Trends this week
Loading…
What do “Own rate”, “Lift” and “Chance” mean?
- The bet
- The selection the record implies, with the number it would be graded against. Period lines carry an est. tag: no bookmaker publishes historical quarter or half lines, so the number is this game’s posted full-game total or handicap scaled by the share of scoring that period actually receives — about 18% for Q1 and 31% for Q2, not 25% each. It is a fair benchmark, not a price, so no value or return is claimed from it.
- Hit rate
- How often the thing happened in the games matching this situation. On its own it says very little — a 73% record means nothing if the team does it 70% of the time anyway.
- Own rate
- The number the trend has to beat. How often this same team does the same thing across all its games, ignoring the situation. It is the matched baseline: comparing a trend to 50% would flag every good team, so each trend is compared to itself instead.
- Lift
- Hit rate minus own rate. The part of the record the situation might explain, rather than the team being the team.
- Chance
- How often luck alone produces a record this lopsided. The team’s real results are reshuffled at random a few hundred times and the same situation re-counted. 0% means no shuffle matched it; 40% means four shuffles in ten did, and the “trend” is noise. Low is interesting, not proven — with this many rows on the page, some will read low by chance.
Break-even at standard −110 pricing is 52.4%, not 50%. A trend hitting 51% is a losing bet at a number that looks like a winning one.
Read the chance column first. It says how often a record at least this good appears when that team’s own results are reshuffled across its own situations — which holds the team’s win, cover and over rate exactly fixed, so a good side’s 80% at a ground is not flagged as a venue effect when it wins 80% everywhere. And it is a per-trend figure. A full round, eleven situations, three markets and up to seven scopes is thousands of lookups, so a trend reading 5% is not a one-in-twenty edge — it is the sort of thing thousands of lookups throw up dozens of times from nothing at all. Nothing on this page has been shown to beat the market. Use it as a shortlist to write down and watch forward.
Checking how these trends have actually done…
How a trend gets onto this page
1. Only completed games count. The fixture being analysed is never part of the history it is measured against, and “after a loss like their last” is read from the game before, never from the one being played.
2. It has to clear a strength band. A perfect run of 5 or more, a perfect run of 8 or more, 80%+ over 10 games, or 70%+ over 20. Records the other way round count too — a side that keeps failing to cover in a situation is the same trend seen from the other side, and is listed as “against”.
3. It has to beat the team’s own rate by at least 10 percentage points. This matters more than it sounds. A quarter total is graded against the full total divided by four, but quarters are not a quarter of the game each — second quarters are nearer 31% and third quarters nearer 21% — so every team clears a second-quarter over about 69% of the time. Without this step the page would fill up with “80% over in Q2” rows that say nothing whatever about the team. The “own rate” column is there so you can see the gap for yourself.
4. Then it is permuted. That team’s results are shuffled across its own situations 200 times and the record recounted. The chance column is the share of those shuffles that matched or beat the real record, so it moves in steps of 0.5% and “0%” means none of the 200 — not that it is impossible.
And the last two columns are not two votes. The shuffle is centred on the team’s own rate, which is exactly what the lift is measured against, so a big lift and a low chance are largely the same observation reported twice. Read them together as one finding, not as one result confirming another.
Trends by game, for every market and every scope
Most trend lists you will find are a single market on the full game: a team’s record against the line, or its head-to-head history with an opponent. This page covers head to head, the handicap and total points, and runs each of them over the full game, the first half, the second half and all four quarters for AFL and NFL — rugby league plays halves, so the NRL has halves only and no quarter rows are invented for it. Every row also names the situation it comes from, so you can tell a venue record from a timeslot record from a response to last week’s hiding.
Why the chance column is the only one that settles anything
Take one code and count the combinations: around seventeen teams, forty-odd grounds, twenty-seven rounds, seven days of the week, thirty officials, three markets, up to seven scopes. That is tens of thousands of team-situation cells. A five-from-five run turns up by chance roughly once in thirty-two tries on a coin-flip market, so a table that size holds hundreds of perfect five-game runs even if no ground, round or official makes the slightest difference anywhere. The question is never how long the streak is. It is how many streaks that strong you would expect to find if nothing were real — which is what the permutation measures and what the chance column reports.
What this is not
It is not a tip sheet and it is not a model. Our own walk-forward testing puts every priced market below the break-even a bookmaker’s margin demands, and the league-wide verifier finds no situational family producing strong records more often than chance in the NRL, the AFL or the NFL. A trend clearing every filter here is a candidate worth recording and following forward, nothing stronger.
What is this page?
The trend verifier asks a league-wide question and answers it honestly: across every team and situation, strong records turn up about as often as chance makes them. That is worth knowing but it is no use on a Saturday, because it never tells you which records are live for the games actually coming up. This page starts from the fixtures instead. For each game it works out the situations both sides fall into — this ground, this round, this day, this official, this opponent, how each team responded to a result like its last one — then lists every historical record attached to them, for head to head, the line and total points, across the full game, both halves and (where the sport plays them) all four quarters.
Do betting trends actually beat their own baseline?
| Code | Predictions graded | Trend hit rate | Matched baseline | Edge |
|---|---|---|---|---|
| NRL | 8,251 | 55.3% | 57.5% | -2.1 |
| AFL | 18,291 | 53.6% | 56.6% | -2.9 |
| NFL | 17,168 | 56.5% | 59.6% | -3 |
The answer measured here is no. Across all four codes, situational trends hit at or below the rate the same teams were already achieving over the same matches, so the apparent edge is the baseline, not a finding. The comparison is against each team's own rate — never 50%, never the league average — because a trend that fires mostly on good teams will look predictive against any other yardstick.