Guides / How our NRL predictions work

Methodology

How our NRL predictions work

Every match on MyNRL.tips carries two numbers: a win probability for each team and a predicted margin. This page explains, in plain English, where those numbers come from — and, just as importantly, what they can and can't tell you.

The two numbers you'll see

A win probability is our estimate of how likely each side is to win the match, expressed as a percentage. If we show the home team at 62%, we're saying that in a game shaped like this one — these teams, this venue, these conditions — we'd expect the home side to come out on top roughly six times in ten. The two percentages always add to 100%.

A predicted margin is the number of points we expect to separate the sides. A predicted margin of 8 doesn't mean we think the game will finish by exactly eight; it's the centre of a wide range of plausible results. Rugby league is a high-variance sport — a couple of tries in five minutes, a sin-bin, a wet ball — so treat the margin as a lean, not a lock.

The signals behind the numbers

Rather than lean on a single stat, we combine several signals that each capture a different part of what decides a rugby league match:

Team strength rating

Each club carries a rolling strength rating that rises when a team beats sides above its level and eases when it loses to those below. Because it updates after every game, the rating reflects a team's current level rather than its reputation or where it sat on the ladder in March. The gap between the two teams' ratings is usually the single biggest driver of a prediction.

Recent form and scoring trends

We look at how each side has tracked over its most recent games — not just wins and losses, but how comfortably. A team grinding out narrow wins and a team posting big margins can have identical records while being in very different shape. Points scored and conceded over the recent stretch help separate the two.

Home ground and venue

Home advantage is real in the NRL, but it isn't uniform. Some venues and some teams carry a much stronger home record than others, and neutral-venue rounds (such as Magic Round) remove the effect entirely. We factor venue in rather than assuming every home team gets the same bump.

Rest, travel and the draw

A six-day turnaround after a Monday night game, a flight across the country, or a run of games against top-eight sides all leave a mark. Short rest and heavy travel tend to shade a team's chances down; a bye or a soft patch in the draw can shade them up.

Team news and availability

Losing a key spine player — a halfback, hooker, five-eighth or fullback — changes a team far more than losing a rotation forward. As team lists firm up through the week, we weigh who's actually named in the 17, with the heaviest emphasis on those spine positions.

Market context

Where betting markets are available, they aggregate a lot of information very efficiently. We use the market as one signal among several — a useful sanity check on the other inputs, particularly for games that look close.

How it comes together

These signals are combined into a single model that outputs the win probability and margin. The model is refreshed as the season progresses, so it keeps learning from recent results rather than staying frozen at its pre-season assumptions. Predictions update through the week as new information arrives — team lists, late changes, conditions — and are locked once a game kicks off so the pick you saw pre-match is the pick that gets recorded.

What the predictions are not

• Not betting advice. Nothing here tells you to place a wager. The numbers are for information and interest.
• Not a guarantee. Upsets are part of the game. A confident prediction can still be wrong, and that's expected over a season.
• Not insider information. Everything is built from publicly available data — results, team lists, the draw and conditions.