Every watch in one rankable database.
Candidates and the collection share one master list with research, variants, media, and citations attached. Everything stays in this file until you export a backup.
Filters
Personal-fit lens & buy-first order
weights, rules, and the value model
Eight editable weights. They recalculate the fit score live and never overwrite your manual rank.
Buy-first order
Blends personal fit, absolute execution, value for money, and which of your five tiers a watch sits in, then ranks candidates by which to buy next. Value for money means cheaper is better if it scores well: price sits on a fixed $100–$100,000 log scale and that cheapness credit is scaled by the watch’s fit and execution, so a mediocre bargain earns less of it than an excellent one. Sort the database by Buy first to use it. Weights are normalised, so they need not total 100.
Personal rules and current thinking
Collector profile
what you're drawn to, computed from your own ranking — no AI involved
Discover
the agent proposes watches you don't track yet, from your profile and ranking
Master database
Ranks are yours alone. Provisional rows keep a deterministic order until you place them. The retired diver-research list is not imported.
Drag candidate rows to reorder, or expand a row and use the ↑ ↓ rank buttons. Set an exact rank number in Edit.