https://thearcfooty.com/2018-afl-projections/

The Arc have released their projected ladder for 2018 and they have the Blues finishing in 14th place.

We have a 17.2% chance of finishing in the top 8.

I would like to see us do a bit better and tackle the 4 teams projected just above us who are:

13. Bulldogs
12. Pies
11. Dees
10. Eagles

The other interesting thing is it shows the teams we must beat,

15. Kangas
16. Dockers
17. Suns
18. Lions

....starting this weekend.

CB

Cool, I was waiting for the Arc

_________________

I Noah how you feel.

Noice methodology.

Predicting Carlton with 9 wins and 13 losses.

We play

Kangas x 1
Dockers x 2
Suns x 2
Lions x 1

I assume the above are where they get 6 wins.

Who do they predict the other 3 wins come from?
We play...

Bulldogs x 2
Pies x 2
Dees x 1
Eagles x 1

I assume they predict 3 wins come from the above 6 games.
If we win all those games that gives us 12 wins (if we beat the teams below us on prediction ladder)

Last year we beat

Swans x 1
GWS x 1
Hawks x 1

Add 3 unlikely wins and there's 15 wins.

Anything can happen.

_________________
Everyone looks good in Navy Blue

Brisbane won’t finish last, surely ?

Oh no. Are we about to get flooded by Ark puns now?

_________________
Never argue with idiots. They drag you down to their level and then beat you with experience!!!

After Monday and Tuesday, even the calendar says W T F .........
Visit http://fromthemoshpit.com/

Hi,
Glad you like my site! Cheers for reading.

In response to bondiblue who was wondering where my model gets about 9 wins for Carlton from:

My predictions are probabilistic. By that I mean that for each game, I estimate Carlton's chance of winning - it can be somewhere between 0% and 100%. The projected wins are then a sum of all these probabilities. For example, say I gave the Blues a 70% chance of beating the Dockers (hypothetically) and a 20% chance of beating the Crows. The expected wins from these 2 games is therefore 0.7 + 0.2 = 0.9.

This sort of probabilistic calculation gives you more accurate forecasts than if you just look for the likely wins (eg. Kangaroos, Dockers, Suns, Lions) and give them a probability of '1' and all the others a probability of '0'. The Blues, like any team, will win some games you expect them to lose and vice versa.

For what it's worth, this is my current estimate of the Blues' chance of winning each game. Add up all the probabilities and you should get 8.7. These estimated probabilities will change after each game, because my team ratings 'learn' from real world results.

R1 - RI - 0
R2 - GC - 0.69
R3 - CW - 0.4
R4 - KA - 0.46
R5 - WC - 0.51
R6 - WB - 0.4
R7 - AD - 0.17
R8 - ES - 0.39
R9 - ME - 0.39
R10 - GE - 0.24
R11 - SY - 0.18
R12 - bye
R13 - FR - 0.66
R14 - CW - 0.42
R15 - PA - 0.44
R16 - BL - 0.45
R17 - SK - 0.4
R18 - HW - 0.5
R19 - GC - 0.42
R20 - GW - 0.43
R21 - FR - 0.43
R22 - WB - 0.43
R23 - AD - 0.43

WOW.

Thanks for the explanation.

Fascinating, and Interesting

Good effort.

I'm in the camp who believes we can win more than 10

... and if we improve our periods of momentum we could win 13.

Glass half full....especially with maturing quality kids

_________________
Everyone looks good in Navy Blue

Cheers! I've gotta say - I'm not a Blues fan but I like the positive, constructive tone on this board. It's a refreshing change.

I think the Blues *can* win 10+ games as well. There's a huge range of uncertainty around any forecast of the season, including mine. At my projections page, you can see the range of uncertainty around my forecast for each team's win total: https://thearcfooty.com/2018-afl-projections/

I estimate that Carlton has a 70% chance of winning somewhere between 5 and 12 games this year. That's a super wide range - this early in the season, anything can happen. The predictions get a lot more accurate after a few rounds.

Stats and probability are interesting, I will follow your updates.

Interesting to see we only have a 45% chance of beating the Lions when they are expected to come last on your predictions and a 50% chance of beating Hawthorne, who you have in the eight.

I would like to hear the method you use to come up with these numbers

Cheers
GMC

Thanks for contributing Bruce, fantastic to hear directly how statistical analysis plays out in modern sport and specifically how it applies to the Blues.

One thing about your approach that interests me, and I have read an article you have written on the topic, is that you do your analysis on the teams lists and not the team that is selected from week to week.

I understand that over the longer run there isn't a great deal of difference in the statistical results, although I suspect if you were using the data for gaming purposes there may be some shorter term variance.

Therefore, if you can reasonably accurately determine a teams likely outcome at the start of the season it shows that the rate of improvement of a list over one season usually can't be material and confirms that a rebuilding process takes time.

Thanks again for producing the Arc Footy.

CB

Hi,
Home-ground advantage accounts for the difference. The Blues play the Lions at the Gabba, but play the Hawks at Etihad. Even though my model rates the Lions as a worse team than the Hawks, the home-ground advantage means the Lions have a better chance of beating the Blues than the Hawks do.

In my system, home-ground advantage is calculated based on the distance each team has to travel from their home city to the city of the game, as well as each team's experience at the venue in the past 3 seasons.

I explain my system for rating teams and predicting results here: https://thearcfooty.com/2016/12/29/intr ... gs-system/

Cheers.

Actually, my system doesn't use any data on player lists at all. It's just based on teams' past results. That means:

1) My system doesn't know about off season changes (retirements, trades, etc.) that could affect your team's performance this year; and
2) My system doesn't know about injuries that could affect your team's performance through the year.

The first problem isn't as big as you'd think. After a few weeks' worth of results, the model has a reasonably good read on where each team is at. The second issue can be a problem for week-to-week tipping, but it's far less of a problem than you might think. Individual player injuries are often less important to a team's chances than fans reckon.

The only situations in which my model struggles (temporarily) are where there are really big changes in a playing list, like Essendon* 2016 and Essendon* 2017, or to a lesser extent North Melbourne 2017.

FYI Champion Data maintains a model which is based on teams' likely best 22 players. The accuracy of their predictions is in the same ballpark as mine - and my simulations outperformed theirs last year https://thearcfooty.com/2018/02/08/how- ... -stack-up/.

Cheers
Matt

Matt's economics tweets/blog are worth a look too.

_________________
"You are being watched. The government has a secret system. A machine that spies on you every hour of every day. I know because I built it." - Finch

Noice

Out of interest, how close to the mark were your 2017 projections?

_________________
WADA medical director Dr Alan Vernec describes Essendon* FC drug case as biggest scandal in team sport the world of sport has seen. #WC2WB

#GUILTY

You arked for it.

_________________
I am as mad as hell and I'm not going to take it any more!!!

..if it does, armegeddon outta here..

_________________
..if you can't be good, be good at it..

I'll just wait to see what BigBlueWave does I guess

_________________
#DonTheStash

I'm confused

Are we playing the Tigers AFL and AFLW side, the Lions AFL and AFLW side, the Cats AFL and AFLW sides, the Bulldogs AFL and AFLW side...it's been factored into the ark I assume, 2 of each???

_________________
If I want your opinion, I'll give it to you!

& will Geoff Raines be playing?

Sent from my XT1033 using Tapatalk

_________________
#DonTheStash