Spending Three Units in Three Installments Has No Automatic Meaning
At 10:30 Monday, Ye Cheng began returning the V2 conditions proposal to the stored Version 1 checklist. The client wanted receipt between 14:00 and 16:00 Wednesday, with no added scope, and had proposed a fee, payment in halves, and item-by-item acceptance. The station had not accepted or committed production resources.
Tang Ke suggested dividing preparation into three parts and doing only a little each time. Gu Ning asked what would become known at the end of each part. If the answer was only “we are closer to finishing everything,” staging might change the appearance of payment and scheduling without giving feedback the authority to stop later commitment.
The stages in this chapter are divided by questions, not equally by quantity. Confirmation asks whether both parties accept the same V2 and stage rules. The sample asks whether one critical connection on the checklist can pass the agreed procedure. Formal scheduling becomes eligible only after the first two questions receive the required answers.
The labor, working capital, sample identifier, and test result below remain fictional. They serve the action structure and do not represent real costs, durations, or acceptance standards for any technique.
Write Down the Complete Commitment First
Using the retrievable checklist, Ye Cheng estimated for the case that the local sample would need one half-day slot and one working-capital unit. If it passed, completing the checklist would need two more half-day slots and two more units, including the handoff occupation in Wednesday's receipt window.
The full path totals three schedulable slots and three deployable units. Monday morning has four of each available, so the point estimate appears to preserve the six existing slots and eight protected units, leaving one apparent unit of margin in each resource.
“It appears to fit” is not yet a reason to adopt it. Estimates may change, and stages may share the same material, people, or time constraint. Gu Ning first preserves the complete path so the team can see where the initial approval might lead instead of approving the first step while refusing to look at total commitment.
The First Gate Confirms the Object and Rules
Stage G0 is conditions confirmation. It occupies no half-day slot, uses no working-capital unit, and has a thirty-minute communication ceiling. Its question is whether both parties preserve Monday's V2 conditions proposal and the added staging arrangement as one retrievable version.
Confirmation covers the Version 1 checklist identifier, Wednesday 14:00–16:00 receipt, checklist total, payment in halves, item-by-item acceptance, and the rule that “a local sample comes first; passing it only makes formal work eligible for another review.” If any item is not jointly preserved, G0 does not pass.
G0 also states its stopping consequence. Without written agreement by 11:30, the station makes no sample, receives no stage payment, and reserves neither of the two later slots. Communication cannot quietly become the start of production.
The Second Gate Tests Only One Local Relation
Stage G1 is the local sample. Its object is the connection check marked S-3 on the checklist. Its question: after one local sample is made with the confirmed material and checklist procedure, can S-3 receive an adjudicable result?
The book does not invent S-3's real technical content. The case stipulates that both parties can return to the same procedure and that results are pass, fail, or procedure unable to adjudicate. The sample does not prove that the whole checklist can be completed or Wednesday's handoff will succeed.
G1 is capped at one half-day slot Monday afternoon and one working-capital unit. If the material identifier differs, the procedure cannot run, consumption would exceed the ceiling, or an existing job develops a problem, Ye Cheng stops and reopens with Gu Ning rather than modifying the sample indefinitely in pursuit of a pass.
The Third Gate Is Not an Automatic Reward
Stage G2 is formal scheduling, currently estimated at two half-day slots and two working-capital units. Entry requires at least: continuing validity of G0 confirmation, G1 passing under the original procedure, actual receipt of the first payment, a resource review that preserves protection and margin, and no changed scope or window.
Even if all conditions hold, Gu Ning must create a release record. A gate grants candidacy rather than pressing Start for the authorizer. Existing work and cash may change between completion of the sample and formal commitment.
G2's production details are not defined here, and it is not executed. Budget, shared exposure, and the waiting window require review first. Writing an unstarted stage as “estimated,” rather than “committed,” is central to preserving choice through staging.
A Stage Must Produce New Material
G0 produces a jointly retrievable version and stage rules, closing the gap in which each party believes it confirmed something different. G1 produces an actual S-3 record under specified material, procedure, and time, replacing a merely verbal assumption about local feasibility.
If later executed, G2 will produce process and handoff material for the complete checklist. But that future output cannot become the reason G1 must start. Each stage's output enters the next gate's input field instead of merely increasing a progress bar.
If completing a segment changes no condition of the next step, the team must ask whether it has merely sliced the original total. A small amount is not information gain. A stage gate has value because its result can stop, modify, or release later action.
Installment Payments Are Not Staged Decisions
Dividing a total into a first and final payment changes cash timing but may not change action commitment. If both parties unconditionally agree to the full scope when the first payment arrives, the station may lack a choice to stop G2 even before later costs are paid.
This case separately states that G1's result and resource review are conditions for entering the formal stage so that payment and action structures remain visible. The first payment covers authorized sampling and conditional later work; the unused balance has an exit treatment. It does not buy permission to ignore the sample and begin everything.
Conversely, action can be staged without installment payment. A subject can use bearable resources to perform one check and decide from its result. Staging depends on whether feedback can change commitment, not on the number of invoices.
Passing One Gate Does Not Remove Uncertainty from the Next
G0's passing shows joint records of version and stage rules. It does not make S-3 pass. G1's passing shows that the local sample obtained a result under the current procedure. It does not remove deviation from the full checklist, Wednesday timing, or handoff.
A passed stage closes some questions and makes the rest more specific. Saying “two gates have passed, so overall risk is low” needs separate material. Gate count is not a probability scale, and completion percentage cannot be translated directly into success probability.
The stage table therefore lists closed and unresolved questions. G2 still faces its own execution, other checklist items, resource conflicts, and time window. Earlier results become inputs without painting later unknowns as zero.
Stage Gates Can Also Produce Confirmation Bias
After spending one slot on a sample, participants may want it to justify continuation. If the same person adjusts the procedure, interprets ambiguity, and authorizes passage on the spot, prior commitment may affect adjudication.
Gu Ning fixes the S-3 procedure before starting. Ye Cheng executes and preserves originals; she checks the object without changing the standard. This does not guarantee perfectly unbiased observation, but leaves traces of changes and prevents “close enough” from silently becoming a pass.
If a result lies outside the procedure's categories, its proper state is unable to adjudicate. A gate is not a green-light machine. Material must be allowed to reject the current classification and force reopening.
Unreleased Resources Must Remain Available to Other Arrangements
Calling G2 unstarted while internally keeping two Tuesday slots empty already creates opportunity occupation. Truly unreleased means the schedule still shows them as available and permits other suitable work unless a separate reservation is explicit.
This creates a real cost: while G2 awaits review, another job may take those slots. Stage design preserves exit choice but may lose speed. The station cannot claim that resources are completely free while guaranteeing they are instantly available later.
Gu Ning accepts this tradeoff because the goal hierarchy first protects existing fulfillment and survival, then pursues a suitable new commission. A later reservation must state its quantity, release time, and missed opportunity; a “stage plan” cannot reserve capacity invisibly.
Staging Still Requires Total-Exposure Review
Writing three units as one plus two does not turn the total into one. If beginning G1 makes refusal of G2 impossible, exposure remains three. If two later slots have already displaced other work, calling them “unreleased” on paper preserves no capacity.
Gu Ning requires unreleased stages to stay outside the resource ledger. Tang Ke cannot tell others Tuesday is full, and Ye Cheng cannot purchase from the later two units. Only a G2 release record changes those slots and units from available to committed.
At the same time, the team must see the possible full path when approving G1. Otherwise every gate asks for “just a little more” until the total crosses the original limit. Local authorization and full-path visibility must coexist.
Payment Cannot Replace Stage Judgment
The Version 1 checklist total is six normalized working-capital units in this case. It enters Part 2 only after Monday's reply made the checklist retrievable; it is not backfilled as a price used in Saturday's judgment.
The client proposes three units after V2 confirmation and three after item-by-item receipt. The first payment can improve cash timing and creates a corresponding fulfillment obligation. All three units are not free gains.
Gu Ning makes actual receipt a necessary condition for G1 but not a sufficient condition for automatically starting G2. Money arriving does not make the sample pass in advance or release the six existing tasks and eight protected units.
State Who Bears the Sample Cost
The stage proposal assigns G1's one-unit cost to the first payment. If G1 passes and both parties continue, it enters the complete commission's costs. If it fails or cannot be adjudicated under the original procedure, the station stops and charges no extra amount for the one unit used; treatment of the remaining first payment follows the stage-confirmation terms.
To keep the case adjudicable, the fictional terms say that any balance not corresponding to executed work is returned through the original channel on the same day after stopping. This belongs only to fictional V2 and is not a real contractual or legal rule. Actual projects must follow applicable agreements.
The station therefore cannot treat a received three-unit payment as compensation in full, nor call the sample free merely because part may be refunded. Receipt, use, balance, and exit time enter the ledger separately.
Forming a Stage Proposal at 11:00
At 10:50, Gu Ning completed the draft gates. Tang Ke sent it through the original channel: the station could first confirm V2, then use one half-day slot for the S-3 sample; only an original-procedure pass, resource review, and separate release would permit two half-days of formal work.
The draft kept Wednesday's window as a V2 condition and left the two later slots unreserved. On stopping, executed work would be settled and unmatched balance returned. The client could not treat a local pass as full acceptance, and the station could not treat first-payment receipt as formal-stage approval.
Sending the draft was part of G0 and fit within its thirty-minute ceiling. It used no production slot or working-capital unit and did not promise that G1 would start.
What Counts as the Client Confirming the Same Version?
At 11:12, the client replied item by item in the original channel: it confirmed the checklist identifier, Wednesday window, total, and two-part payment; agreed to a local sample under the S-3 procedure and to separate release of formal work; and accepted settlement of executed work and return of balance if sampling stopped.
Tang Ke placed this reply beside the station's draft rather than extracting only the word “confirmed.” Gu Ning checked that no item had been added and both parties gave the stage result the same role. At 11:20, she changed V2 from “client conditions proposal” to “stages jointly confirmed.”
The confirmation commits only G0 rules and eligibility to start G1 after payment. G2 remains conditional. V2 is now jointly preserved, but this does not make all production an unconditional obligation.
Recheck Resources After the First Payment Arrives
At 11:35, the case's three-unit payment arrived through the agreed method, and Tang Ke saved the receipt. Visible cash rose from twelve to fifteen, but the total included funds connected to V2 obligations and could not justify expanding other spending.
Gu Ning released one unit from the first payment for G1. Eight-unit protection remained intact and the original four deployable units were not forced across the line. The other two payment units remained tied to later work or refund and did not become free balance in this chapter.
Ye Cheng confirmed that Monday afternoon remained available and the six existing tasks had no anomaly. At 11:40, G1 had version confirmation, received payment, material identity, and resource conditions, and received its own start authorization.
Write the Stopping Point Before Starting
G1 begins at 13:30 Monday and ends no later than 17:00. Ye Cheng follows and records the checklist procedure, Tang Ke preserves jointly confirmed files, and Gu Ning does not change the passing standard during the sample.
If completion would exceed one unit, the ceiling triggers and stops the stage. If the procedure lacks a decision condition, the result is unable to adjudicate. If the observation does not meet S-3, it fails. None authorizes adding a second unit automatically and trying again.
A pass answers only the S-3 result for this local sample under these conditions. It does not remove other unknowns in the full scope. Recording stopping points first prevents an adverse result from turning sampling into miniature full production through repeated additions.
A Sample Does Not Display Confidence
If the team uses a sample to demonstrate enthusiasm, it may select its best-looking object, change the procedure, or hide failure. G1 exists to expose one critical local relation to an actual check, not to perform confidence in exchange for commitment.
The client may receive the result record but cannot add adjudication items during execution. A new demand pauses G1 and reopens V2. The station likewise cannot skip the original procedure because handling the material feels smooth.
Information value requires that a result can oppose the plan. A sample that can only say “continue” is ritual, not a test. Failure and inability to adjudicate must truly prevent G2.
Actual Commitment Begins at 13:30
At 13:30 Monday, Ye Cheng released one of the four schedulable slots to G1 and drew the case material corresponding to one working-capital unit. Resource state changed from zero production input to one slot and one unit used.
He checked the material identifier against the confirmed version, then followed the S-3 recording sequence. Tang Ke saved time, batch identifier, and process originals without adding a prediction that it “should pass.” This chapter adopts no numerical probability for the sample result.
At 15:40, the local sample was complete within the one-unit and half-day ceilings. Completion of fabrication did not equal passing; the result required separate adjudication under the original procedure.
A Limited Pass at 16:00
At 16:00, Ye Cheng checked against S-3 and recorded conformity with its passing conditions. At 16:10, Gu Ning verified the object, procedure, and originals and marked G1 passed.
This pass is an expressly fictional result. It supports “this local sample passed S-3 under the current material and procedure.” It does not support a risk-free checklist, general feasibility of similar jobs, or certain Wednesday acceptance.
If later formal work uses different materials, environment, or steps, transfer from G1 must be checked again. One local result can enter the next gate without expanding its scope indefinitely.
Why G2 Does Not Start After the Pass
Tang Ke proposed using the rest of Monday to begin formal preparation. Gu Ning rejected the on-site upgrade. G2 still required a resource review, shared-exposure check, and separate release, and this chapter had assigned it no Monday capacity.
One slot and one unit were used, leaving three of the originally schedulable slots; two units of the first payment remained tied to later work or refund. The point estimate seemed exactly sufficient for G2, but “exactly” did not address unexpected rework, shared materials, or the waiting window.
Letting the result enter the next round means it can make G2 eligible and must also accept rejection or modification by other conditions. If a pass starts work automatically, the gate merely delays the total commitment by a few hours.
What Would Have Happened on Failure?
Gu Ning did not delete adverse paths after the pass. Had G1 failed, G2 could not be released. The team would preserve one unit and one slot of actual loss, settle executed work from the first payment, and return unmatched balance under the confirmed terms.
If the procedure could not adjudicate, “no problem found” could not be called a pass. The station might propose a revised procedure or another sample, but that would be a new stage with new information and resources; G1 would remain closed as unable to adjudicate.
Those unrealized paths were part of why G1 was bearable before starting. Keeping only the successful route afterward would falsely imply that staged action never had a stopping possibility.
One Pass Cannot Prove Staging Was Unnecessary
Because the first sample passed, someone might say direct formal work would have succeeded and the stage wasted one slot. That counterfactual has no execution evidence. G1 produced a local S-3 record and bounded exposure on an adverse path.
Staging does not need failure every time to have value. It makes failure visible earlier if it occurs and carries success forward within an explicit scope. One smooth result does not prove that an exit was unnecessary beforehand.
Nor would failure prove that staging always saves a project. The used slot and unit remain losses, and the sample may miss other failures. Stage design changes the connection between information and commitment; it does not promise costless trial and error.
Stage Gates Connect Provisional Convergences into a Cycle
Observational convergence forms temporary certainty at limited points. Stage gates implement this by locking only the resources and commitments presently needed, making each result an observation input for the next stage rather than expanding a local pass into certainty about the whole path.
G0, G1, and G2 thus form a short cycle of expectation, reality, and renewed expectation. Staged commitment lets low loss, feedback, and continued correction work together, preserving margin to stop, reduce, or change paths when new material appears.
Preserving State in the Stage Gates
At 16:20 Monday, Gu Ning updated the table. G0: joint stage confirmation, passed, actual communication within its ceiling. G1: one half-day slot and one working-capital unit used, S-3 passed under the original procedure, closed.
G2: estimated two half-day slots and two units; state “some entry conditions met, not released.” It has G0 and G1 results but still awaits budget margin, shared-exposure, waiting-window, and start-time resource checks.
Of the four schedulable slots, one has been used for the sample and three remain uncommitted. Starting cash twelve increased by the three-unit payment and fell by one for sampling, for a visible balance of fourteen. Eight remain protected, the original four remain unallocated, and the two-unit balance of the first payment must be classified in the next chapter. “Net plus two” is insufficient.
Minimum Fields for a Staged-Commitment Record
Each stage preserves at least question, input, resource ceiling, start authority, stopping condition, result type, actual use, and effect on the next gate. Without a question it is only a time period; without stopping conditions the budget expands; without a next-gate effect feedback cannot change action.
Stages must also preserve unreleased state. Estimated, reserved, used, and refundable resources differ. Writing all of them as “the project needs three units” mixes present exposure with future plans again.
G0, G1, and G2 were designed only for V2, not as a universal three-stage method. Other problems may need two or more gates. Stage count follows questions awaiting answers, not the neatness of a template.
Handing Off to the Margin Ledger at 16:20 Monday
Staging has bought two kinds of new information: whether both parties accept the same staged version and the actual S-3 result for the sample under the specified procedure. It has also produced real cost: one slot and one unit are no longer unused.
V2 is not complete, formal scheduling has not begun, and Wednesday receipt has no result. The G1 pass makes G2 a comparable candidate without carrying it over the remaining checks. First-payment receipt improves the cash position while creating execution or refund relations.
The next chapter separates the twelve-unit opening balance, three-unit first payment, one used unit, fourteen-unit current balance, eight-unit protection, two-unit first-payment balance tied to later work or refund, and remaining labor states. A budget does not mean spending whatever appears in the balance. Stage gates retain force only when the record distinguishes commitments, releasable resources, and resources that must wait.