I will be simulating season 2020 in real-time as if the games were really being played

[Capital Magpie]

I thought I was the only weirdo growing up. Used to do exactly the same thing for years when I was a kid with the same dice rules, with a few tweaks for the better teams. Had extra teams from Hobart & Canberra. Would design a full draw and player lists and Browlows and AA Teams too. Had about 20 books of results over the simulated years. Then when my parents would come in my room I'd be really embarrassed and pretend I was doing homework or something. Of course I was the Hawthorn Full-Forward, and the bully at school was the Essendon Full-Back. I'd kick 100 goals every year, won a couple of Brownlows and we'd more often than not win the flag

Hey Shotgun, not only were you not the only weirdo growing up, but judging from the no. of Likes and responses from other posters, our simulated footy concepts were quite commonplace and probably good mental health practices.

I must say that I like your modified rules and your simulated league was more advanced than mine, what with Hobart and Canberra being added as extra national league teams; well done! That never occurred to me when I was a kid, although my concept of promotion/relegation from/to Div 1 and 2 (VFL/VFA) was innovative n the extreme in 1968!

 

[TGDesigns]

Speaking of simulated football leagues, for the last year, I have developed a football league of a fictional country, which have the first thirteen seasons written up. All cities are named after generated Madden players from Franchise modes, and the league currently has ten teams.

 

[telsor]

I actually do have to make one, at least for my own benefit, because I need to be able to to read in virtual results in order to predict the next round.

So you'll have a computer generated prediction of how a computer generated virtual season will go....

 

[Anti Septic]

Great work, Final Siren.
I can't wait to see the Tiges back in virtual action on Sunday!

 

[Final Siren]

For some reason, it seems Pies vs Tigers was "played" with full length quarters but Swans-Bombers used shortened quarters?

Yeah, Pies v Tigers went slightly long. It was supposed to be 16 mins + time on, and the generated scores were based on this, but the timing was a bit too spaced out. I corrected my model for the next Squiggle-hosted game (Carlton v Bulldogs).

 

[Final Siren]

Here are the Round 2 results! You can inspect the scoreworm of any particular game by clicking through from https://squiggle.com.au/



A thriller in Perth with the Dockers hanging on against the Hawks.

And here is the ladder, which is real R1 results + virtual R2 results:




Just summarizing how it works:

• There are five participating computer models
• Each game, one is randomly chosen to host that match
• That model produces a randomized simulation of the match - which means it's not the most likely outcome (where the favourite would always win), but rather one where more likely outcomes are more likely to be chosen. In simple terms, if the model thinks that a team is 70% likely to win, there is a 70% chance that it will toss up a sim in which this happens.
• Each round, the models take these results as inputs for the next round. For example, since Brisbane obliterated North, the models will consider Brisbane to be a stronger team when producing sims for Round 3.

 

[Toffee Lion]

What did North do to the sim gods?

 

[spudmaster]

I'll be simulating watching.

 

[VicBased]

Now come on F.S, your a closet P.A supporter ........ Yeah

 

[Final Siren]

Now come on F.S, your a closet P.A supporter ........ Yeah

I do have a soft spot for Port Adelaide. I like how sometimes they are brilliant and cut sides to ribbons, then the next week it's like they forgot how to play football. It speaks to some kind of deep psychological trauma, which, as a Richmond supporter, I identify with.

The Port match was modeled by Nick Fredriksson from The Flag, though, so no Squiggle bias there.

 

[Billy ray]

I do have a soft spot for Port Adelaide. I like how sometimes they are brilliant and cut sides to ribbons, then the next week it's like they forgot how to play football. It speaks to some kind of deep psychological trauma, which, as a Richmond supporter, I identify with.

The Port match was modeled by Nick Fredriksson from The Flag, though, so no Squiggle bias there.

If anyone leans more to the ‘deep psychological trauma’ side of the spectrum, well do i have an offer for you

 

[Topkent]

If anyone leans more to the ‘deep psychological trauma’ side of the spectrum, well do i have an offer for you

Pffft
Try experiencing if for another 50 years
You haven't even done half of Melbourne's finals first finals drought and you still haven't matched our second.

 

[B4Bear]

What did North do to the sim gods?

Refused to go to the Gold Coast.

Those sim gods hold a grudge for a long time.

 

[hawkman]

I used to do the same thing when I was a kid, would roll 2 dice. Left dice was the goals right the behinds. Obviously ment you couldn't come from more than 35 points down in the last quarter.

I did the same but subtracted 1 from each die so you could get 0.0
Also, if a 6 is rolled, that die is rolled again and added on until a 6 isn't rolled. This way a team could conceivably come back from any deficit.

My version had a breeze blowing to 1 end. If you had the breeze 6 was actually a 6, kicking into the breeze a 6 was a 0.
So a team with the breeze could come from 41 behind to win while a team against the wind could only come from 27 down

 

[Rod Stroker]

Refused to go to the Gold Coast.

Those sim gods hold a grudge for a long time.

So the simulator is really just Caro?

 

[silverscope]

Final Siren

I love this idea but I'm also curious as to if it could be turned into a game (Think Football Manager) where you can control a teams players their ratings and then pit them against other teams and have the simulation determine the result.

Does the sim rely on overall team ratings and previous form? Is there a avenue to explore having player ratings affect the team/form to then simulate the results?

I think this type of game would be amazing for a lot of people who like to think they would be great football managers.

 

[Final Siren]

Check this out, I made the AFL website!



Article: https://www.afl.com.au/news/435062/top-eight-full-of-surprises-stats-gurus-create-virtual-season

 

[B4Bear]

How does your simulation hold up for a 17 round season?

 

[Final Siren]

How does your simulation hold up for a 17 round season?

Do you mean in terms of real predictions or simulating a virtual league? Because the virtual stuff has randomness in it, so favourites don't win every game.

 

[Kapow!!!]

Check this out, I made the AFL website!

View attachment 878353

Article: https://www.afl.com.au/news/435062/top-eight-full-of-surprises-stats-gurus-create-virtual-season

Final Siren has appeared on the AFL website: He has sold out, the squiggle has lost all credibility and this thread has lost all meaning...

....in all seriousness though good stuff mate , have been following this for years and it's always either helped sway my tips or just been a good conversation starter.

 

[Anti Septic]

Finally Final Siren is acknowledged as "stats guru".
But did he predict Ken Hinkley would be so happy?

 

[Billy ray]

Sooooo..... is there anyway to weight recent high draft picks so the suns get higher up ?

 

[B4Bear]

Do you mean in terms of real predictions or simulating a virtual league? Because the virtual stuff has randomness in it, so favourites don't win every game.

When the season restarts, noting the first round results and the remaining 16 games.

 

[Final Siren]

When the season restarts, noting the first round results and the remaining 16 games.

When the season restarts, I'll be back to predicting real results. The 17-game fixture is no problem EXCEPT for the possibility that the AFL might not release it all at once. If they don't, it'll be impossible to predict who's going to win what, and what the final ladder will be. We'd only be able to rate teams in terms of their overall strength, without factoring in who plays who and where. Which would be bad, because that's one of the strengths of a computer model: being able to factor in a lot of different things at once.