JUDGMENT QUALITY / EVIDENCE & GROWTH

What is a high-quality decision? The Howard Capital OS standard

This is not a feature list. It is the public standard by which Howard Capital OS asks to be tested: how judgment quality is defined, compared and audited before outcomes arrive.

PUB-DECISION-QUALITY-0001V1.0PUBLIC CONTROLLEDSEE PUBLISHING TERMS2026.08.12

5 MINUTES / FIVE-MINUTE VIEW

Eight necessary attributes; ten increments share one test

The causal chain is: the standard constrains analysis; analysis creates observable increments; increments change real choices; outcomes then validate or revise the standard.

FORMAL DEFINITION / 正式定义

Outcomes validate; they do not rewrite ex-ante quality

A high-quality decision uses the information, stated or explicitly unconfirmed objectives, authority, constraints and time window available at the time to compare viable alternatives, inaction and opportunity cost; identify decisive variables, mechanisms, risks and dissent; and choose or defer an executable, stoppable and reopenable path more likely to improve long-run risk-adjusted outcomes.

The eight attributes are not equally weighted score items. They are necessary gates: material failure in goal, problem, facts or authority cannot be averaged away by excellence elsewhere.

EIGHT NECESSARY ATTRIBUTES / 八项必要属性

High-quality judgment must survive eight kinds of scrutiny

01DQA-01

Goal alignment

Align with the enterprise's long-run, systemic, risk-adjusted end state and make trade-offs explicit.

FAILURE RISK
Signing, scale, profit, process or a short-term KPI substitutes for the real objective.
PUBLIC EVIDENCE
Whether objectives, constraints, the inaction baseline and long-run outcomes are preserved together.
RELATED PUBLIC EVIDENCE
02DQA-02

Problem framing

Preserve the original question while identifying the source problem and competing frames.

FAILURE RISK
Producing more precise and longer analysis under a false premise.
PUBLIC EVIDENCE
Whether reframing changes alternatives, evidence or stop conditions.
RELATED PUBLIC EVIDENCE
03DQA-03

Evidence integrity

Separate confirmed facts, unverified facts, assumptions, expert opinions and inference.

FAILURE RISK
Documents, experience, conjecture and conclusions are blended into ‘facts’.
PUBLIC EVIDENCE
Whether source, fact cut-off, unknowns, versions and evidence limits can be checked.
RELATED PUBLIC EVIDENCE
04DQA-04

Complete alternatives

Compare the recommended path, viable alternatives, inaction and major alternative uses of capital.

FAILURE RISK
Treating the proposed transaction as the only option.
PUBLIC EVIDENCE
Whether alternatives, opportunity cost and inaction appear in the same decision record.
RELATED PUBLIC EVIDENCE
05DQA-05

Correct weighting

Identify decisive variables and compress low-weight issues and noise.

FAILURE RISK
Mistaking long checklists, more data or averaged opinions for judgment.
PUBLIC EVIDENCE
Whether variable priority changes time, resources, alternatives or the conclusion.
RELATED PUBLIC EVIDENCE
06DQA-06

Mechanism penetration

Explain how value arises, who bears risk, when the thesis fails and the strongest counter-explanation.

FAILURE RISK
Relying on revenue, profit, multiples or a single-period cash view.
PUBLIC EVIDENCE
Whether causal chains, dissent, failure conditions and later outcomes can be compared.
RELATED PUBLIC EVIDENCE
07DQA-07

Timely action

Judge decision sufficiency under incomplete information and define action, waiting, switching and stopping.

FAILURE RISK
Waiting indefinitely for perfect information or acting hastily under time pressure.
PUBLIC EVIDENCE
Whether critical gaps, waiting costs, stage gates and reopening conditions are recorded ex ante.
RELATED PUBLIC EVIDENCE
08DQA-08

Auditable accountability

System advice, formal decisions, authority, dissent, actions and ex-ante versions remain traceable.

FAILURE RISK
AI advice, team opinion or after-meeting consensus is presented as the formal decision.
PUBLIC EVIDENCE
Whether roles, authority, dissent, formal decisions and action ownership remain distinct.
RELATED PUBLIC EVIDENCE

TEN OBSERVABLE INCREMENTS / 十类可观察增量

Judgment improves only when the decision object changes

An increment must change at least one of the alternative, price, terms, capital use, authority, resources, sequence, time window, stop condition or inaction baseline. Longer text, more information or more risk items do not qualify on their own.

01DQI-01

Information

Find an omitted fact or critical gap.

NOT AN INCREMENT
Information grows without affecting any decision object.
DECISION EFFECT
Can change conclusion, price, terms, sequence or stop condition.
RELATED PUBLIC EVIDENCE
02DQI-02

Weighting

Reorder the importance of variables without new facts.

NOT AN INCREMENT
Merely extending a risk list or averaging opinions.
DECISION EFFECT
The new weighting changes resources, time or comparison of alternatives.
RELATED PUBLIC EVIDENCE
03DQI-03

Problem framing

Reframe the original question so the real decision object becomes visible.

NOT AN INCREMENT
Only restating the same question.
DECISION EFFECT
Changes object, alternatives, evidence, accountability, sequence or stopping.
RELATED PUBLIC EVIDENCE
04DQI-04

Goal calibration

Detect that an action has drifted from the north-star objective.

NOT AN INCREMENT
Mistaking a stage deliverable or gameable metric for terminal value.
DECISION EFFECT
Redefines the value that must actually be protected or created.
RELATED PUBLIC EVIDENCE
05DQI-05

Risk

Identify risk not yet carried by an actor, price or mechanism.

NOT AN INCREMENT
Simply stating that risk exists without changing the design.
DECISION EFFECT
Risk enters price, terms, stage gates, exit or stop rights.
RELATED PUBLIC EVIDENCE
06DQI-06

Mechanism

Move beneath surface numbers to a falsifiable causal chain and alternatives.

NOT AN INCREMENT
Describing the phenomenon without explaining how the result arises.
DECISION EFFECT
The mechanism can be supported, challenged or reopened by later facts.
RELATED PUBLIC EVIDENCE
07DQI-07

Cycle

Separate cycle performance from long-run capability.

NOT AN INCREMENT
Extrapolating peak-cycle outcomes into durable capability.
DECISION EFFECT
Changes scenarios, action conditions or capital cadence.
RELATED PUBLIC EVIDENCE
08DQI-08

Action

Turn judgment into a concrete change of path.

NOT AN INCREMENT
Analysis improves while action remains unchanged.
DECISION EFFECT
Changes design, price, terms, responsibility, resources or sequence.
RELATED PUBLIC EVIDENCE
09DQI-09

Time

Reduce unproductive delay or premature action.

NOT AN INCREMENT
Equating speed with quality or delay with prudence.
DECISION EFFECT
Waiting cost, lost windows or rework materially decline.
RELATED PUBLIC EVIDENCE
10DQI-10

Confidence

Strengthen or weaken confidence without necessarily changing the conclusion.

NOT AN INCREMENT
Treating support as no increment, or agreement as validation.
DECISION EFFECT
Adoptability changes after evidence, dissent and mechanism testing.
RELATED PUBLIC EVIDENCE

HOW TO USE / 如何使用

Compare judgments with one standard—not by report length

  1. 01

    Compare the team's original view, its general-AI-assisted view and the Howard-confirmed Capital OS view

  2. 02

    Review which material increment a decision process produced

  3. 03

    Test whether an error created a new gate or operating constraint

  4. 04

    Distinguish more content from an actual change in the decision object

WHAT THIS CANNOT PROVE

A standard alone does not prove that Capital OS is mature, and complete attributes do not guarantee success. Execution, environment, cycles and luck still affect outcomes. Weights, thresholds, scoring engines and full evaluation sets remain private.

SOURCE · VERSION · MACHINE

Source, version and machine interfaces

PUB-DECISION-QUALITY-0001 · V1.0Howard-confirmed current public standard; reviewed 2026-08-12.

decision-system.json →open-os.json →