The analysis in one sentence用一句话定义这次分析
To analyze a stock, write one falsifiable claim about how the company creates per-share value, assemble the original disclosures needed to test it, reproduce the key operating and capital-allocation calculations, challenge the result with a downside case, and state a dated conclusion with unresolved questions. The output is not a score; it is an evidence trail from company mechanism to shareholder outcome.
分析一只股票，先写出一个关于公司如何创造每股价值、而且能够被证伪的命题；再收集足以检验它的原始披露，重做关键经营与资本配置计算，用下行情景攻击初步判断，最后写下带日期、保留未决问题的结论。最终产物不是一个分数，而是一条从公司经营机制通向股东结果的证据链。
Start with the mechanism, not the ticker从经营机制开始，而不是从股票代码开始
Before opening a valuation model, complete this sentence: “This company can increase value per share if ___, because ___, provided that ___.” Name the operating engine, its economic link, and its constraint. A retailer might depend on mature-store contribution repaying setup cost quickly. The sentence narrows the analysis to evidence that can change your mind.
在打开估值模型之前，先打开一张空白纸，补完这句话：“如果____，这家公司就能提高每股价值，因为____，前提是____。”第一处填写经营引擎，第二处说明经济联系，第三处写下约束条件。连锁零售商可能依靠新门店在成熟后产生足够贡献，迅速收回开店投入；平台企业可能依靠买家重复交易，而不必同比例增加获客支出。这个句式会把分析范围收窄到真正可能改变判断的证据。
Harbor & Hearth, a fictional specialty retailer, can create value by opening stores only if mature-store cash contribution remains repeatable, the cash payback period stays within management’s hurdle, and expansion does not weaken the balance sheet. Every number below is invented for teaching and describes no real issuer or security.
贯穿案例的工作命题虚构的专业零售商 Harbor & Hearth 只有在成熟门店现金贡献可复制、现金回收期不超过管理层门槛、且扩张不削弱资产负债表时，开店才可能创造价值。下文所有数字均为教学而虚构，不对应任何真实发行人或证券。
New-store expansion compounds value rather than merely making reported sales larger.
命题新店扩张带来价值复利，而不只是让报表收入变大。
A store reaches stable sales and contribution, then repays buildout, opening, and initial working-capital cash.
可观察机制门店达到稳定销售额与贡献利润后，逐步收回装修、开业和初始营运资金投入。
Payback lengthens because sales mature below plan, store margin falls, required investment rises, or closures absorb the supposed gains.
失败条件若成熟销售不及计划、门店利润率下降、所需投入上升，或关店损失吞噬收益，回收期就会拉长。
Store-level gains must survive central costs, financing, taxes, dilution, and poor reinvestment elsewhere.
股东连接点门店层面的收益还必须经得住总部成本、融资、税费、股份稀释和其他低效再投资的消耗。
Build a claim-sized evidence pack按命题大小组装证据包
For a U.S. issuer, begin with the latest Form 10-K and subsequent 10-Q and 8-K filings. Add the earnings release, investor presentation, proxy statement, debt documents, or older filings only when they answer the claim. Other markets use different forms, so identify the issuer’s home reporting system. Save each file with its filing date and preserve the page or note for every extracted item.
不要因为文件存在就全部下载。若分析美国发行人，可先取最新 Form 10-K，以及此后提交的 10-Q 和 8-K；随后只在它们能回答命题时，补充业绩新闻稿、投资者演示、委托书、债务文件和更早披露。其他市场采用不同表格与监管体系，应先确认发行人的主要报告制度。保存文件时写入提交日期，每个摘录都保留页码或附注位置。
| Question | Primary location | Record before interpreting |
|---|---|---|
| How does a store earn? | Business, MD&A, segment and revenue notes | Store definition, sales measure, cost inclusions, fiscal period |
| Is expansion repeatable? | Store-count table, property note, capex discussion | Openings, closures, relocations, construction timing, leased versus owned |
| What cash is committed? | Cash-flow statement, leases, commitments, debt notes | Buildout, fixtures, opening expense, inventory, lease obligations, funding source |
| Does management benefit from growth at any price? | Proxy statement and compensation discussion | Targets, measurement windows, adjustments, equity awards and share count |
| What would contradict the thesis? | Risk factors, impairments, restructuring and later filings | Closures, asset write-downs, revised targets, covenant or liquidity pressure |
| 问题 | 优先查找位置 | 解释前先记录 |
|---|---|---|
| 一间门店如何赚钱？ | 业务说明、MD&A、分部与收入附注 | 门店定义、销售口径、成本包含项、财政期间 |
| 扩张能否复制？ | 门店数量表、物业附注、资本开支讨论 | 新开、关闭、迁址、建设时点、租赁或自有 |
| 承诺了多少现金？ | 现金流量表、租赁、承诺与债务附注 | 装修、设备、开业费用、存货、租赁义务、资金来源 |
| 管理层是否会不计代价追求增长？ | 委托书与薪酬讨论 | 考核目标、计量期间、调整项目、股权奖励与股数 |
| 什么证据会推翻命题？ | 风险因素、减值、重组和后续披露 | 关店、资产减记、目标修订、契约或流动性压力 |
Use four note columns: Reported fact, Calculation, Assumption, and Judgment. “The company opened 18 stores” is reported if the filing says so. “Upfront cash per opening was $2.05 million” is calculated if its components are disclosed. A 15% mature-store margin is an assumption unless management reports the cohort result. “Expansion appears attractive” is a judgment. Keep these distinctions visible.
在笔记里固定设置四列：已披露事实、计算、假设、判断。“公司新开 18 家门店”在财报明确披露时属于事实；如果投入组成均有来源，“每家新店平均前置现金投入 205 万美元”属于计算；除非管理层披露了同一批门店结果，“成熟门店可维持 15% 贡献利润率”仍是假设；“扩张看起来有吸引力”则是判断。不能让流畅叙述抹掉这些层级。
Reconstruct the store engine before judging growth先重建门店引擎，再评价增长
Harbor & Hearth hypothetically reports 120 stores at the start of the year and 132 at year-end, after 18 openings and six closures, plus $48.0 million of total capital expenditure. A standard opening uses $1.20 million of buildout and fixtures, $0.50 million of opening expense, and $0.35 million of initial net working capital. First reconcile 120 + 18 − 6 = 132. If the bridge failed, stop until acquisitions, relocations, or definition changes explain it.
在虚构披露中，Harbor & Hearth 年初有 120 家门店，全年新开 18 家、关闭 6 家，年末为 132 家；公司还披露全年资本开支 4,800 万美元。另一项教学性披露称，标准新店需要 120 万美元装修及设备、50 万美元开业费用、35 万美元初始净营运资金。第一项核对只是机械勾稽：120 + 18 − 6 = 132。若这个桥接不成立，就应暂停分析，查找收购、迁址或门店定义变化，而不是继续做精细回报计算。
| Item | Status | Illustrative amount | Analytical treatment |
|---|---|---|---|
| Year-end stores | Reported fact | 132 | Reconcile opening and closure bridge |
| Total company capex | Reported fact | $48.0m | Do not assign all capex to new stores |
| Standard buildout and fixtures | Reported fact in the fictional case | $1.20m per opening | Compare with property additions and capex narrative |
| Opening expense | Reported fact in the fictional case | $0.50m per opening | Expense is still a cash cost of creating the store |
| Initial net working capital | Reported fact in the fictional case | $0.35m per opening | Inventory may be recoverable, but cash remains committed |
| Mature sales and margin | Analyst assumption | $6.40m and 15% | Stress rather than present as achieved |
| 项目 | 性质 | 教学数值 | 分析处理 |
|---|---|---|---|
| 年末门店数 | 已披露事实 | 132 | 与新开及关闭数量勾稽 |
| 公司总资本开支 | 已披露事实 | 4,800 万美元 | 不能全部归因于新店 |
| 标准装修及设备 | 虚构案例中的披露事实 | 每家 120 万美元 | 与物业增加额和资本开支叙述核对 |
| 开业费用 | 虚构案例中的披露事实 | 每家 50 万美元 | 即使计入费用，仍是创建门店的现金成本 |
| 初始净营运资金 | 虚构案例中的披露事实 | 每家 35 万美元 | 存货可能回收，但现金仍被占用 |
| 成熟销售额及利润率 | 研究者假设 | 640 万美元与 15% | 必须压力测试，不能冒充已实现结果 |
Upfront cash per store is $1.20m + $0.50m + $0.35m = $2.05m, or $36.90m for 18 openings. Do not subtract this from $48.0m of total capex and label the $11.10m difference “maintenance”: opening expense and working capital do not necessarily sit in capex. Estimate maintenance spending separately from remodels, equipment, depreciation, asset ages, and management commentary.
计算所得每家新店前置现金为 120 + 50 + 35 = 205 万美元，18 家合计 3,690 万美元。它与 4,800 万美元总资本开支看似相容，但开业费用和营运资金通常不会全部出现在资本开支项目中。因此，4,800 万减去 3,690 万所得的 1,110 万并不是“维护性资本开支”，只是两个范围不同数字之间的差额。维护投入应结合翻新、设备、折旧、资产年限及管理层说明单独估计。
Calculate the expansion hurdle without renaming it ROIC计算扩张门槛，但不要把它误称为 ROIC
Assume a mature new store produces $6.40 million of sales at a 15% contribution margin after merchandise, store labor, and occupancy, but before central overhead, taxes, financing, and shared costs. Assume $0.12 million of annual maintenance capex. Annual store cash contribution = Sales × Contribution margin − Maintenance capex: $6.40m × 15% − $0.12m = $0.84m. Simple payback = Upfront cash ÷ Annual cash contribution: $2.05m ÷ $0.84m = 2.44 years.
这里不是声称，而是假设一家成熟新店年销售额为 640 万美元，门店贡献利润率为 15%；该利润率已扣除商品、门店人工和场地成本，但未扣除总部费用、税费、融资成本及部分共享成本。再假设每年维护性资本开支为 12 万美元。用于初筛的年度门店现金贡献 = 销售额 × 门店贡献利润率 − 维护性资本开支，即 640 × 15% − 12 = 84 万美元。简单现金回收期 = 前置现金 ÷ 年度门店现金贡献，即 205 ÷ 84 = 2.44 年。
| Scenario | Sales | Contribution margin | Maintenance capex | Annual cash contribution | Simple payback |
|---|---|---|---|---|---|
| Base assumption | $6.40m | 15% | $0.12m | $0.84m | 2.44 years |
| Slower maturity | $5.60m | 15% | $0.12m | $0.72m | 2.85 years |
| Margin pressure | $6.40m | 11% | $0.12m | $0.584m | 3.51 years |
| Combined downside | $5.60m | 11% | $0.12m | $0.496m | 4.13 years |
| 情景 | 销售额 | 贡献利润率 | 维护性开支 | 年度现金贡献 | 简单回收期 |
|---|---|---|---|---|---|
| 基准假设 | 640 万美元 | 15% | 12 万美元 | 84 万美元 | 2.44 年 |
| 成熟速度较慢 | 560 万美元 | 15% | 12 万美元 | 72 万美元 | 2.85 年 |
| 利润率受压 | 640 万美元 | 11% | 12 万美元 | 58.4 万美元 | 3.51 年 |
| 组合下行情景 | 560 万美元 | 11% | 12 万美元 | 49.6 万美元 | 4.13 年 |
A 2.44-year base payback supports investigation, not proof of value creation. It omits ramp timing, taxes, lease liabilities, central overhead, failed sites, refurbishments, inflation, and the cost of capital. Calling it return on invested capital would overstate the completeness of both numerator and denominator.
这个计算能说明什么、不能说明什么基准回收期 2.44 年只支持继续调查，不能证明公司层面已经创造价值。这个初筛没有完整纳入爬坡时间、税费、租赁负债、总部费用、失败门店、后续翻新、通胀和资本成本。若把它称为投入资本回报率，就会让不完整的分子与分母显得过度完整。
Try to break the thesis before you summarize it在总结之前，先主动击穿命题
Total sales can rise while mature locations weaken. Look for comparable-store measures, openings by quarter, and enough history to separate ramping stores from established ones.
批次掩盖总销售额可能在成熟门店走弱时继续增长。应寻找同店指标、按季度开店数据和足够长历史，把爬坡门店与成熟门店分开。
Management may change which stores enter the comparable base or which costs enter contribution. Recalculate prior periods under one definition when the disclosure permits.
定义漂移管理层可能改变纳入同店基数的门店，或改变贡献利润包含的成本。披露允许时，应按统一定义重算前期。
Landlord allowances, leases, inventory credit, and corporate systems can make store cash needs look smaller without removing the economic commitment.
隐藏资本房东补贴、租赁、存货信用和公司系统可能让门店现金需求看起来更小，却没有消除经济承诺。
A mature-store average may exclude failed locations. Add closures, impairments, and abandoned openings to the denominator of program success.
幸存者偏差成熟门店平均值可能排除了失败门店。评价扩张计划时，应把关店、减值和放弃开业项目加入分母。
Use a disconfirmation table with four columns: thesis, strongest support, credible alternative, and next observable test. Here, “base payback is below three years” supports the claim; “the margin is an unverified target excluding shared costs” challenges it; “publish cohorts by opening year and reconcile contribution to operating profit” is the next test. More notes do not equal stronger evidence.
使用反证表：第一列写命题，第二列写最强支持观察，第三列写最可信的替代解释，第四列写下一项可观察检验。对本案例而言，“基准回收期低于三年”是支持；“利润率只是未经验证的目标，而且排除了共享成本”是替代解释；“按开业年份披露成熟批次，并把贡献利润调节到营业利润”是下一项检验。这样可以避免把更多笔记页数误当成更强证据。
Write a verdict that can survive the next filing写出经得住下一份财报的结论
Date the closing memo and use five labels: Finding, Support, Counterevidence, Unknown, and Update trigger. Avoid “great company” or “cheap stock” unless business quality and valuation have been separately defined. Company analysis and a price decision are connected but not identical; sound reinvestment can be unattractive at a price that assumes flawless execution.
有用的结论备忘录可以很短，因为背后的证据工作已经可见。写明日期，并固定使用五个标签：发现、支持证据、反方证据、未知事项、更新触发器。除非已经分别定义企业质量和估值，否则不要直接写“优秀公司”或“便宜股票”。公司分析与价格决策互相关联，却并不相同；即使再投资合理，当股价已经假设执行毫无失误时，也未必有吸引力。
Finding: Harbor & Hearth’s disclosed setup costs and our base mature-store assumptions produce a 2.44-year simple payback, but this is not yet evidence of achieved cohort returns. Support: store counts reconcile and the modeled base case clears the stated three-year hurdle. Counterevidence: the combined downside stretches payback to 4.13 years, and closures may bias surviving-store economics upward. Unknown: actual cohort margin, ramp duration, shared costs, lease economics, and failed-site capital. Update trigger: revise after the next filing provides openings, closures, comparable-store sales, capex components, or a changed expansion target.
教学性最终备忘录发现：Harbor & Hearth 披露的开店成本与研究者的成熟门店基准假设，得到 2.44 年简单回收期，但这还不是门店批次已实现回报的证据。支持：门店数量能够勾稽，基准模型通过假设的三年门槛。反方证据：组合下行情景把回收期拉长至 4.13 年，关店还可能使幸存门店经济性偏高。未知：实际批次利润率、爬坡时长、共享成本、租赁经济性和失败选址资本。更新触发器：下一份披露出现新开与关闭数量、同店销售、资本开支组成或扩张目标变化时重做分析。
Assign each conclusion a status—supported, contradicted, or unresolved—not a bullish or bearish color. Keep inputs and source locations beside the memo. When a new filing arrives, update affected facts and calculations first, then ask whether the judgment still follows. Evidence can change the thesis while the audit trail remains intact.
最后，为每条结论标记“有支持”“被反驳”或“未解决”，而不是涂成看多或看空的颜色。把电子表格输入和来源位置与备忘录放在一起。新财报出现后，先只更新受影响的事实和计算，再判断原结论是否仍然成立。可重复性由此产生：证据变化时命题随之变化，但审计轨迹始终保留。
Use uploaded files as an evidence workspace把上传文件用作证据工作区
Prepare a bounded packet: the latest filing, later quarterly or event filings, the earnings release, the relevant store or segment table, and your exact claim. Include prior-period pages for comparison. Name files with issuer, form, reporting date, currency, and units; remove irrelevant personal or confidential material.
Request reported facts, calculations, assumptions, contradictory evidence, missing information, and source locations as separate groups. Verify every material number, sign, date, unit, definition, and page reference in the uploaded originals. Reperform the payback arithmetic and reconcile management-defined measures to reported measures. If a citation is missing or periods are mixed, leave the claim unresolved.
使用 Stock Explained 前，先准备有边界的文件包：最新完整披露、之后发布的季度或重大事项文件、业绩新闻稿、相关门店数量或分部表格，以及你要检验的准确命题。加入比较所需的前期页面。文件名写明发行人、表格类型、报告日期、币种和单位，并移除与分析无关的个人或机密材料。
要求输出按已披露事实、计算、假设、反方证据、缺失信息和来源位置分组。随后打开上传的原始文件，逐项核验所有重要数字、正负号、日期、单位、定义与页码；自己重做回收期算术，并确认管理层自定义指标能否调节到最接近的报表指标。如果找不到引用，或期间被混用，就把观点标为未解决，不要自行补齐。
Upload the source set with a narrow claim, request citations and uncertainty labels, and keep the originals open while checking the result.
上传带有明确命题的来源文件，要求标注引用和不确定性，并在核验结果时始终打开原件。
Open Stock Explained打开“一眼看懂这只股票”Questions that arise during a real stock analysis实际分析一只股票时会遇到的问题
Use enough comparable history to observe the tested mechanism. Three to five years can be a starting window, but store ramps may need cohort data and cyclical companies a full cycle. Explain why your period can test the claim.
分析一只股票应该看多少年的披露？应覆盖足够长的可比历史，让所检验机制能够显现。三至五年常可作为起点，但门店爬坡可能需要批次数据，收购可能破坏可比性，周期企业则可能需要一个完整周期。关键是说明为何所选期间足以检验命题。
Do not manufacture precision. Build a range from disclosed inputs only when scopes align. Label allocation assumptions, test a wide downside, and leave the thesis unresolved if consolidated disclosures cannot support a credible unit model.
如果公司不披露单店利润率怎么办？不要制造精确感。只有在范围能够统一时，才用已披露销售额、毛利率、场地成本、人工、资本开支和门店数量建立区间。标清每项分摊假设，进行幅度足够大的下行测试；若合并披露不足以支持可信单店模型，就让命题保持未解决。
Test the business thesis first so valuation assumptions have an operating basis, then value a range of outcomes. Price matters, but starting with the quote can encourage a narrative chosen to justify it.
应该先分析估值，还是先分析经营命题？先定义并检验经营命题，让估值假设拥有经营基础；随后对一组结果估值，而不是只估一个预测。价格对最终判断仍然重要，但从当前报价出发，容易让研究者选择一套替价格辩护的叙述。
No. Simple payback omits cash-flow timing, post-payback cash, financing, taxes, leases, central costs, and the cost of capital. Use it to expose sensitivities, then connect unit economics to consolidated cash and per-share results.
简单回收期足以支持扩张命题吗？不足。简单回收期只是初筛计算，没有反映回收期内的具体现金时点、回收后的现金流、融资、税费、租赁承诺、总部成本和资本成本。它适合暴露关键敏感项，之后仍要把单店经济性连接到合并现金和每股结果。
Stop when a missing input controls the conclusion and no defensible range exists—for example, an undefined contribution measure or irreconcilable store count. Record the missing source and the event that could resolve it.
什么时候应该停止分析，而不是继续估算？当缺失输入决定结论，而且无法建立有依据的区间时就应停止。例如贡献指标没有定义、门店数量无法勾稽、单位不明，或文件包缺少债务与租赁承诺。把缺失来源和可能解决问题的未来事件记录下来。
Primary sources and analytical boundary第一手来源与分析边界
The SEC materials explain where U.S. filings can be found, what a Form 10-K contains, and why company-defined measures need labeling and reconciliation. They do not validate the fictional case or prescribe a conclusion. For a real company, use current issuer filings and its home regulator; check later amendments and events before relying on an earlier document.
This guide teaches an analytical process, not individualized investment advice. It does not account for your objectives, risk capacity, taxes, portfolio concentration, trading costs, or time horizon. A defensible company thesis can still lead to loss, and a source-grounded analysis cannot remove market, business, accounting, or estimation uncertainty.
SEC 资料说明了在哪里查找美国公众公司披露、Form 10-K 通常包含哪些主要信息，以及为何公司自定义指标需要清楚标注并进行调节。它们并不验证本文虚构案例，也不规定投资结论。分析真实公司时，应使用发行人的当前披露和所属市场监管规则，并在依赖早期文件前核验之后的修订与事件。
本指南讲解分析过程，不构成个性化投资建议，也没有考虑你的目标、风险承受力、税务、组合集中度、交易成本或时间范围。即使公司命题论证充分，投资仍可能亏损；以来源为依据的分析也不能消除市场、经营、会计和估计不确定性。