Watch the evidence that can change the claim关注能够改变主张的证据
Before the next report, return to the latest annual and quarterly filings, earnings release, presentation, and call. Extract each unresolved question or dated commitment. Convert it into five fields: claim, metric, primary source, threshold, and interpretation. Label the metric leading or lagging, point-in-time or flow, quarterly or year-to-date, GAAP or management-defined. Choose comparisons that respect seasonality and base effects. A threshold may come from company guidance, contract terms, comparable periods, or an analyst assumption; never blur those origins. Prewrite three outcomes—evidence that supports, leaves ambiguous, or weakens the claim—and specify cross-checks in margins, cash, inventory, debt, or notes. Lock the record with a timestamp. After release, fill only sourced actuals and calculations, without moving thresholds. The output is a falsifiable update plan, not an earnings prediction or post-release checklist.
在下一份财报前，回到最新年报、季报、业绩稿、演示与电话会，提取每个未决问题或带日期的管理层承诺，再转成五字段记录：claim、metric、primary source、threshold 与 interpretation。注明指标是领先还是滞后、时点还是期间流量、单季还是年初至今、GAAP 还是管理层定义。比较期间必须尊重季节性与基数效应。阈值可以来自明确公司指引、合同条款、历史可比期或分析者假设，但来源不能混淆。提前写好三种结果：支持、模糊或削弱主张，并指定利润率、现金、库存、债务或附注交叉检查。给记录加时间戳并锁定。发布后只填有来源的实际值与计算，不得为了保护论点移动阈值。最终产物是可证伪更新计划，不是业绩预测，也不是发布后通用清单。
Begin with open loops, not a fresh list of ratios从未闭环问题开始，而不是新列一排比率
SilverRim Cans is an invented beverage-can manufacturer. All Q2 20X5 facts, contracts, commitments, thresholds, and future scenarios are teaching data in US dollars; volumes are billions of cans. No future actual has been invented.
虚构证据集SilverRim Cans 是虚构饮料罐制造商。所有 20X5 Q2 事实、合同、承诺、阈值和未来情景均为美元教学数据；销量单位为十亿只罐。未来实际值没有被编造。
| Open loop from Q2 | Evidence already known | Question for Q3 |
|---|---|---|
| Demand versus price | 1.20bn cans; $180m net sales; contract escalator effective July 1 | Did volume hold while contracted price reached reported revenue? |
| Aluminum pass-through | $82 aluminum cost per thousand cans; supplier increase of $5 scheduled | Did realized pricing offset the cost step after contractual lag? |
| New-line ramp | 92.0% production yield; management committed to above 94% in Q3 | Did yield improve without hidden scrap, overtime, or quality cost? |
| Seasonal inventory | $180m Q2 ending inventory; management target below $160m at Q3 end | Did cans ship and cash release, or did stock age? |
| Expansion and funding | $120m first-half capex; $180m–$200m full-year guide; $310m debt | Was the line commissioned within budget, and did debt begin to fall? |
| Full-year outlook | Volume growth 4%–6%; adjusted EBITDA $150m–$165m | Were ranges maintained, changed, or redefined? |
| Q2 未闭环事项 | 已知证据 | Q3 问题 |
|---|---|---|
| 需求与提价 | 销量 12.0 亿只；净销售 1.80 亿；合同调价 7 月 1 日生效 | 销量能否保持，合同提价是否进入报告收入？ |
| 铝成本传导 | 每千只铝成本 82 美元；供应商计划上调 5 美元 | 考虑合同滞后后，实现价格是否抵消成本台阶？ |
| 新产线爬坡 | 生产良率 92.0%；管理层承诺 Q3 高于 94% | 良率是否改善，而没有隐藏废料、加班或质量成本？ |
| 季节库存 | Q2 末库存 1.80 亿；管理层目标 Q3 末低于 1.60 亿 | 罐体是否出货并释放现金，还是库存老化？ |
| 扩产与融资 | 上半年 capex 1.20 亿；全年指引 1.80 亿–2.00 亿；债务 3.10 亿 | 产线是否按预算投产，债务是否开始下降？ |
| 全年展望 | 销量增长 4%–6%；调整后 EBITDA 1.50 亿–1.65 亿 | 区间是维持、改变，还是重新定义？ |
The open loops form a mechanism: contracts set price, aluminum and yield shape unit cost, volume converts inventory, capex creates capacity, and debt funds timing gaps. A “revenue and EPS” watch would miss whether these links worked. Record the Q2 file, page, speaker, period, and unit behind each premise. If a commitment cannot be located in the original, downgrade it to an unverified note.
这些未闭环事项形成一条机制：合同决定价格，铝价与良率塑造单位成本，销量转化库存，资本开支形成产能，债务填补时点缺口。通用“收入与 EPS”观察会漏掉这些连接是否成立。每个前提都要记录 Q2 原文件、页码、发言人、期间与单位。若无法在原件定位承诺，建立检验前先降级为未验证笔记。
Write every observation as a five-field contract把每个观察项写成五字段合约
| Field | SilverRim example | Control question |
|---|---|---|
| Claim | The new line can lift usable output without destroying unit economics | Is this a management claim, reported fact, or analyst hypothesis? |
| Metric | Q3 production yield; scrap cost; volume; gross margin | Does one metric need a second metric to prevent gaming? |
| Source | Q3 release plus 10-Q inventory/cost note and cash-flow statement | Where exactly should the evidence appear? |
| Threshold | >94% yield is management’s commitment; margin cross-check is analyst-designed | Who set the threshold, on what date and period? |
| Interpretation | Above 94% with stable cost supports; improvement with weak margin is ambiguous; below 92% weakens | Were all three meanings written before release? |
| 字段 | SilverRim 示例 | 控制问题 |
|---|---|---|
| Claim | 新产线能提高合格产出而不破坏单位经济 | 这是管理层主张、报告事实还是分析假设？ |
| Metric | Q3 生产良率、废料成本、销量、毛利率 | 是否需要第二指标防止单指标被优化？ |
| Source | Q3 业绩稿、10-Q 库存/成本附注与现金流量表 | 证据应准确出现在哪里？ |
| Threshold | >94% 是管理层承诺；利润率交叉检查由分析者设计 | 谁在何日为哪个期间设定阈值？ |
| Interpretation | 高于 94% 且成本稳定则支持；良率改善但利润率弱则模糊；低于 92% 则削弱 | 三种含义是否在发布前写好？ |
Q2 volume, sales, yield, inventory, capex, debt, and guidance are fictional source inputs that must remain unchanged.
报告事实Q2 销量、销售、良率、库存、capex、债务与指引是虚构来源输入，必须保持不变。
Q2 net sales per thousand cans = $180m ÷ 1.20bn × 1,000 = $150; aluminum spread is $150 − $82 = $68.
计算Q2 每千只净销售=1.80 亿÷12.0 亿×1,000=150 美元；铝价差=150−82=68 美元。
An analyst threshold such as $67 spread or Q3 debt no higher than $310m is not company guidance.
假设67 美元价差或 Q3 债务不高于 3.10 亿等分析阈值，并非公司指引。
Whether a miss is timing or structural depends on cross-evidence; the threshold alone cannot decide.
判断未达标属于时点还是结构问题，要靠交叉证据判断，阈值本身不能决定。
Thresholds are decision rules, not forecasts. Use a company commitment where one exists; otherwise justify an analyst threshold from a comparable quarter, physical capacity, contract, or thesis break point. Avoid false precision. Use a band when measurement is noisy and state how rounding affects boundaries. Do not change a threshold after release unless a documented restatement changes its historical input; preserve the original and revision reason.
阈值是判断规则，不是预测。有公司承诺就优先使用；没有时，分析阈值应来自上年可比季、物理产能、合同或论点失效点，并写明依据。避免虚假精确；测量噪声较大时使用区间，并说明舍入如何影响边界。发布后不能随意移动阈值，除非有正式重述改变历史输入；即便如此也要保留原版本与修订原因。
Prewrite the volume–price–metal–yield chain预写销量—价格—铝材—良率链
| Test | Support zone | Ambiguous zone | Weakening zone |
|---|---|---|---|
| Q3 volume; prior Q3 1.28bn | ≥1.33bn; about ≥3.9% year-over-year | 1.25bn to <1.33bn; inspect mix and timing | <1.25bn without documented shipment shift |
| Net sales per thousand; Q2 $150 | ≥$154 with stable contract mix | $150 to <$154; partial realization or mix | <$150 despite July escalator |
| Sales less aluminum cost per thousand; Q2 $68 | ≥$68 | $65 to <$68; expected lag may explain | <$65 without offsetting documented benefit |
| Production yield; Q2 92% | >94% plus stable scrap/quality cost | 92% to 94%, or >94% with cost contradiction | <92% |
| Gross margin | Improves with yield and spread | Stable while ramp costs offset gains | Falls while management claims full price and yield progress |
| 检验 | 支持区 | 模糊区 | 削弱区 |
|---|---|---|---|
| Q3 销量；上年 Q3 12.8 亿 | ≥13.3 亿；同比约 ≥3.9% | 12.5 亿至 <13.3 亿；检查组合与时点 | <12.5 亿且没有有文件支持的出货后移 |
| 每千只净销售；Q2 为 150 美元 | ≥154 美元且合同组合稳定 | 150 至 <154 美元；部分实现或组合影响 | 7 月调价后仍 <150 美元 |
| 每千只销售减铝成本；Q2 为 68 美元 | ≥68 美元 | 65 至 <68 美元；预期滞后可能解释 | <65 美元且无有文件支持的抵消利益 |
| 生产良率；Q2 为 92% | >94% 且废料/质量成本稳定 | 92% 至 94%，或 >94% 但成本反证 | <92% |
| 毛利率 | 随良率与价差改善 | 爬坡成本抵消收益，利润率稳定 | 管理层声称提价与良率完全实现但利润率下降 |
Do not read these rows independently. Volume above 1.33bn with net sales per thousand below $150 may indicate mix, rebates, contract lag, or price leakage. Price realization without aluminum spread can leave profit unchanged. Yield above 94% with rising quality claims may shift rather than solve cost. Check revenue disaggregation, cost of sales, hedge and purchase-commitment notes, quality accruals, and plant commentary. The analytical spread is not reported gross margin: labor, coating, energy, freight, and depreciation remain.
不能逐行孤立阅读。销量高于 13.3 亿而每千只净销售低于 150 美元，可能反映组合、返利、合同滞后或价格泄漏；价格实现但铝价差没有改善，利润可能不变；良率高于 94% 但质量索赔上升，可能只是移动成本。应联读收入拆分、销售成本讨论、套期与采购承诺附注、保修或质量应计和工厂评论。价差公式只是分析构造，不是报告毛利率，因为人工、涂层、能源、运费与折旧仍未扣除。
Make inventory, capex, and debt cross-examine the story让库存、capex 与债务交叉询问经营故事
| Claim and source period | Pre-release threshold | Required cross-check | Future actual |
|---|---|---|---|
| Q2 inventory $180m should convert by Q3 end | ≤$160m supports; $160m < inventory ≤ $180m ambiguous; >$180m weakens | Volume, receivables, payables, inventory reserve, OCF | Enter the actual after the Q3 filing |
| First-half capex $120m funds September 30 startup | Line on time and Q3 YTD capex ≤$180m supports; on time with $180m < capex ≤ $200m mixed; late or >$200m weakens | Construction-in-progress, commitments, Q3 capex = Q3 YTD − $120m | Enter the actual after the Q3 filing |
| Q2 debt $310m should fall as inventory converts | ≤$310m with inventory decline supports; $310m < debt ≤ $330m mixed; >$330m weakens | Cash, revolver, covenants, OCF, payables and capex | Enter the actual after the Q3 filing |
| Year-end debt commitment below $280m | No Q3 pass/fail; require a quantified bridge to year-end | Maturities, projected OCF, capex remaining, working-capital release | Pending until the year-end commitment can be tested |
| 主张与来源期间 | 发布前阈值 | 必须交叉核验 | 未来实际 |
|---|---|---|---|
| Q2 库存 1.80 亿应在 Q3 末转化 | ≤1.60 亿支持；>1.60 亿至 1.80 亿模糊；>1.80 亿削弱 | 销量、应收、应付、存货准备、OCF | Q3 文件后填写 |
| 上半年 capex 1.20 亿支持 9 月 30 日投产 | 按时且 Q3 YTD capex ≤1.80 亿支持；按时但 >1.80 亿至 2.00 亿混合；延期或 >2.00 亿削弱 | 在建工程、承诺、Q3 capex=Q3 YTD−1.20 亿 | Q3 文件后填写 |
| Q2 债务 3.10 亿应随库存转化下降 | 库存下降且债务 ≤3.10 亿支持；>3.10 亿至 3.30 亿混合；>3.30 亿削弱 | 现金、循环信贷、契约、OCF、应付与 capex | Q3 文件后填写 |
| 年末债务承诺低于 2.80 亿 | Q3 不直接判定；要求量化到年末的桥 | 到期、预计 OCF、剩余 capex、营运资金释放 | 待定；年末检验承诺 |
Inventory and debt are point-in-time balances; sales, capex, and cash flow are period flows. An inventory decline with weak volume may be liquidation, not successful demand. Higher debt can reflect capex timing rather than working-capital failure. If Q3 year-to-date capex is $175m, standalone Q3 capex is $175m − $120m = $55m, assuming compatible classifications. Do not subtract quarterly from year-to-date amounts or confuse committed capex with cash paid.
库存是时点余额，销售、capex 与现金流是期间流量，债务又是时点要求，必须尊重这些属性。库存下降但销量弱，可能是清仓而非需求成功；债务增加可能来自 capex 时点，而非营运资金失败。若 Q3 YTD capex 为 1.75 亿，则在分类兼容时，Q3 单季 capex=1.75 亿−上半年 1.20 亿=5,500 万。不能用单季数减 YTD，也不能把已承诺 capex 与已付现金混淆。
The $175m capex figure is a calculation illustration, not a scenario outcome stored as fact. The locked watch record keeps every future-actual cell blank until a cited Q3 source exists.
不编造 Q3 实际1.75 亿 capex 只是计算演示，不是被存为事实的情景结果。锁定观察记录会让所有未来实际栏保持空白，直到存在可引用 Q3 来源。
Separate early clues from reported outcomes and calendar effects分开早期线索、报告结果与日历影响
| Evidence type | SilverRim item | Use | Do not claim |
|---|---|---|---|
| Leading | Contract escalator, customer releases, line commissioning date, purchase commitments | Sets timing and possible future activity | Guaranteed revenue, margin, or cash |
| Intermediate | Production yield, scrap, finished-goods inventory, receivables | Shows conversion before full-year outcome | Durable economics from one quarter |
| Lagging | Reported volume, net sales, gross margin, OCF, debt at quarter-end | Confirms what occurred in the period or at the date | Cause without notes and cross-checks |
| Management outlook | 4%–6% volume growth and $150m–$165m adjusted EBITDA | Records management’s current expectation | A reported future fact or standardized GAAP measure |
| 证据类型 | SilverRim 项目 | 用途 | 不能主张 |
|---|---|---|---|
| 领先 | 合同调价、客户 releases、产线投产日、采购承诺 | 设定时点与潜在未来活动 | 保证收入、利润率或现金 |
| 中间 | 生产良率、废料、成品库存、应收 | 在全年结果前显示转化 | 用一个季度证明持久经济性 |
| 滞后 | 报告销量、净销售、毛利率、OCF、季末债务 | 确认期间内或日期上的实际发生 | 缺少附注和交叉检查时证明原因 |
| 管理层展望 | 销量增长 4%–6% 与调整后 EBITDA 1.50 亿–1.65 亿 | 记录管理层当前预期 | 报告未来事实或统一 GAAP 指标 |
Compare Q3 volume first with prior Q3 because beverage demand is seasonal; sequential Q2 comparison answers another question. Count shipping days and fiscal weeks, and identify acquisitions, outages, strikes, or customer calendar shifts. A low prior-year base can inflate percentage growth, so retain absolute volume. Keep quarterly revenue and margin separate from year-to-date capex and cash. For guidance, compare range, midpoint, definition, period, and assumptions—not just “raised” or “maintained.”
饮料需求有季节性，Q3 销量应首先与上年 Q3 比较；对 Q2 环比回答的是另一个问题。要计算发货日与财政周数，并识别收购、停产、罢工或客户日历移动。上年低基数会放大百分比增长，所以保留绝对销量。单季收入和利润率要与 YTD capex、现金分开。指引要比较区间、中点、定义、期间与假设，而不只写“上调”或“维持”。
Lock three outcome meanings before the release发布前锁定三种结果含义
Volume clears the comparable threshold, price offsets aluminum, yield exceeds 94% without cost contradiction, inventory falls below $160m, the line starts within the declared budget, and debt does not rise.
支持情景销量超过可比阈值，提价抵消铝成本，良率高于 94% 且无成本反证，库存降至 1.60 亿以下，产线在声明预算内投产，债务不再上升。
Some operating thresholds pass, but mix, lag, higher capex, receivables, or margin contradicts them. Keep claims open and request the specific bridge.
模糊情景部分经营阈值通过，但组合、滞后、更高 capex、应收或利润率形成反证。保持主张开放，并要求具体桥接。
Volume, spread, and yield miss; inventory or debt rises; startup slips or exceeds the full-year capex ceiling; guidance falls or definitions become less comparable.
削弱情景销量、价差与良率未达标，库存或债务上升，投产延期或突破全年 capex 上限，指引下降或定义变得不可比。
At publication, duplicate the locked record into a results version. Fill actual, source locator, calculation, and exception; never overwrite the threshold. A boundary result stays in its prewritten band. Label any new management explanation and identify corroboration. The final update should say which claim changed and why: for example, “yield commitment met, but aluminum spread and inventory conversion did not confirm the expected economics.” It should not convert one outcome into a buy-or-sell instruction.
发布时复制锁定记录，生成结果版本。填写实际值、来源定位、计算与例外，绝不覆盖阈值。恰好落在边界的结果仍按预写区间归类。管理层提供新解释时，标成解释并寻找支持。最终更新应说明哪项主张为何改变，例如“良率承诺达成，但铝价差与库存转化没有确认预期经济性”，不能把单一结果转成买卖指令。
Upload the old evidence before building the new test建立新检验前，先上传旧证据
Upload the latest 10-K and 10-Q, prior-year comparable quarter, earnings releases, presentations, call materials, guidance and non-GAAP reconciliations, pricing and customer-concentration disclosures, aluminum purchase and hedge notes, inventory reserves, construction-in-progress tables, capex commitments, debt and covenant notes, cash flow statements, and later amendments. Preserve filenames and filing dates. Tell Stock Explained the fiscal calendar, units, seasonality, claim owner, and lock cutoff. Do not ask it for live expectations or future actuals.
Request one row per claim with metric, source locator, threshold origin, period, unit, leading/lagging and point-in-time/flow labels, three prewritten interpretations, cross-check, and blank future-actual field. Separate reported fact, calculation, management commitment, analyst assumption, and judgment. After release, upload the filing package as a new version and request a difference report. Open each material locator in both originals; recalculate per-thousand amounts, growth, margins, YTD-to-quarter bridges, capex, inventory, and debt. Verify guidance definitions, contract timing, hedges, quality accruals, and amendments. Never move a locked threshold silently.
上传最新 10-K 与 10-Q、上年可比季、业绩稿和演示、完整电话会材料、指引与非 GAAP 调节、合同定价与客户集中披露、铝采购和套期附注、库存与准备附注、物业和在建工程表、capex 承诺、债务与契约附注、现金流量表及后续修订。保留文件名与申报日。向 Stock Explained 说明财政日历、单位、季节性、主张负责人和观察记录锁定截止时间。不要要求实时预期或未来实际。
要求每项主张一行，包含指标、主要来源定位、阈值来源、准确期间、单位、领先/滞后标签、时点/流量标签、三种预写解释、交叉检查与空白未来实际栏；报告事实、计算、管理层承诺、分析假设与判断必须分列。发布后把新申报包作为新版本上传并生成差异报告。逐个打开新旧原件重大定位，复算每千只金额、增长、利润率、YTD 到单季桥、capex、库存和债务，核验指引定义、合同时点、套期、质量应计与修订。不能静默移动锁定阈值。
Upload prior filings, commitments, and operating notes; lock thresholds and three interpretations before adding the next report as a separate version.
上传此前申报、承诺与经营附注，在把下一份财报作为独立版本加入前，锁定阈值与三种解释。
Open Stock Explained打开“一眼看懂这只股票”Questions before the next numbers arrive下一组数字出现前的常见问题
No. A threshold is a prewritten rule for how evidence changes a claim. SilverRim’s >94% yield is a management commitment; the $67–$68 spread bands are analyst-designed. Label origin, period, and rationale.
阈值等于业绩预测吗？不等于。阈值是证据怎样改变主张的预写规则。SilverRim 的良率 >94% 是管理层承诺，67–68 美元价差区间由分析者设计。必须标明来源、期间与理由。
Use the smallest set that can confirm or falsify the operating mechanism. For SilverRim, volume alone is insufficient; price, aluminum, yield, inventory, capex, debt, and guidance cross-check one another. Remove metrics that cannot change a claim.
应该观察多少项？选择能确认或证伪经营机制的最小集合。对 SilverRim，只有销量不够；价格、铝材、良率、库存、capex、债务与指引相互核验。不能改变任何主张的指标应删除。
Match the business and claim. Seasonal Q3 volume should lead with prior Q3; sequential change can still explain ramp direction. Retain absolute amounts and fiscal-day differences so a low base or calendar shift does not manufacture growth.
阈值应使用环比还是同比？取决于业务与主张。季节性 Q3 销量首先对比上年 Q3，环比仍可解释爬坡方向。保留绝对金额与财政日差异，避免低基数或日历移动制造增长。
Do not force it into the old threshold. Preserve the old definition, reconcile or recast history if the company supplies support, and mark comparability broken until a bridge exists. Definition change is itself an observation.
如果管理层改变 KPI 定义怎么办？不能强行套入旧阈值。保留旧定义；若公司提供支持则调节或重述历史；没有桥时标记可比性中断。定义变化本身就是观察结果。
Create a new version, but never overwrite the locked one. Fill sourced actuals first, document any restatement, and explain every new test. This prevents hindsight from turning a failed threshold into an apparent success.
财报发布后可以修改观察清单吗？可以建立新版本，但绝不能覆盖锁定版本。先填有来源实际值，记录任何重述，并解释每个新检验，防止后见偏差把失败阈值改成表面成功。
Primary sources for constructing the evidence record构建证据记录的一手来源
The sources support filing locations, periodic evidence, event disclosures, EDGAR retrieval, and non-GAAP reconciliation principles. Exact reporting, contract, guidance, hedge, covenant, and accounting treatment depends on issuer, jurisdiction, facts, dates, and effective rules. They do not validate fictional SilverRim or analyst thresholds.
This page is educational, not accounting, legal, tax, forecasting, market-data, or individualized investment advice. Passing a watch threshold does not establish fair value, future performance, or a trading decision.
以上来源支持申报定位、定期证据、事件披露、EDGAR 检索与非 GAAP 调节原则。准确报告、合同、指引、套期、契约与会计处理取决于发行人、司法辖区、事实、日期与有效规则；来源不验证虚构 SilverRim 或分析阈值。
本页仅供学习，不构成会计、法律、税务、预测、市场数据或个性化投资建议。通过某个观察阈值不能确立公允价值、未来业绩或交易决定。