Tuga sniffs her food before she eats it. Every single time, including food she has eaten six hundred times, from the same tin, in the same bowl, put down by the same person.

It looks like neurosis and it isn't. It's a cheap check that runs before an expensive commitment, and it costs her two seconds. The two seconds don't make her right — she still eats things she shouldn't find outdoors — but they catch the obvious problems, and the obvious problems are most of them.

Here is the crypto equivalent. Twenty minutes, nine questions, applicable to anything anyone sends you.

1. What Problem Does This Solve, in One Sentence?

If you can't state it plainly after reading the front page, that's the answer. Projects that do something specific can say what it is. Projects that can't will give you a paragraph containing "ecosystem", "infrastructure layer" and "empowering".

Red flag: the description only makes sense to people already inside it.

2. Who Uses It Who Isn't Being Paid To?

The single most informative question on this list, and the one most often dodged.

Rewards attract users. The honest version of the metric is: what remains three months after the incentives stop? If every number on their dashboard is from an incentive programme, you're looking at a payroll, not a product.

3. Would This Exist Without a Token?

If the token is a fee discount, a governance instrument, or genuinely necessary to the mechanism — fine. If you remove it and the product still makes sense, ask why it's there. If you remove it and the product makes more sense, you have your answer.

Red flag: the token exists to be sold to you.

4. What Does the Supply Actually Look Like?

Not the price. The structure:

• How many units exist, and how many are circulating?
• What's the unlock schedule? Who receives tokens over the next two years, at what rate, and at what cost basis?
• What share went to insiders and investors versus to the public?

This is published, findable, and routinely ignored. A low circulating supply with a cliff of insider unlocks ahead of it explains an enormous share of charts that mysteriously go down.

5. Who Is Selling Into Your Buying?

Follow directly from the above. Someone with tokens at a fraction of your price, unlocking next quarter, is not your partner. They are the other side of your trade, and they were there first.

6. Who Built It and Who Funded It?

Names, histories, previous projects — including the ones that failed. Anonymous teams aren't automatically disqualifying (Bitcoin's creator is anonymous), but anonymity plus a token sale plus marketing spend is a specific combination with a specific history.

Check what the funders bought at, and when they can leave.

7. What Would Make This Fail?

Ask it of the project, then look for whether they ever discuss it. A team that publishes its own risks is operating in reality. A communication channel with no bad news in it has an editorial policy, not a product.

If you can't name a failure mode after twenty minutes of reading, you haven't understood it yet — you've absorbed the marketing.

8. How Old Is the Story?

By the time something has a name, a dedicated index, and daily coverage, the positioning is largely complete. You are not early to a thing you discovered in a feed. This is the laser-pointer article in checklist form.

9. Who Benefits From Me Believing This?

The last question, applied to whoever put it in front of you. Affiliate link, referral code, existing bag, paid placement, platform where engagement is the revenue. None of those make someone a liar. All of them shape what gets said.

Apply it to me too. I hold bitcoin and I'd rather you held some.

The Scoring

Not a model, just a rule I'd actually use:

• Two red flags: walk away. There will be another one next week. There is always another one next week.
• No red flags: you've eliminated the obvious frauds, and nothing more. This filter removes bad things; it does not identify good ones. Most projects that pass it will still fail, because most things fail.
• Size it like a bet regardless. Passing a checklist is not conviction, and conviction is not a sizing input anyway.

And the honest overall position: for most people, most of the time, the correct output of this exercise is no. That's not cynicism — it's the base rate. The twenty minutes aren't there to find you a winner. They're there to stop you funding something you'd be embarrassed to explain.

The Point

The sniff doesn't tell her the food is good. It tells her whether anything is obviously wrong with it, in two seconds, before she commits her face to the bowl. She still gets it wrong sometimes. She gets it wrong much less than an animal that didn't bother.

Twenty minutes, nine questions. You'll still back things that don't work. You'll stop backing the ones that were never going to.

Sniff first. It costs you nothing and it catches the obvious. 🐾⚡