I Asked an AI for My Fundamental Watch Archetypes. Then I Graded Its Answer.

I Asked an AI for My Fundamental Watch Archetypes. Then I Graded Its Answer.

The final question from my Instagram question box series. @thatdavidmerlin asked: "You know this one: what are your fundamental watch archetypes?" He's right, I do know this one. But instead of navel-gazing, I ran an experiment.

I fed my entire collection to an AI and asked it to tell me what my fundamental archetypes are. Not brand, not complication, not price tier. The actual roles and emotions underneath the buying.

This is a trick I've used before. When I culled from 140 watches to 40, an AI analysis of my own collection helped me see patterns I was too close to see. The machine has no nostalgia and no ego. It just reads the data.

It came back with six archetypes. Some of them nailed things I'd never articulated. One of them was my own thesis sold back to me under a new name. It cited a watch I no longer own. And it completely missed a seventh archetype that I'd argue is the most durable category in my entire collection.

Here are the six, with my verdicts.

1. The Quiet Masterpiece

Watches that reveal themselves slowly instead of demanding attention. Exceptional finishing, restrained design, little or no logo flex, appreciated most by enthusiasts. The AI's examples from my box: the A. Lange & Söhne Lange 1 Moon Phase, the Laurent Ferrier Sport Auto, the H. Moser & Cie Mayu Marrone, the Grand Seiko SLGW007 Moonlit Birch. Its summary: this archetype says "I know exactly what I'm looking at, and that's enough."

Verdict: right, and also, I've heard this one before. This is understated luxury. It's on the masthead. It's the lane I've been writing about since this blog started. The AI analyzed a few dozen watches and confidently rediscovered my own collecting thesis, then handed it back to me with a better-sounding name. I'll admit "Quiet Masterpiece" is a good name. But the machine didn't find this pattern. I built the collection on it. The pattern found the machine.

It also called this my favorite archetype. Close. It's my biggest archetype. Favorite and biggest aren't always the same thing, and we'll get to that.

2. The Collector's Collector

Watches that make another collector stop and smile. Not mainstream grails, but pieces people discover after years in the hobby. They communicate curiosity more than wealth. Examples: the Cartier Tortue Monopoussoir CPCP, the Credor Pacifique, the Pequignet Concord Titanium, and from my wishlist, the Daniel Roth Chronograph S247.

Verdict: right, and this one I hadn't named. There is a specific pleasure when someone who's been in the hobby for fifteen years clocks the CPCP across a table. It's a different transaction than a stranger noticing a watch. It's recognition between people who took the same long road. I knew I felt it. I'd never filed it as its own category of buying. Point to the machine.

3. The Pure Design Object

Watches purchased because every proportion feels correct. The AI's examples: the Nomos Zürich Worldtimer, the CPCP again, the Parmigiani Fleurier Tonda PF Minute Rattrapante, the Laurent Ferrier Sport Auto again. Then it made its sharpest observation: I almost never buy a watch because it's complicated. I buy it because the design is elegant.

Verdict: correct, full stop. I never buy a watch for its complications. If a watch I love happens to have complications, great. But the rattrapante didn't come home because it's a split-seconds chronograph. It came home because it's one of the most beautifully resolved objects I've ever held, and the split-seconds is along for the ride. Complications are seasoning. Design is the meal.

Notice, also, that the CPCP and the Sport Auto each show up in two archetypes. That's not sloppy analysis. That's the tell. These aren't bins, they're lenses, and the untouchable watches in my collection are the ones that score through several lenses at once.

4. The Honest Tool Watch

Icons in their purest form. The Rolex Submariner 5513. The Tudor Submariner 7016. The Rolex Daytona 126500LN. The Omega Speedmaster Tokyo 2020. The AI pointed out what's missing: huge dive watches, extreme complications, modern hype pieces.

Verdict: right. I love tool watches, but only when they're honest. A tool watch that has been inflated to 45mm and draped in marketing about extreme professionals is no longer a tool. It's a costume. The 5513 is a tool watch the way a good hammer is a hammer.

5. The Story Watch

The most personal category, per the machine. Watches purchased not because they're objectively best but because they represent something. The Tokyo Speedmaster. The Tudor Heritage Chrono from the same trip. The vintage Rolexes.

Verdict: right, but out of date, and here the machine embarrassed itself. Its list of my story watches included a Glashütte Original PanoMaticLunar, which it described as a chapter of my life. I sold both of my PanomaticLunars this year to fund the Lange 1. The AI wrote a moving tribute to a watch that had already left the building. Which is a useful reminder about these tools: they read the records, not the life. The records lag.

6. The Beautiful Misfit

Watches that deserved more attention than the market gave them. The Pequignet Concord Titanium. The Mayu Marrone. The Citizen AQ4100-22L. The Credors. The AI called this the most "Midlife Crisis Watches" archetype of all, the pieces that punch above their weight, the ones I most like writing about.

Verdict: guilty. These are the watches I evangelize hardest, precisely because nobody else will. A Beautiful Misfit is a Collector's Collector that hasn't been collected yet. Championing them is half the fun of writing this blog.

What the machine missed

This is where the grading gets interesting, because the machine didn't make one kind of mistake. It made three, and they escalate.

Miss one: it skipped the best examples of its own categories. My Laurent Ferrier Classic Moon Silver appears nowhere in its analysis. That's the watch I've called my goosebumps piece, the one I named in my three-for-life post as the "special" watch. It is arguably the purest Quiet Masterpiece I own, and the machine left it off the list of its own favorite archetype. Same story with my neovintage Rolex GMT-Master II 16713, the Rootbeer, which slots into Honest Tool Watch without a fight. If a human analyst presented this framework and whiffed on the single best supporting example in the room, you'd question the whole deck.

Miss two: it deleted my outliers. Nowhere in the six archetypes is there a home for the Zenith Chronomaster Lapis Lazuli or the Behrens NaviGraph. These are the boldest watches I own, the two most likely to start a conversation, the ones I've admitted I stare at constantly. They don't fit the understated thesis, so the machine quietly left them out of the taxonomy entirely. Think about what that means. It didn't just read my collection. It found a pattern it liked and then discarded the evidence that complicated the pattern. The analysis flattered the thesis. If I graded my own portfolio work that way, I'd be fired.

Miss three: it cannot see the milestone watch. Six archetypes, and not one of them covers the most durable category I own.

I've written about this before in this series: I use watches as prizes for achieving things, and I have never culled a single one. Not the original Panerai Radiomir from the sale of my first company. Not one. In a collection where roughly a hundred watches have come and gone, the milestone watches are batting a thousand.

The AI couldn't see this category because it isn't visible in the specs. A Radiomir looks like a big tool watch to a machine. It can't see the company sale attached to it. The most important data about a collection is the data that never makes it into the spreadsheet.

So the real list is seven, plus an asterisk for the outliers. Six the machine found, one it couldn't, and two watches it pretended not to notice.

The verdict on the verdict

The AI tried to close its analysis with a tidy three-word summary of my collecting philosophy. It was the kind of line that sounds profound on first read and like a fortune cookie on the second. I'm not printing it. If a collecting philosophy fits in three words, it isn't a philosophy, it's a slogan.

Here's what I'll say instead. The experiment was worth running. The machine found one genuinely new category, sharpened several I knew by instinct, and confirmed that I've never once bought a watch for its complications. It also eulogized a watch I'd already sold, skipped my goosebumps watch, edited out the evidence that didn't fit its story, and was blind to the category that matters most.

Which is about the right expectation for these tools, in collecting and everywhere else: excellent at reading the metadata, no access to the meaning. The meaning is still my job.