An NRL betting strategy is a repeatable way to price a game yourself and bet only when a bookmaker offers more than your price. In rugby league the method rests on four things: final margins bunch on even numbers, field goals make one-point results common, the final 19 is confirmed only 90 minutes before kick-off, and wingers score far more often than forwards.
The margins below come from the NRL's results for all 2,227 premiership games from 2016 to 2026, finals included, and the try rates from its match records for the 426 games of 2025 and 2026. None of it picks a winner; it prices the bet, the part of NRL betting you control.
A simple NRL betting model: inputs and outputs
A betting model turns inputs into fair prices. For the NRL it has five parts:
| Input | What it prices |
|---|---|
| A dated margin table, recent seasons first | Lines and margin bands |
| Each team's expected points, from a line and total or your own ratings | Totals and try counts |
| Try rates by position | Try scorer bets |
| Team news, up to the final 19 | When to bet, and which bets can be voided |
| Settlement rules in your bookmaker's terms | Extra time, draws and named players who sit out |
From a line and total, expected points are Team A = (total + start) / 2 and Team B = (total - start) / 2. With an illustrative total of 47.5 and Team A giving 9.5 points, Team A is on (47.5 + 9.5) / 2 = 28.5 and Team B on (47.5 - 9.5) / 2 = 19.0.
Key numbers: where NRL margins land
Final margins, either way, as a share of each period's games:
| Final margin | 2016 to 2020 (973 games) | 2021 to 2026 (1,254 games) |
|---|---|---|
| 1 | 7.0% | 5.3% |
| 2 | 9.1% | 8.8% |
| 3 | 0.9% | 0.4% |
| 4 | 8.6% | 8.0% |
| 5 | 0.8% | 0.5% |
| 6 | 7.6% | 6.0% |
| 7 | 1.6% | 1.0% |
| 8 | 7.9% | 7.8% |
| 9 | 0.8% | 0.4% |
| 10 | 4.4% | 5.4% |
| 11 | 0.2% | 0.3% |
| 12 | 7.2% | 4.4% |
| 14 | 5.8% | 6.4% |
In both periods every even margin in the table landed more often than the odd ones beside it, and all but 2 against 1 at least four times as often, because tries are worth 4 and goals 2. In both, 2, 4 and 8 were the three most common margins. Odd margins of 3 to 11 stayed rare: 43 of 973 games (4.4%) in 2016 to 2020 and 32 of 1,254 (2.6%) in 2021 to 2026.
One pattern did not hold: 12, two converted tries, fell from fifth of the margins in the table in 2016 to 2020 to eighth in 2021 to 2026, below 10 and 1, while 14 rose to fourth. Games averaged 40.9 points in the first period and 45.9 in the second. Rebuild the table each season and weight recent seasons more. NRL line betting shows how each margin settles a line.
Pricing a half point between two bookmakers
Neighbouring lines differ only on the exact margin between them, so take the harder line at price B over the easier one at price A only when:
B / A > 1 + p / q
where p is the chance of that exact margin and q is the chance the harder line wins. In an illustrative game, Team A has a 50.0% chance of winning by 8 or more. As a rough start, give Team A half of each 2021 to 2026 share: 1.0% / 2 = 0.5% for winning by exactly 7, and 7.8% / 2 = 3.9% for exactly 8.
| Team A line | Wins if Team A wins by | Chance | Price | Average return per $1 |
|---|---|---|---|---|
| -6.5 | 7 or more | 50.5% | $1.90 | 0.505 x 1.90 = 96.0c |
| -7.5 | 8 or more | 50.0% | $2.00 | 0.500 x 2.00 = 100.0c |
| -7.5 | 8 or more | 50.0% | $1.90 | 0.500 x 1.90 = 95.0c |
| -8.5 | 9 or more | 46.1% | $2.00 | 0.461 x 2.00 = 92.2c |
Across 7 the extra 10 cents is worth taking: -7.5 beats -6.5 at $1.90 from 1.90 x (1 + 0.005 / 0.500) = $1.92. Across 8 it is not: -8.5 needs more than 1.90 x (1 + 0.039 / 0.461) = $2.06 to beat -7.5 at $1.90. So when a line moves in half points at a flat price, the steps across 3, 5, 7, 9 and 11 cost you little, while the steps across 1 and the even margins from 2 to 14 cost a lot.
Risk: Betting involves risk. A better line lowers the average cost of a bet, but it can still lose, and past margins do not predict the next game. See responsible gambling for limits and support.
Field goals and lines near 1 and 2 points
Under the NRL's 2023 Telstra Premiership interpretations, a field goal scores 1 point, or 2 from a kick in general play taken beyond the opposition 40-metre line. The NRL's match records for 2025 and 2026 show 40 one-point and 7 two-point field goals in 426 games. Of the 40 one-pointers, 34 came after the 70th minute, 9 of them in extra time.
So the smallest lines behave very differently:
| Favourite's line | Also loses on a win by exactly | Share of games, 2021 to 2026 |
|---|---|---|
| -1.5 (against the head to head) | 1 | 5.3% |
| -2.5 | 2 | 8.8% |
| -3.5 | 3 | 0.4% |
| -4.5 | 4 | 8.0% |
So -2.5 and -3.5 are nearly the same bet, while the step from -1.5 to -2.5 is one of the biggest half points there is. An underdog's +1.5, +2.5 and +3.5 read the same way in reverse.
Team lists: price the final 19, not Tuesday's list
Since 2026 clubs name squads on Tuesday, cut two players 24 hours before kick-off and confirm the final 19 at the 90-minute mark, with four of the six on the bench allowed on. Until 2025 it was a final 17 plus an 18th player used only in set circumstances. Prices can move at each step, so bet before the news only when the price is already wrong, and leave all-in player bets until the final 19, as NRL team lists betting explains.
Golden point in your numbers
Scores level at 80 minutes send a regular-season game into golden point, up to 10 minutes in which the first score wins; a final plays a full 10 minutes of extra time before that. Whether those points count is a settlement rule that decides three things. A head to head on a game still level is a dead heat, a refund or a loss, as the terms say. Any line under 4.5 turns on it, because no regular-season golden point win from 2016 to 2026 was by more than 4. And it sets the margin band of a golden point win.
For bookmakers authorised in NSW, the NSW Bookmaker Declared Betting Events Betting Rules count extra time unless something else is agreed (clauses 5.1.2 and 5.1.3), but settle a market that prices a draw at full time (5.1.7). Other bookmakers write their own terms. NRL margin betting covers bands and draws.
Try scorer prices from position base rates
Start a try scorer price from position. For the 426 games of 2025 and 2026, counting players who took the field, by starting position:
| Group | Games | Scored a try | Fair anytime price |
|---|---|---|---|
| Wingers and centres | 3,408 | 38.76% | $2.58 |
| Fullbacks | 852 | 30.99% | $3.23 |
| Halves | 1,704 | 21.07% | $4.75 |
| Starting forwards (props, hookers, back rowers) | 5,112 | 12.93% | $7.73 |
| Bench players who took the field | 3,370 | 7.86% | $12.72 |
Fair anytime price = 1 / scoring rate, so 38.76% gives 1 / 0.3876 = $2.58. On their own, wingers scored in 46.0% of their games and centres in 31.5%; NRL try scorer betting has every position.
A group rate belongs to an average attack, so scale it by the team's expected tries. Teams averaged 4.10 tries and 23.5 points a game, about 5.73 points a try, and a winger 0.643 tries, or 0.643 / 4.10 = 15.7% of his team's. With the model's illustrative line and total:
| Step | Team A winger | Team B winger |
|---|---|---|
| Expected team tries = expected points / 5.73 | 28.5 / 5.73 = 4.97 | 19.0 / 5.73 = 3.32 |
| Winger's expected tries = team tries x 0.157 | 0.780 | 0.521 |
| Chance of a try = 1 - exp(-winger's expected tries) | 54.2% | 40.6% |
| Fair price = 1 / chance | $1.85 | $2.46 |
Treat the formula as a check: at an average 4.10 team tries it gives a winger 1 - exp(-0.643) = 47.4%, against the 46.0% wingers scored, so shade wingers down a little; it also knows nothing about the player. Poisson distribution betting explains the 1 - exp step, the chance of at least one try.
Risk: Betting involves risk. Even at a fair $1.85 the winger fails to score about 46 times in 100, and a price above your estimate can still lose. See responsible gambling for limits and support.
Testing the method, and the mistakes that cost most
Log each bet's price beside the last price before kick-off, its closing price: with illustrative prices, a line taken at $2.00 that closed at $1.85 beat the close by 2.00 / 1.85 - 1 = 8.1%. Beating the close over hundreds of bets is evidence your prices beat the market's; a few dozen bets prove little either way. Closing line value explains the measure, and positive EV betting shows how a fair price becomes an expected value.
| Mistake | What it costs | What to do instead |
|---|---|---|
| Pricing margins off a smooth curve | A curve puts 7 between 6 and 8; from 2021 to 2026, 7 came up in 1.0% of games and 8 in 7.8% | Use the measured table, rebuilt each season |
| Pricing every winger alike | $1.85 and $2.46 for the same position in one game | Scale by each team's expected tries |
Skip a bet when you cannot say which rule settles it, or when no price on offer beats your own fair price.
Where the B337 screeners fit
B337's +EV screener does the comparison step across sports markets: it lists bookmaker prices above an estimated fair price and ranks them by edge. That fair price is an estimate, not a known probability, so check a row with the methods above before betting it.
The +EV, arbitrage and middles screeners come with Terminal Pro and Full Automation, or as their own Screeners plan. A free account sees the live counts with the rows locked, and you place every bet yourself. Bookmakers can limit stakes, void bets or close accounts, many restrict automated betting in their terms, and that risk is yours.
For free and confidential support call 1800 858 858 or visit gamblinghelponline.org.au.
Risk: Betting involves risk. A +EV bet can still lose, and often does, an arbitrage or a middle only works if every bet is accepted at the price shown, and a screener row is a price comparison, not a prediction. See responsible gambling for limits and support.