I read whitepapers backwards. Token allocation first, team section second, roadmap third, and the actual idea last, if I get that far at all. It sounds cynical until you accept what a first read is for. A first pass has one job, which is ruling the paper out. Understanding comes later, if the document earns it. Understanding is expensive and slow, and most papers that get sent to me do not deserve the effort. Ruling a paper out is cheap by comparison. Fifteen minutes, a search engine, and a short list of tells will catch a large share of the garbage before you have read a single sentence about the technology.
The reason this works is that a legitimate team and a scam team write whitepapers for different purposes. A real team writes to explain a system to people who might build on it, invest in it, or attack it. A scam team writes because the token launch checklist includes a whitepaper, so the document exists to be skimmed and to reassure. That difference in purpose leaves fingerprints all over the text, and the fingerprints are much easier to spot than the technology is to evaluate.
The four tells I check first
Plagiarism is the fastest check and the most damning one. Take two or three of the most technical sentences in the paper, wrap them in quotes, and search them. Lifted sections were rampant in the ICO era, and template mills still recycle entire architecture chapters today. Even one copied paragraph is disqualifying, because there is no innocent explanation for it. Teams that did the work describe the work in their own words, badly sometimes, but in their own words.
Second, revenue or yield claims with no mechanism behind them. Any time a paper promises staking rewards, guaranteed returns, or "sustainable yield", I ask the same question, which is who pays that yield and why they are happy to keep paying it. An exchange that charges trading fees has an answer. A paper where the yield comes from token emissions has just described paying you with dilution. A paper that never addresses the question at all has told you the answer is nobody.
Third, look at what the roadmap is made of. Engineering roadmaps are boring lists, testnet, audit, mainnet, a named integration with a specific protocol. Scam roadmaps read like marketing calendars, exchange listings, influencer partnerships, community growth targets, global expansion, a metaverse pivot in some quarter. When more than half of the milestones describe the token's visibility rather than the product's function, the roadmap is telling you what the team is actually building, and it is a price chart.
Fourth, find where the token allocation lives. Credible projects put distribution and vesting in the main body, with real numbers, because they are willing to defend them. Papers that push the allocation into an appendix, compress it into a single pie chart with no vesting table, or omit it entirely are hiding it from exactly the fifteen-minute read you are doing right now. Insider allocations north of roughly 40 percent, no lockups, and a large "ecosystem fund" controlled by the team are the usual finds once you dig it out.
A real paper next to a template one
The comparison that taught me the most is putting the Bitcoin whitepaper next to any template-generated launch paper. Satoshi's paper runs about nine pages. It states a problem, proposes a mechanism, walks through the obvious attacks, and includes the math for an attacker trying to catch up with the honest chain. There is no roadmap, no team section, and no token allocation, partly because there was nothing pre-allocated to disclose. It barely uses adjectives, and every section exists because a skeptical reader would have asked the question it answers.
The template paper is usually forty-odd pages, and half of them are stock diagrams. The problem statement is the market size of some industry rather than a technical problem. The architecture section names components without ever explaining how they interact, hides behind phrases like "proprietary AI-driven consensus", and leans hard on revolutionary, ecosystem, and next-generation. The giveaway is that you could swap the logo and the token name and the paper would describe a thousand other projects equally well. A genuine paper reads like it could only have been written by the people who built that specific system. A generated one reads like it was written to be a PDF that exists.
The fifteen-minute scoring sheet
Here is the pass I actually run, in order. Score each item 0 for clean, 1 for questionable, 2 for a clear fail.
- Plagiarism. Search three verbatim technical sentences. Any match against another project's paper scores 2.
- Yield mechanism. Every return number has an identified payer. Vague "protocol revenue" with no fee source scores 1. No source at all scores 2.
- Roadmap composition. Count marketing milestones against engineering milestones. A marketing majority scores 1. A roadmap that is entirely listings and partnerships scores 2.
- Token allocation. Distribution plus vesting in the main body scores 0. Buried in an appendix scores 1. Missing, or insiders above roughly 40 percent with no lockup, scores 2.
- Team. Named people with verifiable history score 0. First names and avatars score 1. No team information anywhere scores 2.
- Specificity. Read one mechanism paragraph and ask whether it would still be true with a different project's name pasted in. If yes, score 2.
My cutoff is a total of 4 or more, and at that point I close the tab no matter how good the narrative sounds, because narrative is the one thing scams reliably do well. A 2 or 3 means proceed carefully and verify the weak items outside the paper. A 0 or 1 means the paper has earned an actual read, which is where the real work starts rather than ends.
What this filter misses
I want to be honest about the limits. A clean whitepaper does not make a good investment. Plenty of sincere teams with careful papers built products nobody wanted, and the token drifted to zero for ordinary business reasons. The well-funded frauds also hire good writers, so a polished paper from a serious operation will pass a text-only screen. The scoring sheet catches the lazy majority of scams, which is worth a lot, but it does very little against the diligent minority.
The fix for that second group is checking the paper against the chain. The allocation table claims a vesting schedule, so look at whether the insider wallets are actually locked, and whether large holders are quietly distributing while the roadmap talks about growth. The gap between what the paper says and what the wallets do is where the better frauds show up, and it is one of the reasons we built whale and insider wallet tracking into Blockcircle. Paper claims and chain state disagree more often than you would hope.
None of this requires expertise you do not already have. It takes a search engine, a timer, and the willingness to close the tab when the score says to close it. Most papers I run through this fail inside ten minutes, and the small number that pass tend to be the ones worth a slow second read.