What Is a Golf Conviction Score? (And Why We Grade Our Own Signals)

Two players, same week, same event. Both show an identical gap between our model's probability and the market's — call it eight points. One of them carries a conviction score of 9; the other, a 4. If the score didn't exist, those two players would look interchangeable on a ranked list. They aren't remotely, and the difference is the most useful thing we publish.

The problem the score solves

Pre-tournament research keeps producing the same dead end: five or six names you feel vaguely good about and no defensible way to rank them. One has lovely course history and cold recent form. Another is striking it beautifully at a venue that's never suited him. The raw numbers don't resolve those tensions — you do, by gut, at eleven o'clock on a Wednesday night. That synthesis step is where most research quietly turns back into opinion.

A conviction score is that synthesis done explicitly. It's a single number out of 10 expressing how much confidence the model has in its own signal for a specific player at a specific tournament, after weighing all eight factors — strokes gained total, course history, recent form, course fit, strokes gained approach, strokes gained putting, driving and travel impact — together.

What a 9 means, and what a 4 means

Take our player with the 9. His strongest strokes gained category matches what the course rewards; his course history points the same way; his form has been trending up for six weeks; and the field context doesn't muddy any of it. Five inputs, one direction. The data picture is unusually coherent — which does not mean he'll contend. It means that if you're going to spend research time anywhere, this signal has the cleanest foundations.

Now the 4. Same headline gap — but under it, the inputs are arguing with each other. Strong irons, ugly course record. Or fine history, built years ago, contradicted by current form. The model still sees a disagreement with the market worth flagging; it's just telling you, candidly, that the flag is planted in softer ground.

We'd rather publish that candour than a false confidence. A tool that only tells you what it likes, never how sure it is, is asking you to trust it blind.

Not the same thing as an "expert confidence rating"

The phrase "confidence rating" usually means "how much the writer likes the pick" — an editorial judgement, applied to whichever players the writer chose to cover. A conviction score is a model output, applied identically to every player in the field. The 80th name in the market gets scored by exactly the same process as the favourite. That consistency is the point: it's what lets you sort and filter a full field instead of reading around a shortlist.

How it changes the workflow

Ours is the boring version of a workflow, which is rather the aim: open the ranked list; filter to conviction 7 and above; check gap sizes; read the plain-English summary on each card — it states which inputs drove the score, so you can argue with it; then cross-reference anything you know that the model can't. What used to be a ninety-minute synthesis job becomes twenty minutes of checking, and the twenty minutes are better spent because the reasoning is on the table rather than in your head.

Every score we've ever attached to a selection sits in the open at proplace.golf/performance — the misses too. If you'd like to watch conviction scoring run on a live field, the 7-day trial is full-access, no card.

Frequently asked questions

What is a golf conviction score?
A model-generated number out of 10 reflecting how strongly and consistently the data signals align for a specific player in a specific tournament, across strokes gained, course history, course fit, recent form and field strength.

What does a high score actually promise?
Nothing about the result — golf doesn't work that way. It promises that the signal's foundations are clean: multiple independent inputs pointing the same direction, rather than one loud number.

Can two players have the same gap but different conviction scores?
Yes, routinely — and it's the most important distinction on the list. Same disagreement with the market; very different quality of evidence behind it.

Does every player get one?
Every player in every PGA Tour and DP World Tour field, each week. No shortlists.

How should I use it?
As a prioritisation tool, not a verdict. Filter high-conviction names first, read the reasoning, then apply your own judgement. The score tells you where the data is confident — what you do with that is yours.

18+ · Please gamble responsibly · BeGambleAware.org. ProPlace is a statistical-analysis and research product and does not accept stakes. Past performance is not a guide to future results.