The path · The Lab

The Lab: four reps, checked against the rubric

Lessons teach; reps train. The Lab is four defined reps, each built on a fixed dataset — the same archived edition, the same fictional company — so the work is repeatable and you can compare your second attempt against your first. Each rep states its objective, the lessons to have under you first, step-by-step instructions, a response template, a worked example, and the rubric a satisfactory rep is checked against.

How checking works here: there is no grader. Nobody — and no system — marks a Lab rep. You check your own work against the rep’s rubric, line by line, and Rep 4 turns that check into the work itself. A logged rep is self-attested: the record says you wrote it and checked it, nothing more. Four distinct completed reps, together with the core, your track and the finish, are what a certificate certifies.

Drafts save themselves in this browser as you type, so a rep can be started on a lunch break and finished on Sunday. One rep a week is the rhythm the path suggests; the log below is the record.

Checking your reps…

Reps live in this browser, like all your path progress. The counter counts distinct reps — logging Rep 1 twice is one rep done, twice practiced.

Rep 1 — The five-line paper plan Trade

Objective: write a complete five-line paper trade plan for one name, using only facts printed in a fixed archived edition — the plan Lesson T6 teaches, produced under the Lab’s conditions: before any order exists, on paper, in about twenty minutes.

Before you start

The foundation’s risk folder — Lesson 18, Risk per Trade, Lesson 19, Position Sizing, Lesson 20, Stops and Exits and Lesson 22, Your Risk Plan — plus the Trade track’s Lesson T1, Earnings and Catalysts and Lesson T6, The Trader’s Process, where the five lines are defined.

The dataset

One fixed edition: the morning edition of Friday, October 16, 2026. Its Engine List prints 23 names with last price, the day’s change, volume against average, and each name’s distance from its 50-day average; its futures read prints index levels. Pick one name from that list and build the whole plan from what the edition prints about it. The worked example below uses IBM, so your rep should use a different name — the exercise is the writing, not the copying.

Instructions

Open the edition and choose one Engine List name. Note its last price, its 50-day average if the edition prints one (or its distance from it), and anything the edition says about its calendar — IBM, for instance, reports on October 21.
Write Line 1, the idea, in one sentence of structure language, with the calendar checked: what reports, and when, while this plan would be running.
Write Line 2, the level where the idea is wrong — on a close, per Lesson 20, using a level the edition actually prints.
Write Line 3, the size, with the arithmetic shown: a hypothetical planned-risk budget divided by the distance from the reference price to your level, rounded down to whole shares. State the capital the position would require as a separate number. Use a fictional account and say so.
Write Line 4, both exits, each with an execution note (order type, session, price basis) — including what a gap through your level would do to the tidy arithmetic.
Write Line 5, the skip condition: the specific circumstance under which this plan, however attractive, is not taken.
Check all five lines against the rubric below, fix what fails, then log the rep.

Template

Line 1 — Idea: [name] + [the structure fact from the edition, one sentence] + [calendar checked: what reports, and when]
Line 2 — Level: the idea is wrong on a close [below/above] $[level the edition prints]
Line 3 — Size: planned risk $[budget] ÷ $[distance to the level] = [shares, rounded down]; capital required $[shares × price] — planned risk is a budget, not a maximum loss
Line 4 — Exits: [the exit on the level, with order/session note]; [the second exit — time, target, or the earnings line]
Line 5 — Skip if: [the specific circumstance that cancels this plan]

Worked example — IBM, from the same edition

A satisfactory rep, written against the October 16 edition’s own figures. Paper only; the account is fictional.

Line 1 — Idea: IBM closed Thursday at $234.12, up 2.7%, reclaiming its 50-day average at $232.06 — a working setup in the edition’s words — five days before its October 21 earnings report, which the calendar check puts inside this plan’s life.

Line 2 — Level: the idea is wrong on a close below $232.06, the 50-day line the edition printed. A close, not an intraday dip.

Line 3 — Size: fictional $25,000 paper account, planned risk 0.5% = $125. The distance from $234.12 to $232.06 is $2.06 a share, so 60 whole shares (60 × $2.06 = $123.60 of planned risk, rounded down). The capital required would be 60 × $234.12 = $14,047.20 — a different number from the risk. Planned risk is a budget, not a maximum loss: a gap through the level can lose more.

Line 4 — Exits: first, a close below $232.06 ends the plan at the next session’s open, regular hours, market order on paper — the fill is whatever the open prints, gap included. Second, win or lose, the plan ends at the close on October 21, before the after-close report: no paper position is held through the print.

Line 5 — Skip if: IBM gaps below $232.06 before the plan can start, the report date moves into the plan’s first day, or I cannot see a close on any day it runs — a plan that cannot be watched is not taken.

Rubric — a satisfactory Rep 1

All five lines present, labeled, in order.
Line 1 names one name and one structure fact the edition prints, and the calendar is checked in the same line.
Line 2 is a close-based level taken from the edition — not an intraday trigger, not an invented price.
Line 3 shows the division, rounds down, and keeps planned risk, capital required, and actual loss as three different things — no line promises a maximum loss.
Line 4 carries two exits, each with an execution note, and says what a gap does to the plan.
Line 5 names a specific, checkable circumstance — not “if it looks bad.”

Common mistakes

“My stop guarantees I can’t lose more than $125.” Correction: a stop is a trigger, not a promise. Planned risk is the budget you size from; gaps and slippage can carry the actual loss past it (Lessons 18–20).
Treating an intraday dip below the level as the exit. Correction: the plan’s Line 2 says a close. Grade and execute the line you wrote, not a twitch.
Sizing from capital instead of risk. “$14,000 worth” is not a size. Correction: Lesson 19’s division — budget ÷ distance, rounded down; capital required is reported separately.
No calendar in Line 1. Correction: the catalyst check is part of the idea line. IBM reports October 21; a plan silent about that has not been checked.

Your Rep 1

The Paper Book — follow your Rep 1 plan Follow-through

What this is: Rep 1 ends with a five-line plan written against one fixed edition. A plan is not meant to end on the page it was written on. The book carries it forward: as each new edition you read prints your name’s close, log that close here with its date. The book keeps the record a paper journal keeps — the sessions followed, the last close measured against your Line 2 level, and where the plan stands in dollars and in R.

Where the closes come from: you. Nothing on this page fetches a price. Every close in the book is one you read in an edition and typed in yourself — the same self-attested record as the reps above. Read the figure off the edition and check it before you log it: a book built on a mistyped close grades the wrong plan.

The book lives in this browser, like the reps. It keeps no price of its own and sends nothing anywhere.

Rep 2 — The statements check Invest

Objective: read one fictional company’s statements the way Lessons I1–I2 teach — margins as a trend, profit tested against cash, and one deliberate hunt for the line that deserves a harder look — and write it as three findings, in about twenty minutes.

Before you start

Lesson I1, Financial Statements: Income and Lesson I2, Financial Statements: Balance Sheet and Cash Flow. The investor’s research-log form in Lesson 26, The Journal You Keep is the same skeleton pointed at a thesis.

The dataset — Meridian Industrial (fictional)

Meridian is the fictional company from Lessons I1 and I2. The margin and cash figures below are the lessons’ own where the lessons print them; the receivables and debt lines are extensions built for this rep, also fictional. Every number you need is on this page.

Fiscal year	Revenue	Gross margin	Operating margin	Net margin
2023	$39.2B	34.1%	10.9%	7.4%
2024	$42.6B	35.6%	12.1%	8.3%
2025	$45.1B	36.8%	13.0%	9.0%
2026	$48.0B	37.9%	14.2%	9.8%

2026 lines, and the balance sheet	Amount	Note
Net income (2026)	$4.7B	From Lesson I1
Operating cash flow (2026)	$7.9B	From Lesson I2
Capital expenditure (2026)	$2.6B	From Lesson I2
Accounts receivable	$3.1B (2023) → $5.4B (2026)	Extension for this rep
Interest-bearing debt / cash & equivalents (2026)	$6.5B / $2.0B	Extension for this rep

Instructions

Finding 1 — the margins’ trend. Read all four years, not the last one: each margin’s first year, last year, and the change in points, alongside revenue’s growth. Say what the trend is consistent with — and why it is still not proof of it.
Finding 2 — conversion. Compute free cash flow (operating cash flow minus capital expenditure) and divide by net income. State the ratio and what it establishes — and what it never establishes, in either direction.
Finding 3 — the red-flag hunt. Pick the one line in the dataset that grew out of step with the business — compute its growth against revenue’s — and write the question it leaves, framed as a question: a flag locates digging, it does not deliver a verdict. (Net debt, computed Lesson I2’s way, belongs in this finding or the second.)
Check the three findings against the rubric, then log the rep.

Template

Finding 1 — Margins: gross [first year]% → [last year]%, operating [ ] → [ ], net [ ] → [ ], while revenue grew [ ]%. The trend is consistent with [reading] — and is not proof, because [reason].
Finding 2 — Conversion: free cash flow $[operating cash flow − capital expenditure]B ÷ net income $[ ]B = [ratio]. Profit [did / did not] arrive as cash in 2026; that establishes [ ], and never establishes [ ].
Finding 3 — Red-flag hunt: [the line] grew [ ]% against revenue’s [ ]% ([the arithmetic]). The question it leaves: [the question, stated as a question]. Net debt, for the file: $[debt]B − $[cash]B = $[ ]B.

Worked example — Meridian, checked

A satisfactory rep, computed from the dataset above. Check its arithmetic — that, too, is the exercise.

Finding 1 — Margins: gross 34.1% → 37.9% (+3.8 points), operating 10.9% → 14.2% (+3.3), net 7.4% → 9.8% (+2.4), while revenue grew from $39.2B to $48.0B, about +22% over the four years. Every margin widened every year. The trend is consistent with pricing power and a disciplined operation — and is not proof of either, because a favorable product mix, a cheap input stretch, or costs deferred past a period’s end can print the same shape.

Finding 2 — Conversion: free cash flow = $7.9B − $2.6B = $5.3B; $5.3B ÷ $4.7B of net income = 1.13. In 2026 Meridian’s reported profit arrived as cash, with a surplus. That establishes where the year’s earnings went; it never establishes accounting quality or its absence — a ratio above one can also be a business starving its future, which is why it is read with the margins, not instead of them.

Finding 3 — Red-flag hunt: accounts receivable grew from $3.1B to $5.4B — about +74% — against revenue’s +22%; receivables were 7.9% of revenue in 2023 ($3.1B ÷ $39.2B) and 11.3% in 2026 ($5.4B ÷ $48.0B). The question it leaves: is Meridian booking sales faster than it collects them — slower-paying customers, a mix shift toward longer terms, or revenue pulled forward? The dataset cannot answer it; it tells the owner where to dig next. For the file: net debt = $6.5B − $2.0B = $4.5B, interest-bearing debt less cash and equivalents, per Lesson I2.

Rubric — a satisfactory Rep 2

Three findings, labeled, in the taught order: margins, conversion, hunt.
Finding 1 uses the whole table — first year to last, in points, next to revenue’s growth — and names what the trend is consistent with without calling it proof.
Finding 2 shows the formula and the arithmetic (free cash flow first, then the division), and states what conversion never establishes.
Finding 3 computes the suspect line’s growth against revenue’s and ends in a question, not an accusation.
Every figure in the rep can be traced to the dataset on this page.

Common mistakes

Reading the last year only. “37.9% gross margin” is a level, not a finding. Correction: the trend across all four years is the evidence; one period certifies nothing.
Calling conversion a verdict. Above one is not proof of quality and below one is not proof of misconduct. Correction: the gauge locates digging — in both directions (Lesson I2).
Treating the receivables line as proof of invented revenue. Correction: it is a growth mismatch and a question. Slower-paying customers or a mix shift explain the same numbers; the rep’s job is to ask precisely, not to convict.
Quoting margins without the revenue line. Correction: a margin trend read without the top line’s growth beside it hides the denominator. Report them together.

Your Rep 2

Rep 3 — The read: structure, or evidence Your track

Objective: produce one disciplined read — a levels-and-structure read (Trade) or an evidence assessment (Invest) — that separates what the dataset shows from what it cannot show, with the condition that would change the read stated in advance. About twenty minutes. Do Rep 1 or Rep 2 in your track first; this rep uses the same dataset again, on purpose.

Before you start

Trade: Lesson 10, Support and Resistance, Lesson 11, Market Structure and Lesson T5, Sessions and Execution (a close is a session’s verdict, not an intraday poke). Invest: Lessons I1 and I2 again, plus the falsifier form in Lesson 26.

The dataset

Trade: the same archived edition as Rep 1 — the morning edition of Friday, October 16, 2026, whose futures read prints the index levels with their distances, and whose Engine List prints each name against its 50-day average. Invest: the Meridian dataset printed in Rep 2 above — scroll up; nothing new is added, because assessing the same evidence twice is the skill.

Instructions

Choose your track in the form below — the template button loads the matching scaffold.
Trade: pick one level the edition prints (an index level from the futures read, or a name’s 50-day line). State where price sat against it, in percent; what a close through it would change; what a close back the other way would change; and what the level alone cannot tell you.
Invest: take the thesis from Lesson 26’s Meridian entry — margins hold while it grows, profit converting into cash. Marshal the dataset’s evidence for it and against it, state the falsifier in advance (Lesson 26’s own form: two consecutive periods of falling margins and conversion below one), and name the next check and the lines it will read.
In both tracks: no motives from price alone, no verdicts from gauges. Check against the rubric, then log.

Templates

The level: [level, and where the edition prints it]
Where price sits: [last price] — [distance, in %] [above/below] the level; [the edition’s own word for the position, if it gives one]
What changes the read: a close [through $ ] makes [new read]; a close [back below/above $ ] makes [other read]; between them the read stays [ ], unresolved
What this read does not know: [the motives behind the tape, the next move — name what a level cannot show]

Thesis: [one sentence]
Evidence for: [figures from the dataset]
Evidence against / open questions: [figures from the dataset]
What would change the read: [the falsifier, stated in advance, in measurable terms]
Next check: [when, and which lines I will read first]

Worked example — Trade: the S&P future at its shelf

From the October 16 edition’s futures read. The Invest form is the same discipline in Lesson 26’s shape — thesis, evidence both ways, falsifier, next check.

The level: resistance at 7,838.50 on the S&P 500 future, from the week of August 10, as the edition’s futures read prints it; support below at 7,509.25, from the week of September 14.

Where price sits: the future printed 7,805.25 that morning — 0.4% under resistance, 3.9% above support. The edition’s word is “pressing resistance,” and pressing is all it is: no close has been taken through the shelf.

What changes the read: a daily close above 7,838.50 retires “pressing” and starts the argument about whether the old shelf becomes a floor; a close below 7,509.25 fails the shelf underneath and reopens the lower map. Between the two, the honest read stays what the edition printed — pressing, unresolved.

What this read does not know: why anyone bought or sold, whether the volume was conviction or hedging, or what happens next. A level describes where price has reacted before; it assigns no motive and promises no repeat (Lessons 10–11).

Rubric — a satisfactory Rep 3

Trade: the level and the price both come from the edition, with the distance computed. Invest: every figure comes from the Meridian dataset.
The change condition is stated in advance and is checkable — a close through a named level, or the falsifier’s measurable terms.
Observation and interpretation are kept apart: what printed, then what it is read as — labeled as a reading.
The “does not know” side is filled honestly — at least one real unknown, not a disclaimer pasted on.
No motive is assigned to buyers or sellers from price alone; no gauge is treated as a verdict.

Common mistakes

Calling an intraday poke a break. Correction: this path grades structure on closes (Lessons 11, T5). A wick through a level is a visit, not a verdict.
“Buyers are in control” from one green close. Correction: price shows what traded, not who intended what. Write the observation; label any reading as a reading.
A falsifier that cannot fail. “I’d change my mind if things got worse” is not a condition. Correction: measurable terms, stated in advance — Lesson 26’s two-periods-and-below-one is the model.
Cherry-picking one side. Correction: the Invest form has an “against” line on purpose. A read with no evidence against it has not been assessed; it has been advertised.

Your Rep 3

Rep 4 — Review and grade Both tracks

Objective: grade one piece of written work against its own rubric — one of your earlier reps, or the sample below if you have none yet — and produce the corrected line. This is the Lab’s grading mechanism, made explicit: the rubric, your reading of the work against it, and one line made right. About fifteen minutes.

Before you start

At least one of Reps 1–3 logged above (or use the sample). The habit is Lesson 26’s — The Journal You Keep: written before, checked after, graded on process and reasoning, never on whether the market happened to agree.

Instructions

Pick the rep to grade and open its rubric (Reps 1–3 above).
Write what holds: at least one rubric item the work genuinely satisfies, quoting the line that satisfies it.
Write what is missing or wrong: at least one rubric item it misses, quoting the line that misses it — or naming the line that should be there and is not.
Write the corrected line: the missing or faulty line, rewritten to satisfy the rubric. One line, surgical — not a rewrite of the whole rep.
Log the review. If you graded the sample, say so in the first line so the record stays honest.

Template

Rep reviewed: [Rep N, and when it was logged — or “the sample”]
Rubric used: [that rep’s rubric]
What holds: [rubric item] — [the line that satisfies it, quoted]
What’s missing or wrong: [rubric item] — [the line that misses it, quoted, or the absence named]
The corrected line: [the line, rewritten to satisfy the rubric]

Worked example — grading a faulty plan line

The sample under review is one line from an imaginary Rep 1: “Line 3 — Size: 60 shares, stop at $232.06, so I can’t lose more than $125.”

What holds: Rep 1’s rubric asks Line 3 to show a size derived from a level and a budget — this line does carry a share count and the level it was sized from ($232.06, the edition’s 50-day line), so the skeleton of the division is present.

What’s missing or wrong: the rubric’s planning-language item fails outright. “Can’t lose more than $125” turns a planned-risk budget into a guaranteed maximum loss — the exact claim Lessons 18–20 retire. The line also omits the capital required, which the rubric keeps as a separate number.

The corrected line: “Line 3 — Size: planned risk $125 ÷ $2.06 distance = 60 whole shares (60 × $2.06 = $123.60 planned risk); capital required 60 × $234.12 = $14,047.20. The level is a trigger, not a promise — a gap through $232.06 can lose more than planned.”

Rubric — a satisfactory Rep 4

The review names the rep and the rubric it is graded against.
At least one genuine hold and one genuine gap, each tied to a rubric item and a quoted line.
The corrected line actually fixes the gap it names — read it against the rubric item and it passes.
The grade judges process, language and reasoning — never whether a trade made money or a thesis “was right” afterward.

Common mistakes

Grading the outcome. “The plan lost, so it was bad.” Correction: a plan can follow every line and lose; a sloppy one can win. The rubric grades the writing and the reasoning (Lesson 26, T6).
Vague praise. “Looks good overall” cites no item and quotes no line. Correction: name the rubric item, quote the line, then judge.
Rewriting the whole rep as the “correction.” Correction: one line, fixed. Surgery teaches; replacement just hides the fault from the record.
Inventing rubric items. Correction: grade against the rep’s printed rubric only. If the rubric itself is missing something, that is a note for the Lab, not a deduction from the learner.

Your Rep 4

How the Lab record works. Submitted reps and drafts live only in this browser’s local storage — no account, nothing uploaded, no server copy. Clearing this browser’s site data removes them; using another device does not erase this browser’s record, it simply cannot see it. Four distinct completed reps satisfy the Lab requirement on every certificate variant; the certificate page reads the same record this page keeps.

Last reviewed: October 24, 2026.