkets of the world loving this because they're like look a bunch of people get in their minds that they could be brian and daniel and they could have a crew and do a bunch of information and win or i could see them disliking it because they're like wow you know what i'm not going to uh start i'm not going to trade because i do not have anywhere the means to come close to this group in terms of like the resources required to invest to trade consistently well i'm curious like what you've [45:00] And it's like, are stories like this good for the markets or do they would they rather sort of perpetuate the illusion that it really is like totally random? I just made my rent money because I knew it wasn't raining today. And that's simple, which is how some of their ads are. And this is like really like the sort of the advertising, the misadvertising, et cetera. I've seen, you know, you've seen it from both of the two major platforms. At a high level, I'm very, you know, concerned with like, how are people going to read this thing? And I think my hope and what folks have told me is, wow, I didn't really realize that I was the dumb money. And I don't even think I realized at the beginning of this journey that I was the dumb money. And so the hope, I think, is that people read something like this, people listen to a conversation like this, and they think, oh my gosh, okay. I maybe don't stand a chance against Brian or Daniel or the rest of these guys in these Discord groups. But Kashi and Polymark can kind of flip the script and tell you the other story. Yeah I mean I think what is in the long term good of prediction markets continuing to be legal and regulated is stories that are like Adams because a lot of journalism reporting on prediction markets has sort of focused on wow isn it crazy that people are making money on how long a handshake will last and what words someone will mumble at a speech? And, you know, I mean, my personal opinion is that people should be able to bet on things because it's their money. But these are not important questions that prediction markets are built to answer. And I think that in some ways, Calci and Polymarket have been so aggressive in trying to make as much profit as possible, as quickly as possible, that they have gotten deep in markets that are not only pushing against state regulation, but are clearly not the kind of important social questions and economic questions that prediction markets really should be answering. I mean, what I appreciated about Adam's story was that it was essentially about people who are going the extra mile and working really hard to try to answer questions that actually matter. Who will win elections? What prices will look like? And in some cases, people who are working harder than, you know publicly accepted experts on those So I think for the long term existence and regulation and legality of prediction markets you know stories that focus on I don want to say people like Daniel and I because it not about him and I but it about what questions are we all trying to answer and what puzzles are we trying to solve? And if those puzzles matter to the general discourse in the economy, then absolutely it should be legal. But some of the extraneous stuff that is just a silly excuse to gamble, I don't think that that is really the future of where this should be headed. Yeah, I think that's an important point because we talk about the price signal of all of this and the price signal really doesn't matter if it's like a dumb question being asked, right? One more question for me on your research process. How much of this has been enabled by AI and the tools that are now at your disposal? AI is very helpful for getting started on something like an international market or especially for searching in foreign languages where it can intermediate the language barrier for you. There's basically been, I think, very few of us use much AI for modeling. Some of the guys who code more have been using Cloud Code a bunch just to do some statistical stuff you know quicker and easier But the people who just ask LLMs a question and think that that gives them an edge on a market are some of the squarest money out there. You know, these LLMs will tailor their answer to what you ask. And, you know, if you ask the question a certain way, it will tell you this the probability and it will ignore it. What an insightful question. You are on the right track. Wait a second. Is square money, is that what you guys call soft money these days? Or like, is that square money? Is that like the term for like a soft table or whatever? I mean, that's just the opposite of the sharps, right? There's plenty of it. Got it. Oh, of course. The squares are the sharps. But you saw tons of this again, going back to Pratt, you know, all the people on Twitter, my LLM told me the chances of this were one in a trillion, et cetera. I've asked chat GPT things like, what will inflation be next month? And it will give me a number and it'll be like, wow, great question. I think it's going to be this because of these reasons. And then I'll just say without, without anything else, I'll just say that's too high. And it'll say, you're right. I'm glad you brought that up. It actually will be lower for these reasons. And then I'll say that's too low. And it'll be like, you know what? Thanks for bringing that to my attention. So, um, not only, [50:00] Is it telling you what you want to hear? But it only has grounding in other expertise that it gathers. And if that expertise that I'm already trading against is beatable, then I don't really see why the AI is any less beatable than that, at least in its current incarnation. I lied. I have one more question. Very important question. Daniel, why are you called Carnitas Taco? So when I used to play a lot of poker, I would often be in tournaments all night and then stay up for breakfast tacos in the morning. And that has stuck with me as a DJ name, a Twitter name, a prediction market name. Just that's my online name. Brian, Daniel, and Adam, thank you all so much for coming on AdLots. I think that actually worked. That was a little bit complicated to organize, but that was a great conversation and really appreciate all of you taking your time. Thanks for having us. Thanks for having us. Thanks. It was great to be here. Tracy, that was really fun. That was sort of a complicated episode to do but I thought that was like a I actually felt like I learned a lot in that conversation Yeah absolutely I didn realize that the poker boom had sort of gone through a similar thing where you had a bunch of people playing online and then they just kept losing and they left Yeah, because the story was like, do you remember like how the why the online poker boom like really happened? Vaguely, but remind me. Because this guy named Chris Moneymaker won the World Series. His name was Chris Moneymaker. Yeah. And he was a total nobody. and he won the world series of poker in las vegas so everyone got in their head that like actually anyone can win a lot of money and playing poker and that was the sort of catalyst for like poker becoming this thing that like espn would cover and etc it's this huge wave after wave and then eventually like there was i think sometime in 2009 or 2000 i think it was 2009 it was a big government crackdown on some of these like quasi-illegal offshore sites but that was It was already at that point, I think, like the minnows were coming out of, because a bunch of people were losing, that Chris Moneymaker was really a fluke. There were a lot of interesting things on that, including, as to your point, like will eventually the square money or the dumb money or the minnow is just like flush out of the system? And then it sharp versus sharp and only the platforms are making money All this stuff about like okay what is a healthy future for these prediction markets look like But also it is you know the real work involved to actually have an edge It's like if you are listening to this and you think you're going to make money, you probably aren't unless you like actually have some reason to think that you're like putting in work. It kind of emphasizes that in the age of AI, like the edge is still going out and finding new data, like picking up on turning points because most of the LLMs are still very backward looking. And I guess having that sort of like human connection. No, if you think about it, like- You have to know the vibes, right? But if you think about it too, it makes sense because one of the things that like a lot of our AI guests will talk about is the value of proprietary data, right? And so that actually makes sense. Like what is, quote, scarce in the age of AI? Well, someone knocking on doors and asking questions of people rather than someone just asking the model what they think is going to happen. I also think that story. I didn't mean to turn this into another. I was just curious. Yeah, no, it's an important question. I want to I'm curious how much money was lost on that. Twenty twenty five Romanian election because I know that that was like a big upset And so and he walked through you know respect the candor of admitting they really whipped on that one and that all the randos in Romania who are paying attention to it knew more than the sharps on that one. Shall we leave it there? Let's leave it there. All right. This has been another episode of the All Thoughts podcast. I'm Tracy Alloway. You can follow me at Tracy Alloway. And I'm Joe Weisenthal. You can follow me at The Stalwart. Follow our producers, Carmen Rodriguez at CarmenArmand, Dashiell Bennett at Dashbot, Kale Brooks at Kale Brooks, and Kevin Lozano at Kevin Lloyd Lozano. And for more Odd Lots content, you should check out our daily newsletter. You can find that at Bloomberg.com forward slash Odd Lots. And you can chat about all of these topics 24-7 in our Discord, discord.gg slash Odd Lots. And if you enjoy Odd Lots, if you like it when we talk to sharps about beating the squares or the circles, then please leave us a positive review on your favorite podcast platform. And remember, if you are a Bloomberg subscriber, you can listen to all of our episodes absolutely ad-free. All you need to do is find the Bloomberg channel on Apple Podcasts and follow the instructions there. Thanks for listening. [55:00] Whatever your goal, trade show giveaways, client gifts, or team gear, 4imprint has the promo products to match. With thousands of options, from apparel and drinkware to tech and totes, it's easy to find the right fit for your brand and budget, with standout choices at every price point. And with their 360-degree guarantee, you can be 4imprint certain your order will show up just right, right on time. Explore more at 4imprint.com. 4imprint. 4certain. When you're running a business, the best days are the ones where priorities stay on track. For midsize and large companies risk can affect multiple parts of the organization at once from property and liability to cyber and regulatory challenges At that level managing risk becomes an ongoing discipline At the Hartford, the focus is on helping businesses manage risk before it turns into something more disruptive. And when losses do happen, that work is paired with insurance coverage shaped by years of underwriting, risk engineering, and claims experience. Learn more at thehartford.com slash risk mitigation. policies provided by Hartford Fire Insurance Company and its property and casualty affiliates, Hartford, Connecticut. These days, it seems like AI agents are just about everywhere you turn, every field and every function. But without identity, you can't trust they'll serve your business instead of jeopardizing it. Fortunately, Okta helps you get identity right by securing your AI agent's identities, giving you a single layer of control, a single standard of trust. So whether an AI agent supports a single user or your entire enterprise, With Okta, you'll turn risk into opportunity. Secure every agent. Secure any agent. Okta secures AI.