心理账户 · Mental Accounting

"一块钱并不总是等于一块钱——大脑会先给它贴上标签。" — 源自 Richard Thaler 的行为经济学

标准经济学有一条铁律:钱是完全可替代的——工资里的 100 元、奖金里的、捡到的,价值完全相同。心理账户揭穿了这个假设:大脑会把钱分门别类塞进互不相通的心理账本——「生活费」省着花,「中奖的钱」大手大脚,「教育基金」一分不动。账户之间不通汇,于是同一笔钱在不同标签下,行为天差地别。

非平凡点:① 心理账户既是 bug 也是 feature。它扭曲理性(同样的钱该一视同仁),却也是一种自控工具——把钱分桶隔离,是用心理边界对抗意志力薄弱。② 交易效用:你愿意为海滩度假村里的同一瓶啤酒多付钱,不是因为啤酒更好,而是「参考价格」变了——商家正是靠操纵参考价来制造「划算感」。③ 沉没成本谬误其实是心理账户的推论:一个没「结清」的账户让人难受,于是硬把难看的电影看完,只为给账户画上句号。④ 与分布式系统同构:把全局资源人为分区成隔离的桶,牺牲了 fungibility(全局最优),换来可控性与防挥霍——本质是「全局效率 vs 局部可控」的权衡。

实践判别:当你对两笔金额相同的钱态度不同时,问一句——「如果这笔钱换个来源/标签,我还会这么花吗?」答案为否,就是心理账户在替你做决定。

经典例子

经典的「丢票实验」:丢了 100 元现金的人,大多照常买 100 元的演出票;但若丢的是已买好的那张票,多数人就不再补票了——因为「看这场演出」的账户已经花掉 100 元,再花就是 200 元,太贵。总损失一样,行为却相反,只因记进了不同账本。

场景 · BigCat

① 年终奖与工资是同样的钱,但奖金更容易被冲动消费、或投进高风险标的——因为它被记进了「意外之财」账户。反过来用它:发奖金当天就转入长期账户,让「省着花」的标签自动生效。② 把 AI 订阅、云算力统统记成「实验经费」反而最容易超支——「实验」这个标签天然宽容。给它设硬上限,逼它回到统一预算。③ 教孩子理财别讲大道理,给她三个透明罐子:「花的」「存的」「给的」。心理账户从娃娃抓起,是一种看得见的自控脚手架


Mental Accounting — standard economics assumes money is fungible: a dollar is a dollar. In reality the brain files money into non-interchangeable mental "accounts" (rent money is guarded, windfall money is splurged). This is both a bug (identical money should be treated identically) and a feature (partitioning is a self-control device against weak willpower). Key corollaries: transaction utility (you pay more for the same beer at a fancy resort because your reference price shifts) and the sunk-cost fallacy (an unclosed account hurts, so you finish the bad movie just to "settle" it). Structurally it's a global-efficiency vs local-controllability trade-off — like partitioning a shared resource. Test: would I spend this money differently if it carried a different label or source?

中文提示词
我最近在 [某笔钱 / 某项开支] 上的行为是 [描述]。请用「心理账户」帮我诊断: ① 我是不是给这笔钱贴了某个标签(如「意外之财」「实验经费」「沉没的」),导致我对它和普通的钱区别对待? ② 这个心理账户此刻是在帮我(自控)还是在害我(挥霍 / 沉没成本)? ③ 给 1 个具体动作:要么换标签让它服务于我的目标,要么把它并回统一预算重新评估。
English Prompt
My recent behavior around [a sum of money / a category of spending] is [describe]. Use Mental Accounting to diagnose: 1. Have I labeled this money (e.g., "windfall," "experiment budget," "sunk") in a way that makes me treat it differently from ordinary money? 2. Is this mental account currently helping me (self-control) or hurting me (overspending / sunk-cost)? 3. Give one concrete move: either re-label it to serve my goal, or merge it back into a unified budget for honest re-evaluation.

禀赋效应 · Endowment Effect

"一旦它成了『我的』,放弃它就变成了一种损失。"

禀赋效应:一旦你拥有某物,对它的估值会立刻上升。经典的马克杯实验里,刚拿到杯子的人开价要七元才肯卖,没拿到的人却只愿出三元买——同一个杯子,「卖价」常是「买价」的两倍。拥有这件事本身,凭空给物品镀了一层金。

非平凡点:① 它的根源是损失厌恶——失去拥有物的痛,大于得到它的快。所以禀赋效应不是「我更懂它的价值」,而是「放弃 = 损失」,而损失被大脑放大了。② 它远不止于物品,更适用于观点、身份、现状——你对自己持有的想法也有禀赋效应,这正是确认偏误背后的情感引擎:放弃一个旧观点,感觉像在割肉。③ 「拥有」可以纯属心理:免费试用就是在制造禀赋——一旦「它已经是我的」,退订就成了「损失」。整个订阅经济都建立在这台机器上。④ 与佛学的「我执」同构:把无常的占有当成自我的延伸,于是放手即痛。破除禀赋效应,是一种「无我」的微型实修。

实践判别(零基重估法):「如果我现在并不拥有它,我会花多少钱把它买回来?」当「买回价」远低于你舍不得放手的「卖出价」时,差额就是禀赋效应在收的「智商税」。

经典例子

股市里的「死攥亏损股」:买入价 100、现价 60 的股票,很多人死也不卖,宁愿等它「回本」。理由不是基本面——而是卖出 = 兑现损失 = 给账户画上一个痛苦的句号。同样的钱,若问他们「现价 60,你现在会买吗?」答案往往是「不会」。既然不会买,逻辑上就该卖;不卖,正是禀赋效应 + 损失厌恶在作祟。

场景 · BigCat

① 屯了一硬盘「总有一天会读」的论文、一堆没上的课、一段早该重写的旧代码——都舍不得删。用零基重估法:「如果今天从零开始,我还会主动获取它吗?」答案为否,就该清掉。② 你对亲手设计的架构、调好的 prompt 链有强烈禀赋效应,这会蒙蔽客观判断——刻意问「如果这是别人交来的方案,我会给几分?」③ 孩子舍不得扔旧玩具,不是不懂事——禀赋效应从小就满格。理解这点会让你更耐心,也可用「捐给更需要的小朋友」把「损失」重新框成「给予」,化解那份痛。


Endowment Effect — once you own something, you instantly value it more. In the classic mug experiment, sellers demand ~2× what buyers will pay for the identical mug. The root is loss aversion: giving up a possession hurts more than acquiring it pleases, so it's not "I understand its value better," it's "letting go = loss," and loss is amplified. It extends beyond objects to beliefs, identity, and the status quo — you're endowed with your own opinions (the emotional engine of confirmation bias). Ownership can be purely psychological: free trials manufacture endowment, which is why the whole subscription economy runs on it. The de-bias tool is zero-based revaluation: "If I didn't already own this, how much would I pay to buy it back?" The gap between that and your reluctant selling price is the tax the endowment effect collects.

中文提示词
我正纠结要不要放弃 / 卖掉 / 删掉 [某物 / 某观点 / 某方案]:[描述现状]。请用「禀赋效应」帮我做零基重估: ① 如果我此刻并不拥有它,我愿意花多少代价把它买回来?把这个「买回价」和我舍不得放手的「卖出价」做对比。 ② 我对它的高估,有多少来自真实价值,多少来自「放弃 = 损失」的痛? ③ 给一个能把「损失」重新框定的角度(如「释放出空间 / 注意力」),帮我降低放手的心理阻力。
English Prompt
I'm agonizing over whether to give up / sell / delete [an object / belief / plan]: [describe]. Use the Endowment Effect for a zero-based revaluation: 1. If I didn't already own it, how much would I pay to acquire it now? Contrast this "buy-back price" with the "selling price" I'm clinging to. 2. How much of my high valuation is real value vs. the pain of "letting go = loss"? 3. Offer a reframe that turns the "loss" into a gain (e.g., reclaimed space / attention) to lower the psychological barrier to release.

IKEA 效应 · The IKEA Effect

"劳动孕育爱——你亲手组装的,在你眼里就是更值钱。"

IKEA 效应:人对自己亲手参与创造的东西,估值会系统性偏高。自己组装的家具,哪怕歪歪扭扭,你也觉得它比同款成品更好看、更值钱。劳动本身,注入了情感价值。

非平凡点:① 它不只是沉没成本。沉没成本是「我已经投入,不甘心」;IKEA 效应是付出努力的过程本身创造了意义和依恋——是正向的爱,不是被动的不甘。② 它是禀赋效应的「主动增强版」:自己造的东西,禀赋更牢。③ 双刃剑——好的一面:让用户、孩子、团队参与创造,承诺与投入会飙升(半成品蛋糕粉的历史就是证据:全自动版本「太省事」反而卖不动,改成「需要你自己打一个鸡蛋」后大卖)。坏的一面:对自己的作品、方案、代码丧失客观判断力。④ 关键边界条件:只在「成功完成」时生效——半途而废的作品非但不增值,还可能贬值。这解释了为何「让人有参与感」必须配「让人能完成」。⑤ AI 时代的新张力:AI 替你生成的东西,你的 IKEA 效应很弱(你没出力)——这有利于客观评判,却不利于承诺与珍惜。人机协同里要刻意保留「你的劳动印记」,否则产出虽多,归属感却空。

实践判别:评价自己的作品时,套用「换人测试」——「假如这是别人交给我的,我会打几分?」自评与换人评之间的落差,就是 IKEA 效应的加成。

经典例子

折纸实验:让人自己折青蛙或纸鹤,再给作品出价,结果接近专家折的同款——而旁观者只愿出零头。同一张纸,「我折的」三个字就让估值翻了几倍。商业上,宜家把「组装」这个本该是成本的环节,变成了情感黏性的来源。

场景 · BigCat

① 你亲手搭的笔记系统、自己调的 Obsidian 库——客观上未必比现成工具强,但你会高估它,因为是你一砖一瓦垒的。判别:让外人用一周,看真实反馈。② 工程团队里,「不是我们造的」(NIH) 会让人轻视外部方案;它的镜像是「我们造的」被过度珍视,明明该弃用却死守。两者都是 IKEA 效应。③ 育儿里它是宝贵的引擎:孩子自己搭的乐高、自己做的手工,别急着替他「优化纠正」——那份歪扭里住着他的投入和成就感。保护他的 IKEA 效应,就是保护他的内在动机。


The IKEA Effect — people systematically overvalue what they helped create. Self-assembled furniture feels more valuable than the identical pre-built version; in studies, people price their own folded origami near expert-level. It's not mere sunk cost (passive reluctance) — the act of effort itself manufactures meaning and attachment (active love). It's the proactive amplifier of the endowment effect. Double-edged: letting users/kids/teams co-create skyrockets their commitment (the half-baked cake-mix lesson: adding "crack your own egg" made it sell), but it destroys objective judgment of your own work. Crucial boundary: it only fires on successful completion — abandoned projects don't gain value. AI-era tension: AI-generated output triggers weak IKEA effect (you didn't labor), which aids objectivity but starves ownership — preserve your "labor signature" deliberately. De-bias tool: the swap test — "If someone else handed me this, what score would I give it?"

中文提示词
我做了 [作品 / 方案 / 系统 / 代码],正在评估它好不好 / 要不要继续投入:[描述]。请用「IKEA 效应」帮我祛魅: ① 套用「换人测试」:假如这是陌生人交给我的同一份东西,我会客观打几分?我的自评高出多少? ② 我的高估,有多少来自它真好,多少来自「这是我亲手做的」? ③ 如果是我想让别人(团队 / 孩子 / 用户)更投入某件事,请反向设计:在哪个环节加入「让他亲手做成一点」,能制造健康的 IKEA 效应?
English Prompt
I made [a work / plan / system / code] and I'm judging its quality or whether to keep investing: [describe]. Use the IKEA Effect to de-bias me: 1. Run the swap test: if a stranger handed me this exact thing, what objective score would I give it? How much higher is my self-rating? 2. How much of my high valuation is "it's genuinely good" vs. "I made it with my own hands"? 3. If instead I want someone (team / child / user) more invested, design in reverse: at which step should I add "let them complete a piece by hand" to create a healthy IKEA effect?

默认选项与助推 · Default Options & Nudge

"不强制、不改变激励,只改变预设值,就能改变千万人的行为。"

人有强烈的默认偏好:倾向于不动预设选项——惰性、现状偏误,加上「默认 = 推荐」的暗示,三股力合流。于是,改变默认值——不强制、不动激励、不减少自由——就能大幅改变行为。这就是「助推」:一种「自由主义的家长制」,既保留你说不的权利,又把好选项摆在你最省力的路径上。

非平凡点:① 没有「中立」的默认——只要存在预设选项,选择架构师就已在替众人做选择。真问题从来不是「要不要影响」,而是「朝哪个方向、是否透明」。② 行为的最大杠杆往往不是动机,而是摩擦:想让人多做某事,就让它成为默认、最省事;想让人少做,就给它加摩擦。改环境,比改意志力可靠得多。③ 助推与操纵(暗黑模式)的分界线在三条:是否为被助推者的利益、是否透明可逆、是否尊重其自主选择的能力。同一招术,越界即作恶。④ 与软件工程的「合理默认值 / 约定优于配置」同构:好的默认值承载了设计者的智慧,是产品中最强、却最隐形的引导。

器官捐献率:只因默认值不同 约 15% 选择「加入」(opt-in) 约 90%+ 默认捐献,可「退出」(opt-out) 同样的人、同样的偏好——只是把空白表格的默认勾选换了个方向 摩擦 ≈ 0,行为差距 ≈ 6 倍
默认值是最隐形、却最强大的选择架构杠杆
经典例子

器官捐献:需主动「加入」(opt-in) 的国家,登记率常只有两位数低位;而默认即捐献、需主动「退出」(opt-out) 的国家,登记率高达九成以上。人没变、偏好没变,只是表格上默认方向不同。同理,退休储蓄改为「自动加入、可退出」后,参与率从约五成跃升到九成——一个默认值,改写了千万人的晚年。

场景 · BigCat

① 设计自己的环境 = 给自己设默认。想多读书,把书放床头、手机扔客厅;想少刷手机,删掉 app、开灰度屏(给刷手机加摩擦)。改默认,远比每天靠意志力硬扛可靠。② 做产品或带团队:好的默认配置就是最强引导,远胜写一摞文档去「教育」用户——绝大多数人永远用默认值。③ 设计 AI agent 时,给它的默认行为、默认 prompt 模板,是整个系统最大的行为杠杆——改一行默认,胜过事后纠正一百次。④ 育儿不靠天天说教,而靠设计家庭「默认」:零食不进家门,书随手可及。你是孩子环境的选择架构师——请有意识地、为他的利益、透明地设计这些默认。


Default Options & Nudge — people have a powerful default bias (inertia + status-quo bias + "default = recommendation"). So changing the default — without force, without altering incentives, without removing freedom — massively shifts behavior. This is the nudge: "libertarian paternalism." Key points: (1) there's no neutral default — once a preset exists, the choice architect is already choosing for people; the real question is direction and transparency. (2) The biggest lever is often friction, not motivation: make the desired action the default/easiest, add friction to the undesired. Redesigning the environment beats willpower. (3) Nudge vs. manipulation (dark patterns) divides on three lines: is it for the nudged person's benefit, is it transparent and reversible, does it respect their autonomy. (4) Isomorphic to "sensible defaults / convention over configuration" in software — good defaults carry the designer's wisdom so 99% of users get a good outcome with zero configuration. Classic proof: organ-donation opt-in (~15%) vs. opt-out (90%+), and auto-enrollment retirement savings.

中文提示词
我想让 [自己 / 团队 / 用户 / 孩子] 多做 [行为 A]、少做 [行为 B],目前情况是 [描述]。请用「默认选项与助推」帮我设计: ① 当前的「默认值」和「摩擦分布」是怎样的?它在暗中把人推向哪边? ② 给 2 个改默认 / 调摩擦的具体动作:让 A 成为最省力路径,给 B 加阻力(不靠说教和意志力)。 ③ 自检三条伦理红线:这个助推是否为对方的利益、是否透明可逆、是否尊重对方说不的权利?指出任何越界为操纵的风险。
English Prompt
I want [myself / team / users / my child] to do more [behavior A] and less [behavior B]; current situation: [describe]. Use Default Options & Nudge to design: 1. What are the current defaults and friction distribution? Which way are they silently pushing people? 2. Give 2 concrete moves to change defaults / adjust friction: make A the path of least resistance, add friction to B (no lectures, no willpower). 3. Ethics check on three lines: is this nudge for their benefit, is it transparent and reversible, does it respect their right to say no? Flag any risk of crossing into manipulation.