I remind you again this week that graphic design is my passion.

With the opening round being reduced to two games thanks to some guy named Alfred, we only got half the round we were promised. Both tips were correct and that’s really all you can ask for. MAE has blown out a bit to start the season, but that’s not a problem with such a small sample size.

Some movement happened this week, with Hawthorn jumping up two places into first and closing in on having the best midfield rating in the league. Sydney and Collingwood both fell two spots, while GWS didn’t move but still closed the gap to 7th placed Freo. GWS also took advantage of the fact that 14 sides haven’t played yet and took Best Defence honours, but we’ll have to see if that stands up this weekend (it probably will since they’re playing Melbourne’s paltry offence).

Hawthorn, thanks to their head start, are our sole premiership favourites as we wait for the Dogs and Lions to start their 2025 campaigns. GWS also jumped into the top 4 at the expense of the Swans (and footy fans applaud) and the top 8 at the expense of the Blues (and footy fans applaud even louder), while Collingwood tumbled down into 14th (and footy fans start doing celebratory backflips).

Barring any other natural disasters, here’s the tips for the week. The sole upset tip is for Port Adelaide over Collingwood, but we could see favouritism shift depending on injury news. STaRS is also a lot less keen on Geelong than most, so we’ll see how that turns out.

Have a great week!

With all of my attempts to innovate falling flat on their face, STaRS 2025 is largely the same model as STaRS 2024. There are some minor differences (which I hope will be improvements) but not a lot has changed. Brisbane go into the season as the team to beat. However, the Bulldogs and Hawks aren’t too far behind.

I don’t mean to shock anyone but the projected ladder largely mirrors last year’s ladder with some adjustments for teams who were in great form at the end of the season. Port, Geelong and GWS are replaced in the top four by Brisbane, Bulldogs and Hawthorn for late-2024 form reasons.

Will this actually happen? Of course not! The ladder is very dynamic each year and will change drastically from pre-season projections. Did you know that, in this century, there’s only three occurrences of three teams from the bottom 4 returning to the bottom 4 next season? And that all three were in the wake of GWS and Gold Coast entering the competition and sucking up all the good draft picks for a few years? I think this fact demonstrates how much things change year-on-year, and everyone’s shoe-in bottom 3 of Richmond, North and West Coast actually has a fairly small chance of happening.

Graphic design is my passion. Since visuals now take an hour of coding, this is just how I’m showing fixture difficulty this year and you’ll have to deal with it. This will largely be obsolete within a few months and we’ll know how difficult the fixtures actually are at the midway point of the season. If you don’t know how I calculate fixture difficulty, I simulate the fixture and then an ideally competitive 34-round fixture and look at the differences, accounting for the differences in the amount of games played.

The fixture looks to be largely what you’d expect. Collingwood seem to have gotten one a little more challenging than they’d want, but since they are recent premiers, I (and most of the country) will be willing to overlook this. Carlton are on the flip side of this, with a softer draw than would be expected.

Melbourne and Adelaide are the main recipients of the fixture equality system, both getting a 10% better chance of playing finals than a more even fixture. Most of the numbers in the above table are fine. They’ll all buff out, but 10% is a hefty boost and much more significant.

Round 0 (hee hee) tips are in and the games are looking pretty close based on what occurred last season. I personally think that you should vastly overestimate home ground advantage at the start of the season, so I won’t be tipping against Sydney. But you can do what you want.

So, what’s next for Stattraction? My priorities this summer are:
* Get all the reactable visualisations done for AFL/W and Climbing.
* Update STaRS to a 2025 version with the ideas I have for that and optimisation and testing.
* Update ScArPer, my climbing model, to a 2025 version that’s a little less scuffed. Maybe also add speed climbing to the things it can model.
* Create an ice hockey model for the Professional Women’s Hockey League. I have all the data I need and there aren’t many citizen modellers for it at this stage. The season started this morning so I have no deadline now that I’ve missed the opening games.

Tip percentage and MAE were a bit lower than last season. But STaRS also barely finished positive in Bits last season, so finishing at 20.21 Bits is a significant improvement on last season. That’s really positive news!

Once again, we had no movement in the positions. However, North were sharing no. 1 midfield honours with Adelaide last week and they’ve put that debate to bed this weekend by dominating inside 50s against Brisbane.

Best overall, attack, midfield and defence: North Melbourne
Worst overall and midfield: West Coast
Worst attack: Collingwood
Worst defence: Sydney

Above is how teams compared with the re-simulation of the fixture with end-of-2024 data. In it we can see the teams that STaRS thought overperformed and underperformed. The top will almost always have overperformers and the bottom will always have underperformers. But some of the scale here is still worth mentioning.
The biggest overperformers: Hawthorn, Fremantle, Port, West Coast, Carlton (I’d be especially worried if I was one of those latter two).
The biggest underperformers: Geelong, Sydney, GWS, Gold Coast, Collingwood (Geelong in particular should expect to do much better next year).

Over to hindsight fixture difficulty now. The AFL did well in giving the worst four teams of last year the easiest four fixtures, even if the result was probably a little more extreme than they probably wanted. Sydney getting an easy fixture and dropping off the face of the Earth is… not usually meant to happen, especially for an up-and-coming team. Melbourne and Geelong were two teams that likely should’ve made finals, but got the two hardest fixtures and missed finals as a result.

I’ll see you next year, early March at the latest.

Springfield was hosting the danger game and, yep, it was a dangerous. STaRS got the tip wrong and the Lions were comprehensive in their victory to set up a rematch from last year’s grand final.

Pretty straightforward again this week, with no movement. Ratings have gone back to the previous scale before Hawthorn and Port screwed them up.

No surprises for the big one, as STaRS thinks North should win somewhat comfortably in the granny this year. And before you mention that North were also tipped last year in this same game against the same opponent at the same ground, that was a 64% tip by 5.6 points. There’s a lot more confidence with this tip. North have been the team to week for pretty much the whole season, and they’re going to prove a mighty obstacle for the Lions to win their second straight premiership.

This is the best I can do for tipping right now since I haven’t gotten started on this table’s replacement yet. The first game was fairly straightforward and Freo never looked a chance. The second game, however, was a big upset and a comeback for the ages. It cost the model a positive Bits record for the week but oh well.

Now, this is a new graphic. It’s pretty close to the old one with some changes. Luckily, I didn’t have to worry about how to show changes in ranking this week. That’ll be something to tackle next time it happens, but no-one moved this week.

You may have noticed that ratings have plummeted in value. This is largely due to a trend away from the midfield rating being the most important, displayed by Port in winning their game despite having 22 less inside 50s than the Hawks. Midfield is still ~4.5 times more important than the attack or defence ratings, but it’s previously been as high as 8.

Not a lot of changes from the standings last week. North held steady at 66%, and Port went from 0.01% to 0.07%. Adelaide picked up most of the percentage from Hawthorn and Fremantle (which was barely anything) and also took 3% from Brisbane thanks to their big win.

… yeah. Good luck, Port! I can’t say I’ve seen many prelims with a 98% confident tip. The real battle will be over at Springfield, where STaRS is somewhat confident in Adelaide winning. We’ll see how it turns out.

Also I’ve gotten a Bluesky and I’m already more active there than I ever was on Twitter. Follow if you want random slightly interesting tidbits.

Fairly expected results outside of Port coming alive against Richmond. It wasn’t enough to dampen the model’s results this week but it was unexpected.

Wow! Boring! The Roos are now the number one attacking team, becoming best team in all three ratings, but that was the most significant change in the ratings this week.

After getting premierships in 0% of all simulations last week, Port can now proudly say that they won a premiership in exactly 1 simulation. Congrats to them!

North solidified their position as premiership favourite. Hawthorn and Freo are running into them in the prelim, and STaRS doesn’t think either have a particularly good chance of winning that one.

Two matches and two very firm favourites. We’ll find out if anything crazy happens.

An awesome week, mostly due to all the favourites winning. We had only the one AE blowout, too.

West Coast end the season in last place. Again. They got a few more wins this year but ultimately their rating ends in an even worse place.

You can see above that Port are the worst rated team of the remainders, and, honestly, it’s for good reason. They’ve played four top eight teams and lost each one – some by a lot.

STaRS’ belief that there are only three teams in it continues, but maybe a few wins could change things.

I doubt there are any upsets here, but I haven’t checked. I guess there’s a chance Port have been given the edge over Richmond, but I doubt it.

Edit: Yes, I do normally change the team names for Indigenous Round, but I don’t have the time to fiddle with the backend and change things around this week. Just pretend that I changed them.