Model Probability vs Market Consensus: Why the Gap Is the Signal

Every week our model produces a probability for every player in every PGA Tour and DP World Tour field. And every week, the honest answer to "so what?" is: on its own, not much. A number that says a player is 12% to finish top ten means very little until you know what the market thinks. If the market says 16%, the model is telling you nothing you couldn't get from an odds screen. If the market says 8%, now you have something — a disagreement. That disagreement is the entire product.

Consensus is good. That's what makes beating it interesting

Market consensus isn't stupid — it aggregates enormous amounts of information, quickly. But it also aggregates narrative. A player who won two weeks ago and is all over the coverage gets priced for the coverage as well as the golf. A player who missed a high-profile cut at a course that never suited him gets marked down for a result the data largely predicted. A model has no idea who's trending. It processes strokes gained, course history, course fit, recent form and field strength the same way for the 3rd name in the market as the 93rd — and where its output diverges materially from the market's, one of two things is true: the model is missing something, or the market is. Working out which is the research.

We'd add a caution most tools skip: sometimes the model is the one missing something. A withdrawal rumour, a swing change, an injury the data hasn't caught up with. The gap is where to look, not what to conclude.

Not all gaps deserve your attention

This is the part we bang on about internally. A wide gap built on thin inputs — sketchy course history, a small form sample, a field composition the model finds hard to read — is noise wearing a signal's clothes. So every player card carries a conviction score out of 10 alongside the gap: a measure of how cleanly the eight factors line up behind it. A 12-point gap at conviction 4 and a 7-point gap at conviction 9 are different animals, and the second is usually the more interesting one. We've noticed subscribers learn this faster by watching a few weeks play out than by being told, which is partly why the trial exists.

Why we run the whole field

Most research products hand you a shortlist — someone has already decided which 20 or 30 names are worth your time. The trouble is that the decision was made before anyone looked at the data, and the market's blind spots are, almost by definition, not on the shortlist. Some weeks the largest model-versus-market disagreement in the field sits on a player ranked 60th in the consensus. You will never find him in a curated top 20, because nobody curated him in. Full-field coverage isn't a feature bullet; it's the difference between research that confirms opinions and research that occasionally produces a genuinely new one.

And why every card explains itself

Raw probability tables are fine if you enjoy decoding them. Most people doing serious pre-tournament research don't want to build the interpretation layer themselves — so each card states, in plain English, which inputs drove the gap and what the conviction score reflects. If a player's been flagged, you should be able to see why in under a minute, and disagree with it if you know something the model doesn't. That last clause matters. The gap is a research signal, not an instruction.

The full record of how these signals have performed — every selection, every miss — is at proplace.golf/performance. Audit it before you decide anything. Then, if you want to watch a live week's gaps for yourself, the 7-day trial is full-access with no card required.

Frequently asked questions

What is the gap between model probability and market consensus?
The difference between what a data model calculates a player's probability of a top finish to be and the probability implied by market prices. A material gap means the data and the market disagree — which is where research effort pays best.

Is a bigger gap always a better signal?
No. Gap size matters, but so does the quality of the inputs behind it. That's what the conviction score (out of 10) measures: how cleanly strokes gained, course history, course fit, recent form and field strength align. A moderate gap at high conviction usually beats a wide gap at low conviction.

Why does full-field coverage matter?
Because shortlists pre-filter the field before you see any data, and the market's mispricings are rarely among the names everyone is already discussing. Running every player through the same process means the signal can surface anywhere.

How often do the rankings update?
In-week, as new data lands. Monday's list and Wednesday's list can differ.

Can I check the track record first?
Yes — it's published in full at proplace.golf/performance, misses included, before you subscribe.

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.