The N/A Report: What a Nine-Dimension Diligence Template Tells You When Every Field Is Empty

Hook

Over the past seven days I have read forty-one research documents. This is bear-market arithmetic — deal flow dries up, memos multiply. Forty of them were indistinguishable: a thesis, a chart, a valuation range, a conviction sentence. The forty-first had no thesis, no chart, no range, and no conviction sentence. It had nine dimensions. Technical architecture. Token economics. Market structure. Ecosystem position. Regulatory exposure. Team and governance. Risk matrix. Narrative and expectations. Supply-chain transmission. Every heading was present. Every table had borders. Every cell was formatted correctly.

And every single field read: N/A.

Not "unknown." Not "requires further diligence." N/A — not applicable. Structurally unfillable. The analyst had constructed a complete instrument and the instrument had returned nothing. The only section of the entire document that contained actual prose was the section explaining why the rest of the document contained no prose at all.

I have spent twenty-one years watching this industry, and the last nine auditing code inside it. In late 2017 I spent three weeks on an ERC-20 distribution function and found the integer overflow before DragonCoin launched. I have watched contracts fail loudly, at volume, on-chain, with a block explorer as witness. What I had not watched, until this week, was a research framework fail silently — and remain standing afterward.

That is the artifact worth dissecting. Not the missing project. The surviving template.

Context: The Economies That Produce Empty Templates

To understand why an empty diligence report is interesting, you have to understand what the diligence report has become. It is no longer primarily an analytical instrument. It is a product. It has a market, a pricing curve, and a customer acquisition funnel, and none of those three things reward the output "N/A."

Here is the mechanism, and I want to map it the way I map a liquidity pool — because the incentive geometry is the same.

Between 2017 and 2019, crypto research was mostly retail-facing. Twitter threads, Medium posts, occasional PDFs. The audience was people who wanted to know whether to buy. The output that served them was directional: bullish, bearish, or genuinely uncertain — and genuine uncertainty was rare because genuine uncertainty does not go viral.

Between 2020 and 2021, the audience bifurcated. DeFi summer created a cohort of users who needed mechanics, not opinions — impermanent loss math, emission schedules, gas optimization. I was in that cohort. I wrote a Python script in mid-2020 that polled Uniswap and SushiSwap pool states and fired arbitrage trades when the spread exceeded my modeled cost of execution, including gas, slippage, and the probability of a reverted transaction. Five hundred and eleven automated trades. Forty-five thousand dollars. The lesson was not that I was clever. The lesson was that the market was pricing narrative and the pools were pricing spreads, and those two things diverged often enough to be harvested. The script did not read Twitter. The script read reserves.

Between 2022 and 2024, the audience bifurcated again — or rather, it stratified. The collapse of TerraUSD in May 2022 did two things simultaneously. It vaporized a few hundred billion dollars of nominal value, and it created durable demand for documents that looked like they could have prevented it. That demand was institutional, and institutions buy process. They do not buy vibes. They buy templates, frameworks, scoring rubrics, and escalation matrices, because a template is auditable and a hunch is not.

So the template industry was born. Not the research industry — that already existed. The template industry. The business of selling the shape of diligence separately from the substance of diligence.

The shape is genuinely valuable. I mean that. In early 2024 I spent three months inside the prospectuses of the spot Bitcoin ETF applicants, and the substance that mattered was almost entirely structural — custody arrangements, creation and redemption mechanics, the legal question of whether an in-kind redemption was even permissible, the cash-only constraint that most retail readers skipped past because it is boring and because it does not produce a price target. I estimated that structural differences alone would account for roughly two billion dollars in differentiated initial inflows. That estimate came from reading filing documents line by line, cross-referencing custody language against prior custody language, and tracking which clauses changed between amendments.

That work had a shape and a substance. The problem is that the shape is reproducible by anyone with a template and an afternoon, and the substance is not reproducible by anyone without primary sources and a willingness to read them.

When the shape decouples from the substance, you get exactly the artifact I received this week. Nine dimensions of scaffolding. Zero dimensions of load-bearing material. And critically — this is the part that matters — the scaffolding did not collapse. It was submitted. It was formatted. It had section numbers. It is, in every visual and structural sense, a diligence report.

Core: The Nine Gates and What Their Emptiness Discloses

A useful diligence framework is not a questionnaire. It is a sequence of verification gates. Each gate has a pass condition, a fail condition, and — this is the one nobody designs for — a cannot-evaluate condition. Most frameworks treat cannot-evaluate as a temporary state, a placeholder to be filled in later. That is a design error with the same flavor as a smart contract that treats an unset variable as zero.

Let me take the nine dimensions in the order the document presented them, and treat each empty field as evidence rather than absence. Because an empty field is not nothing. An empty field is a reading.

Gate One: Technical

What the technical field needs is a repository. Not a whitepaper — a repository. Commit history with timestamps that show continuous work rather than a single dump. An audit report with a named firm, a named date, and a named revision hash, so I can verify that the audited code is the deployed code. A trust-boundary diagram: who can pause, who can upgrade, who can mint, who can drain. Answers to the three questions I ask of every contract, which are the same three questions I asked in 2017 when I sat with DragonCoin's distribution logic for three weeks.

Who has privileged access? What is the failure mode if that access is used maliciously? Can the failure mode be triggered from outside the trust boundary?

On DragonCoin, the answer to the third question was yes, and the answer to the first was "the distribution function itself." An integer overflow in the minting arithmetic meant that a miner — not the team, not an investor, a miner — could have pushed the supply counter past its maximum and begun minting into an unbounded state. The fix was nine lines. Finding it took weeks, because the vulnerability was not in the logic anyone was proud of. It was in the arithmetic nobody had re-read.

When the technical field reads N/A, the reading is not "no vulnerabilities found." The reading is "no trust boundary has been defined, therefore no trust boundary has been tested." That is a strictly worse condition than "audited and clean" and a strictly worse condition than "audited and found flawed." It is the condition of a system whose assurances have not been specified, which means they cannot be relied upon and cannot be refuted. Unfalsifiable assurances are not weak assurances. They are not assurances at all.

Gate Two: Token Economics

What this field needs: total supply, current circulating supply, emission schedule with cliff and linear components separated, allocation table by category with named counterparties where possible, and — the field almost everyone omits — a value-capture mechanism that does not depend on new buyers entering.

The last one is the whole ballgame. I wrote about this in 2020 while arbitraging emission tokens, and the observation has only hardened since. Yield is not a property of an asset. Yield is a property of a subsidy schedule. When you farm an emission token, you are not earning a return on capital; you are receiving a transfer from the treasury to you, denominated in a unit whose price is set by the marginal buyer of that transfer. The transaction is real. The yield is real. The sustainability is not, because the mechanism requires a continuously increasing buyer base to hold the price of the thing being distributed.

If the tokenomics field is empty, there is no emission schedule to model, which means there is no dilution curve, which means there is no way to distinguish a protocol with genuine fee revenue from a protocol with a well-timed distribution. Both look identical on a price chart for the first two quarters. They diverge in the third.

An empty tokenomics field means the divergence cannot be forecast. It also means — and this is the subtler point — that the framework has no way to detect the specific structure I spent 2020 learning to detect. Without supply data, you cannot compute the ratio of real revenue to emitted value. Without that ratio, you cannot tell whether the protocol is a business or a faucet. Faucets are fine. Faucets that are marketed as businesses are not.

Gate Three: Market

What this field needs: cycle position, funding rates, order-book depth across venues, and the correlation structure between this asset and the rest of the book.

Market data is the one dimension that is almost never genuinely absent, which makes its N/A designation informative. Price exists. Volume exists. Funding exists. If the market field is empty, it is not because the data does not exist — it is because the analyst could not attribute the data to a specific instrument. Which means: no tradable asset has been identified. Which means: the thing under analysis may be a protocol, a proposal, a governance process, or a concept, but it is not yet a position.

I want to be precise about why that matters in a bear market, because this is where I diverge from the standard research voice. In an expanding market, almost everything works. Liquidity is abundant, correlation is high, and the difference between a good protocol and a lucky one is obscured by the tide. In a contracting market, the differences surface violently. Depth evaporates first, before price. Funding rates flip before price. The order book thins before price. Anyone who has executed size in a deteriorating market knows the sequence: the quotes widen, then the size at the quote shrinks, then the quote disappears and you are left staring at a spread that no longer represents anything you can trade against.

Liquidity dries up before the hype does. That is not a slogan; it is a sequencing fact. And an empty market field means the sequencing cannot be observed, because there is no instrument on which to observe it.

Gate Four: Ecosystem

What this field needs: a map. Upstream dependencies — what does this protocol require to function, and who controls those inputs? Downstream integrators — who consumes this protocol's output, and how easily can they switch? And the developer signal: contributor count over time, contract deployment frequency, the ratio of code merged to code proposed.

I care disproportionately about the upstream column, because upstream dependencies are where systemic risk hides. Every protocol that depends on an oracle depends on the oracle's honesty assumptions. Every protocol that depends on a bridge depends on the bridge's validator set. Every protocol that depends on a stablecoin depends on that stablecoin's redemption mechanism. These dependencies form a directed graph, and the graph has degrees of freedom and failure modes that no single protocol's audit will ever cover.

In May 2022 I watched this graph fail in real time. I was on Etherscan before the outlets picked it up, tracking the minting pattern of TerraUSD against the burn mechanics of LUNA. The correlation was visible hours before it was consensus — not because I had proprietary data, but because the on-chain mint-and-burn relationship was arithmetically legible to anyone who sat down and computed it instead of watching it. The mechanism was: burn one dollar of LUNA, mint one dollar of UST. As long as UST is trusted to be worth one dollar, the arbitrage holds the peg. The moment the peg slips below the cost of the burn, the incentive inverts, and the minting becomes self-reinforcing in the wrong direction. The supply of LUNA expands to absorb the redemption pressure, which dilutes LUNA, which increases the required minting, which dilutes further.

That is a closed loop with a sign error. It is not a sentiment failure. Sentiment was the messenger. The mechanism was the message.

An empty ecosystem field means the map does not exist, which means the loop cannot be traced, which means you cannot tell whether you are looking at a system with a stabilizing feedback or a destabilizing one. Both look calm until they do not.

Gate Five: Regulatory

What this field needs: jurisdiction of incorporation, jurisdiction of the team, jurisdiction of the user base, and a legal analysis of whether the asset triggers securities classification under the relevant test.

The document I received had a Howey table. Four rows — investment of money, common enterprise, expectation of profit, efforts of others — and four N/A cells. I find this instructive, because the Howey test is not a test you run on a project. It is a test you run on a transaction, and often on a particular marketing statement about a transaction. The test's fourth prong — reliance on the efforts of others — is famously the one that turns on the specific promises made to specific purchasers at specific times. A token sold with a roadmap and a team allocation is in a different legal posture than the same token sold later on a secondary venue with no promises attached.

This is exactly the kind of nuance I spent three months on in early 2024, and it is exactly the kind of nuance a template cannot capture, because a template asks "does the asset pass Howey" and the honest answer is "Howey is a sentence that has to be completed with facts."

An empty regulatory field means the sentences are uncompletable. Which means the exposure is unmeasured. Which in a bear market is the worst of the available states, because enforcement priorities shift during contractions — regulators, like everyone else, respond to incentives, and the incentive to pursue visible enforcement is highest when the market is loudest about its losses. Understanding the offer's structure is the work — marketing language, contribution mechanics, and the sales narrative.

Gate Six: Team and Governance

What this field needs: named humans with verifiable history, a governance model that specifies where authority actually resides — multisig with known signers, or token vote with known distribution, or something else — and an investor table with round sizes, valuations, and vesting.

The governance question is the one most often answered incorrectly, because the answer is almost never "on-chain." It is "a multisig controlled by four people, one of whom is pseudonymous, holding upgrade authority over the contract that custodies user funds." That is a centralized system with a decentralized interface. It is not a criticism — many such systems work fine — but it is a description, and descriptions are what a diligence report is supposed to contain.

Top-ten holder concentration is the proxy that matters most here, and it is the one the template asked for and did not receive. If you do not know the concentration, you do not know the governance. If you do not know the governance, you do not know who can change the rules while you are holding the token. I have seen this play out enough times to have a default posture: assume the multisig can and will do whatever the multisig is permitted to do, then ask whether you are comfortable with that set of permissions. If the permissions are unknown, the question has no answer, and a system whose rules cannot be stated cannot be underwritten.

An empty team field does not mean anonymous. It means unexamined — the difference between a pseudonymous developer with a nine-year commit history and a pseudonymous developer with no history at all is enormous, and both present as blank cells.

Gate Seven: Risk

The risk matrix is where a template's limitations become definitional. A risk matrix is a two-dimensional object: likelihood and impact. Both axes require a reference class. Likelihood of what, relative to what distribution of similar events? Impact measured how, in dollars or in probability of total loss?

Without a reference class, a risk matrix is a mood board. Every cell gets a color, the colors are generated by feel, and the output looks rigorous because it is gridded.

What an empty risk matrix tells me is subtler than it appears. It tells me the analyst could not populate the columns, which means the analyst could not name the threat model, which means the threat model has not been constructed. And a threat model is not a list of bad things that could happen. A threat model is a statement of the form: given this trust boundary, this asset's value, and this adversary's capability, what is the cheapest path to extracting that value?

The cheapest path is almost never the exotic one. It is the unpatched oracle, the single-key admin, the bridge validator set with a quorum of three, the frontend that loads a script from a CDN. If you have not written the cheapest path down, you do not have a risk matrix. You have an anxiety summary.

Gate Eight: Narrative

This dimension the template handled better than the others, though still emptily. It asked for the current narrative, the narrative's lifecycle position, the fundamental support level, and — the good part — an expectation-gap table, comparing what the market expects against what has actually been delivered across user growth, revenue, and technical milestones.

That is the right structure. Narrative is a real object with measurable properties. It has a birth, a growth phase, a saturation point, and a decay. It has a relationship to fundamentals that changes over the life of the narrative: early on, the narrative leads and the fundamentals follow or fail to; late in the lifecycle, the fundamentals are fully priced and the narrative can only decay.

The expectation-gap table is where money is made and lost. Terra is the canonical example. The market expected an algorithmic stablecoin to hold a peg that its mechanism could only hold under a specific and narrow set of conditions. The mechanism was public. The expectation was public. The gap between them was public. Nobody priced the gap because pricing the gap required admitting the peg was conditional, and admitting the peg was conditional was, for a period, socially expensive.

An empty narrative field means the lifecycle position is unknown, which means you cannot distinguish a narrative in its second week from a narrative in its fortieth. Those two positions have completely different risk profiles and completely different expected returns, and they look identical in a promotional document. The reason they will produce similar journalistic summaries is that the underlying claim cannot be tested — the narrative has no referent, and a narrative with no referent is only a signal about the people repeating it.

Gate Nine: Supply-Chain Transmission

What this field needs is a directed graph of who benefits and who loses when this thing works or fails. Miners, exchanges, infrastructure providers, DeFi protocols, NFTs, traditional finance. The template listed these categories and left them empty.

Transmission analysis is the one dimension I would keep if forced to discard the other eight, because it is the only dimension that captures second-order effects, and second-order effects are where the outsized returns live. When a change happens at the base layer, the biggest moves are rarely in the thing that changed. They are in the things that depend on it, priced by people who did not notice the dependency.

This is also the dimension that explains why so much of what passes for innovation in this industry is really re-labeling. A product that is aimed at a new chain but which inherits the assumptions, the liquidity, and the user base of an existing chain is not a new market. It is a new storefront in front of the same customers. The transmission graph makes this visible: if you draw the arrows and every arrow points back to the same concentrated pool of capital, the product is not creating a market. It is re-partitioning one.

An empty transmission field means the graph is unbuilt, which means the second-order effects are unmodeled, which means the thing could succeed spectacularly and still not move the assets you are holding. That is a real and underanalyzed failure mode: correct thesis, wrong instrument.

The Meta-Finding: The Template Survived

Nine gates, nine N/A cells. The instinct is to call this a failure of analysis. I want to argue the opposite. The report is the most honest document I received all week, and the reason is structural.

Every other report I read filled its gates. Not because the information existed, but because the format demanded content. A template with a blank cell creates pressure to fill the blank cell, and the pressure does not know the difference between data and inference. So the cells get filled with something — a comparable, a projection, a "based on our conversations with the team" — and the filling converts a rigorous instrument into a persuasive one. The report stops describing the world and starts describing the analyst's need to appear productive.

I know this pressure intimately because I have felt it in the other direction. When I audited DragonCoin, I could have written a clean report. The code compiled. The tests passed. The token distribution had a plausible-looking schedule. Nothing in the presentation of the code signaled a problem. The problem lived in the arithmetic, and finding it required me to sit with a function I had already decided was safe and re-derive it from first principles. The moment I decided the function was safe, I stopped being able to see the bug. The only cure was to un-decide.

An N/A report is a report where the analyst refused to decide. That is rare and it is valuable, and it is almost never published, because the market for research pays for conclusions and N/A is not a conclusion.

So here is the reframe, and it is the core insight of this piece: the empty diligence report is not an incomplete document. It is a completed document whose conclusion is that the object of analysis has no verifiable surface. That is a finding. It has an action attached to it — the action is to route the opportunity elsewhere and to spend the diligence budget on something with an attack surface.

Nine gates returning N/A is not nine failures. It is one success: the framework did its job, which was to refuse to manufacture certainty.

Contrarian: The Industry Is Over-Researching and Under-Looking

Here is the counter-intuitive claim I want to leave with you, and it runs against everything the last four years of institutionalization has taught us to believe.

The industry does not have a research deficit. It has a research surplus, and the surplus is actively harmful, because it produces the feeling of diligence without the substance of verification, and the feeling is what gets priced.

Consider the mechanics of information gain. A research document adds value if it contains at least one claim the reader did not have and could not easily get. That is the whole test. Everything else — the formatting, the length, the confidence, the font — is decoration. Run any given research report through that test and the failure rate is spectacular. The reader learns the protocol exists, learns what it says about itself, and learns the analyst's opinion about what the protocol says about itself. No primary source has been consulted. No repository has been opened. No contract has been read. The report is a mirror held up to the project's own marketing, with a byline attached.

I did the counter-example work in 2024 and it was mostly unglamorous. The information gain came from custody language, not from price targets. It came from noticing that a particular amendment changed a particular clause in a way that had downstream implications for how redemptions would clear. That insight was worth more than every opinion in every note I read that quarter, and it was obtained by comparing two documents at the sentence level.

The surplus has a second harmful effect, which is that it manufactures collective confidence about things nobody has checked. When forty analysts publish forty reports about the same eight protocols, the industry's attention is allocated by the volume of documents rather than the quality of evidence. The protocols that get researched are the protocols that are already known. The protocols that need scrutiny are the ones too small, too weird, or too early to have attracted a byline.

Which brings me to the market structure the surplus is built on top of — and here my contrarian reading gets specific.

There is a widely repeated claim that this industry's central problem is liquidity fragmentation, and that the solution is more infrastructure to route liquidity across venues. I have watched this claim circulate for six years, and I have watched it be singularly productive — not at solving the problem, but at generating products. Each new protocol that promises to unify liquidity creates a new pool that must itself be routed toward. The fragmentation is not being solved. It is being subdivided, and the subdivision is being sold as the solution to the previous subdivision.

The same geometry applies to the scaling layer. There is a fixed population of capital and users in this market, and it is contracting, not expanding. Adding more execution environments to a contracting pool does not increase the pool. It slices it. Each new environment requires its own liquidity, its own security budget, its own bridge surface, and its own research coverage. The research requirement grows faster than the user base, which is precisely why the research surplus exists: the infrastructure is generating diligence obligations faster than it is generating adoption.

And the labeling gets worse the further you get from the base layer. There is a category of assets claiming to extend Bitcoin that are, on inspection, execution environments built with the tooling and the assumptions of a different chain family entirely, with a Bitcoin-shaped brand applied at the front. The brand is doing the analytical work that the architecture should be doing. The Bitcoin community has noticed. The template has not, because the template has a field labeled "upstream dependencies" and the field got an empty cell, and the empty cell was formatted correctly, and nobody read it.

That is the shape of the trap. The format is professional. The fill is opportunistic. The reader is reassured by the first and cannot detect the second.

The defense is not a better template. Templates are shape, and shape is reproducible. The defense is a behavioral commitment: no claim without a primary source; no score without a stated rubric; and — the hardest one — a willingness to publish N/A and let it stand.

I will be blunt about how rare that commitment is. The incentive structure pays for ranked lists, for conviction calls, for the sentence that says a thing is worth attention. It does not pay for the sentence that says the thing cannot be evaluated, even when that sentence is the truth. The analysts who write the second sentence do not get invited back. The ones who write the first sentence get quoted. This is not a failure of individual integrity. It is a failure of market design, and market design failures do not fix themselves.

Takeaway: Null Results Are Signals

Here is where I place my own bet, and it is the same bet I have been making since I put a testnet wallet in front of an autonomous agent in 2026 and let it negotiate a data-access fee against a contract with no human in the loop.

The agent did not care about the narrative. It did not care about the formatting. It read the contract, priced the call, signed the transaction, and moved on. Ten thousand dollars in testnet funds, a set of rules, and an outcome. What emerged from that experiment was not a product. It was a preview of how machine readers behave: they do not reward shape, they reward legibility. A field labeled N/A is legible. A field filled with a confident inference and no source is noise, and noise is the thing agents are built to filter.

So the direction of travel is clear to me. As machine-mediated evaluation grows, the value of structural honesty rises and the value of persuasive formatting falls. The N/A report that a human committee would discard is exactly the document that a machine reader would route on, because it tells the downstream process where the missing information lives — which gate failed, which field is empty, which primary source would need to be consulted to close the gap.

That is the difference between a null result and an absence. A null result is a measurement that says: I looked, with this instrument, in this place, and found nothing. An absence is a blank. The document I received is a null result. It knows what it looked for and it knows it did not find it, and it recorded both.

Which raises the question I cannot answer and will not pretend to. If the industry's research output is measured by volume, then null results are worthless and the surplus will keep growing. If it is measured by information gain, then the surplus is a liability and the null results are the most valuable thing any of these desks produced this quarter.

Every desk I know is currently measuring the first thing and claiming the second. The gap between those two metrics is the same gap as the one inside the report: a template that looks complete, sitting on top of an object that has no verifiable surface.

The difference is that the report admitted it.

That is the only reason I read it twice.