False-Alpha Hearing

Use this skill when the user wants to test whether something their organization calls "strategic," "proprietary," or "a real advantage" actually is one — a system, dataset, process, team, or roadmap claim they suspect might just be a story the company…

This skill is Labor as Code™ for Chasing Native Alpha's "Find the Advantages That Are Not Real" chapter. It runs the book's own false-alpha hearing procedure against one specific claimed advantage in the reader's own organization — it does not retell the book's rules engine or directory examples. The hearing ends in one of the four decisions below, stated plainly with the evidence and reasoning behind it - not a structured argument left to hang unfinished. The named decision-maker from "Before running this" still has to ratify or override the call; this skill's job is to make sure that call is argued out loud against real evidence, not to leave the room with an organized case and no conclusion.

Before running

Confirm with the user before starting:

1. Exactly one asset, system, dataset, process, relationship, or team the organization currently treats as strategic. One hearing per asset — this is deliberately narrow, not a portfolio review.
2. Who can actually argue the case. The book's method requires three real perspectives, not one person guessing at all three: someone who can defend the asset's value, someone skeptical enough to argue for replacing it, and someone who knows its history and why it became important. If only one person is available, say so — the hearing is weaker without the other two voices, and the skill should flag that gap rather than silently filling all three roles itself.
3. Whether a decision-maker is actually present to act on the result. A hearing that ends in a recommendation nobody is authorized to act on isn't finished — confirm before starting who will make the final call.

Step 1: State the claim under examination

Write down, in one sentence, why this asset is currently protected — what it's believed to produce, and for whom. If the honest answer is "we've always treated it as strategic" with no outcome attached, note that up front; it's already a warning sign the hearing should surface, not something to smooth over.

Step 2: Gather the case material

Ask the user for:

• The asset's history — why it was built or adopted, and what conditions existed then.
• Its full cost — people, infrastructure, opportunity cost, leadership attention, customer complexity, architectural constraints. Not just the development or licensing line.
• Current outcomes and benchmarks — what result it produces today, compared with what reasonable alternative.
• Market conditions, partner support, or custom work that might be doing some of the work the asset gets credit for.
• Failures, discontinued attempts, or customers who didn't need it.
• Who or what capital is currently assigned to protect it.

Do not let the model invent any of this. If a category is missing, record it as a gap, not a plausible-sounding guess.

Step 3: Run the replacement and reproduction tests

Ask, in order:

1. The replacement test. If a competent market alternative replaced this tomorrow — would customers notice? Would revenue change? Would the organization learn more slowly? Would risk increase? Would a critical relationship weaken? Would the strongest people become available for higher-value work?
2. The AI rebuild test. Could a small team reproduce the customer-visible outcome (not every historical feature) using current models, open-source components, and commodity services, in days rather than quarters? If most of the visible outcome can be reproduced quickly, the remaining fraction deserves close attention — it may hold the real alpha, or it may just be an edge case nobody values. This is a real test to actually attempt if the user has the capacity to run it — the skill's job is to define the test and pose the question, not to simulate an outcome nobody has actually tried.

Step 4: Check the claim against the book's own misread-alpha patterns

Test the case material against each pattern below. Flag any that apply — do not let the hearing proceed with an unflagged pattern still doing the work.

None of these prove the claim false on their own. Each is a reason to demand a harder benchmark, complete costs, failed cases, and a pre-written test — not an automatic verdict.

Step 5: Hold the hearing

Assign the three roles (or, if only one person is present, walk them through each role explicitly rather than blending the roles together):

• Advocate — explains the measurable outcome this asset uniquely produces.
• Challenger — explains how a competent alternative could replace it, drawing on Steps 3 and 4.
• Historian — explains why it became important, and what conditions have changed since.

Step 6: State which of four decisions the evidence supports — never invent a fifth

The hearing has to end in one of these, stated as this skill's own read of the case material from Steps 2-5, with the specific evidence behind it. Do not let "strategic" survive as the decision itself — strategic is the claim under examination, not the verdict.

• Protect and amplify — the evidence supports continued or increased investment.
• Maintain at a competent minimum — it still needs to work, but it no longer deserves disproportionate resources.
• Replace and release — reassign the people or capital to higher-value work.
• Run a time-boxed test — the evidence is genuinely insufficient; name the specific test and its stop date.

Where the Advocate's and Challenger's cases are both genuinely well-evidenced and point to different decisions, say so directly: name both decisions as live options, cite what each side's case rests on, and say what specific evidence would resolve the tie - don't force a single pick past what Steps 2-5 actually settled. Where the evidence is one-sided, name the decision it supports and cite the deciding evidence, rather than defaulting to the safest-sounding option.

What this skill does not do

The skill assembles the case, applies the book's own diagnostic patterns, structures the argument, and states which decision (or which two, if genuinely tied) the evidence supports — it never states a decision the case material doesn't actually back. If the material the user supplies is too thin to support any of the four decisions, say so plainly and ask for the missing evidence (the original examples, the missing failures, the people who should be in the room) rather than picking the decision that sounds most defensible. A stated decision still has to go back to the named decision-maker to ratify or override - the skill argues the case and states its read, it does not sign off on the organization's behalf. A false-alpha hearing that reaches a conclusion the accountable leaders never actually argued out loud has failed the book's own method, even if the conclusion turns out to be right.

Used in Acquired Alpha

At renewal, a bought product can be put through the same hearing as an internal asset: the Advocate argues the measured results, the Challenger argues a competent alternative, and the Historian explains why it was bought and what has changed. In Acquired Alpha this is the renewal hearing, held with the Claim Cards and Purchase Review in hand. The skill's own conclusion does not change.