It is the final pick after the main signals have all had their say.

The model compares both teams, turns that read into a win chance, then tips the side on top.

It is the bottom line: who the site thinks you should back.

Higher means the edge looks clearer. Lower means the game looks tighter.

After weighing the matchup, the model converts that read into a percentage chance.

It tells you whether the tip looks solid or feels more like a 50-50.

It is not trying to nail the exact scoreline, just the likely size of the win.

A separate margin model looks at strength, form and match setup, then estimates the scoreboard gap.

Useful for margin comps and for spotting games the model thinks could open right up.

It is the easy-to-read version of the win probability.

The site sorts the percentage into simple bands like high, medium or coin toss.

It helps you know which tips look sturdy and which ones need caution.

It is a team-strength input, not the public ladder-style rank.

It blends slower strength with faster form from the true-form model, before the public rank is anchored back to record and margin.

It helps the tip keep a stable view of underlying quality, even when the ladder or one huge win is noisy.

It is the credibility-friendly public view: model power plus what the season scoreboard has actually shown.

It starts with conservative model power, then anchors it to win-loss record and average margin.

It stops a side with a poor record from sitting artificially high just because the raw model still likes them.

It measures acceleration, not whether a team is simply hot on the scoreboard.

We compare fast recent form with slower underlying strength.

A dominant side can have negative model momentum if its latest rating has cooled from an even higher peak.

It answers the plain question: are they winning now, and by how much?

It looks at the latest few results, streak and margin, with the newest game carrying the most weight.

This is the signal to use when a team is plainly red hot even if raw model momentum says their acceleration has slowed.

A side can look flat because it has faced contenders, or look shiny because it has dined out on strugglers.

We look at the quality of opponents each team has played and give recent games more say.

It gives fairer credit to teams that have had the tougher draw.

This captures true home-deck advantage and the extra edge at fortress grounds.

We check whether the home side actually owns the venue and whether that ground has historically helped locals.

Certain clubs are simply harder to beat at their own patch.

It looks at interstate travel plus who has had the cleaner break between games.

We compare travel load and rest days for both sides, with extra penalty for backing up quickly after travel.

Tired legs can turn what looks like an even game.

Wet, windy or roofed conditions can help one style and blunt another.

We pair the forecast with how each team tends to defend, move the ball and score.

Ugly conditions can drag a polished side back to the pack or suit a tougher, scrappier one.

This is about the recent history between the two clubs, not their whole season.

We look at recent meetings and give the newer ones more weight than the older ones.

Some matchups stay awkward even when the ladder says they should be even.

It cares about the quality of the missing cattle, not just the raw count.

We compare the likely side with the club's recent best team and mark down tests, managed players and confirmed outs.

Losing two stars hurts more than losing two fringe players.

It turns absences into a rough points penalty.

The model estimates how much quality is missing from the expected lineup and converts that into a simple impact number.

It helps separate a manageable injury list from a genuine personnel problem.

Lower is healthier. Higher means the next game looks more affected by outs.

It rolls the injury model into one easy number for that club's next matchup.

It lets you glance at which sides are going in well stocked and which ones are stretched.

It is a quick view of whether the announced crew has historically suited one side more.

We look at how each club has gone under that crew, especially against expectation.

It can be a useful tie-breaker when the other factors look close.

This is the model's conservative read on home-whistle and margin tendencies.

We shrink umpire history heavily so one odd patch does not get overplayed.

It adds a small nudge rather than driving the whole tip.

It shows whether a team usually wins or loses the free kick count.

We average free kicks for and against across completed games.

It is handy context when you are weighing up umpire trends and game style.

Wins get four points, draws get two, and percentage breaks ties.

Exactly the same way the AFL ladder works.

It is a quick guide to the season picture, even if it is not the whole story.

High percentage usually means a team scores well and does not leak easy points.

It compares total points scored with total points conceded.

It helps separate strong sides from teams just scraping home.

It is the scoreboard gap you usually expect from that team.

We average the margin from completed games.

It can show that a side is stronger or weaker than its record suggests.

It tracks how ladder spot, power and injury picture have moved across rounds.

We plot those numbers after each completed round.

It shows whether a side is building, sliding or just holding steady.

Some teams travel to certain venues better than others.

We use past results at the venue to show win rate and average margin.

Ground history can add confidence to a venue edge.

This is the bookmaker read once odds are turned into a clean win chance.

We convert the head-to-head odds into probabilities.

The market is a strong reality check and usually carries good information.

If the line moves, the market grew more confident in one side.

We compare later prices with earlier ones.

Late moves can hint at team news or sharper market opinion.

A roughie is an underdog the model rates better than the market does.

We compare the model's win chance with the market's win chance.

It helps spot the bold tip that is not just a blind dart throw.