Lessons from the back room
Ten things about betting prices that the people who set them would rather you worked out slowly. None of it is advice to bet. Most of it is reason not to. Read them in order; each one leans on the last.
About twenty minutes. 18+ only. Nothing here promises a result, and no method removes the risk of losing money.
Lesson 1: A price is a probability wearing a disguise
Before anything else, learn to read the number.
A price of $1.50 is not a reward. It is the bookmaker saying this side wins about two times in three. Divide one by the price and you have the probability they are selling you: 1 ÷ 1.50 = 66.7%. A price of $4.00 says 25%. A price of $1.05 says 95%.
Do this every time until it is automatic. You cannot judge whether a price is fair until you can see the claim hiding inside it.
Lesson 2: The house takes its cut before the match starts
Add up both sides and you will see it.
Two evenly matched sides might be offered at $1.90 each. Convert both: 52.6% and 52.6%. Add them: 105.3%. Real probabilities add to 100, so the extra 5.3% is the bookmaker's margin, built into every price on the board. Some call it the overround, some the vig. It is the rent.
Strip it out and you see what the bookmaker actually thinks: 50% each. Everything you are offered is worse than that by design, and over time the margin is the reason most people lose. A bet only makes sense when you have good reason to think their price is wrong by more than it, and that is rarer than it feels.
Lesson 3: Value is the gap, not the winner
Judge the decision, not the result.
Suppose you have good reason to think a side wins half the time, and the bookmaker offers $2.20. Their price says 45%. Your view says 50%. That gap is what punters call value: if your view is right, then over many bets like it you come out ahead even though you lose half of them. If your view is wrong, the same bet loses money for years.
The side can lose tonight and the bet was still right. The side can win at $1.30 when it deserved $1.60 and the bet was still wrong. If you cannot separate those two things, no model will help you.
Lesson 4: A 70% chance loses three times in ten
Losing runs are not a sign something is broken.
If every bet you place is a 60% chance, you will still lose four in a row about once every forty sequences. Over a season of a hundred bets, a run of four straight losses is close to certain, and a run of six is more likely than not to show up somewhere.
Most people quit, double up, or change their method exactly when the maths says to do nothing. The run tells you almost nothing about whether your prices are good. A hundred bets tells you a little. A thousand tells you something.
Lesson 5: Stake small enough to survive being wrong
The size of the bet matters more than the pick.
A method that is genuinely good can still send you broke if you bet too much of your bank on each play. The Kelly formula gives the stake that grows a bank fastest given your edge and the price, but it assumes your probability is exactly right. It never is.
That is why serious punters bet a fraction of Kelly, a quarter or less, and why the staking tools on this site refuse to stake more than 5% of a bank on anything, whatever the number says. A cap is not caution for its own sake. It is what keeps a losing run survivable.
Lesson 6: Nothing is ever due
The coin does not remember.
A side that has lost five in a row is not more likely to win the sixth because of it. A number that has not come up is not owed to you. Each match is priced on its own, and the past only matters in so far as it changes what you know about the sides.
Bookmakers know most people feel otherwise, and the prices on a side that is "due" often tighten for exactly that reason. The crowd's feeling is in the price. Your edge, if you have one, is in what the crowd is feeling wrong.
Lesson 7: The bet you place to get it back is the one that gets you
Know what chasing feels like from the inside.
It feels like urgency. A loss, then a need to fix it tonight, then a bigger stake on a match you had not planned to bet. Nearly every account that goes to zero does so in a handful of sessions like that, not through a long slow leak.
The defence is boring. Decide the stake before the match. Decide the number of bets before the round. Set a deposit limit with your bookmaker, which the law requires them to offer. When a session is not going to plan, the plan is to stop.
Lesson 8: The closing line is the honest scorecard
Measure yourself against the final price, not the scoreboard.
By kickoff, a price has absorbed every bet placed on the match. The closing price is the sharpest estimate available, and it is a far better judge of your bets than whether they won.
If you took $2.20 and the side closed at $2.00, you beat the market on that bet, whatever the scoreboard said. Beating the close consistently is the only real evidence that a punter has an edge; without it, a winning month is luck. Take $2.00 on sides that close at $2.20 and you are losing even in the weeks you win.
Lesson 9: Most matches are a pass
The house's biggest edge is that you bet every game.
On a typical round, a good model and a sharp bookmaker will agree on nearly every match within the margin. That means no value, which means no bet. A punter who bets eight games a weekend is paying the margin eight times for the pleasure of watching.
Discipline is mostly the ability to look at a card of eight matches and bet none of them.
Lesson 10: Judge a model by its misses
Anyone can show you the hits.
The international model on this site was checked against matches it had never seen, with its settings frozen before the test. Its record, including its worst stretch, is on the method page. It is right about the favourite roughly two times in three, and wrong one time in three. That is what a good model looks like. Anyone claiming much better is selling something.
Treat every tipster the same way. Ask for the whole record, the prices taken, and the closing prices. If they will not show the misses, you have your answer.
The record, so far
The international model named the right favourite in 65.6% of 489 matches it had not seen, with a Brier score of 0.203 against 0.250 for a coin toss. The full method and every stretch of the record.
When the doors open
Members get the model's price on every NRL match from Round 1, the tools to compare it with a bookmaker's, and a tracker that scores their bets against the closing line. The lessons stay free for everyone.
By putting your name down you agree to one email from Matchday Maths when memberships open, with an unsubscribe link, and nothing else unless you ask. Your address is stored on our own server, never sold or shared, and deleted on request. 18+ only.
Chances are you're about to lose. 18+ only. Matchday Maths gives information, not betting advice, and no model can promise a result. Set a limit and stick to it. For free, confidential support call Gambling Help on 1800 858 858 or visit gamblinghelponline.org.au. To block yourself from all licensed Australian online and phone wagering, register at betstop.gov.au.