Methodology
How the scores work.
The whole point of this site is trust, so here is the entire system. Nothing on this page is simplified for marketing. This is the actual math.
The short version
We watch the trusted reviewers so you don't have to. Each one's verdict, from loved it to avoid, gets averaged into the score.
- +90+ means they basically all loved it. Around 50 means they're torn.
- +Real tests count more than passing mentions, and no reviewer is counted twice.
- +No brand can pay for a number. Every quote links to the video, so you can check any of it.
that's it. the fine print below is for the skeptics
Where verdicts come from
We track 10 YouTube channels we consider the most credible club reviewers working: Rick Shiels, Club Champion (TXG), Mark Crossfield, Peter Finch Golf, James Robinson Golf, Golf Monthly, MyGolfSpy, TrottieGolf, Today's Golfer, 2nd Swing Golf. That list is editorial judgment. Curation is the quality bar. We watch every equipment video they publish (360 videos so far this season) and extract what each reviewer actually said about each club, from the full spoken review, not the video title or thumbnail.
Extraction is done by AI reading the video transcripts, and every verdict links back to its source video so you can check us.
Each verdict gets a sentiment
Every (reviewer, club) pair is scored on a five-point scale: −2 avoid, −1 lukewarm, 0 mixed, +1 liked it, +2 loved it. A clickbait title with a positive review inside scores positive. The verdict is what they said, not how the thumbnail sold it.
Depth matters
A full on-camera review counts at full weight (1.0). A head-to-head comparison counts 0.8. A passing mention counts 0.3. A rave in passing should not outweigh a critical full test.
One verdict per reviewer
If a channel covers a club five times, only their deepest take counts. No reviewer is ever counted twice on one club, so a channel that uploads constantly cannot dominate a score.
The consensus score
The score is the depth-weighted average of all sentiments, which lands between −2 and +2, mapped linearly onto 0–100. A club at 93 means the average verdict was +1.72, deep into “loved it” territory. 50 means genuinely mixed.
score = (Σ sentiment × weight ÷ Σ weight + 2) ÷ 4 × 100
One more guard: rankings order clubs by a confidence-adjusted version of the score, where small samples get pulled toward the average. Two perfect reviews can look like a 100, but they will not outrank seven near-perfect ones. The number you see is always the raw score.
Agreement and top picks
“Unanimous” means every tester landed within one point of each other. “Split” means the most and least positive testers were 3+ points apart, often the most interesting clubs. A top-pick vote is counted only when a reviewer explicitly names the club a winner or category best, never inferred.
What we don't do
We don't hit clubs ourselves. The reviewers do the real work, and every quote links to them. We don't take money from manufacturers. Retailer links may earn us a commission, and that commission never touches a score: the pipeline that computes them has no idea which clubs pay.
Known limitations
AI extraction from auto-generated captions can misread a garbled club name (we catch and correct these, but not instantly). Clubs marked “early signal” have one reviewer and will move as reviews land. And the channel list is our editorial choice. A different list would produce different scores.