Play Nice 2018 Ladder Predictions

[Donakebab]

Seems reasonable.

View attachment 464371

Their predictions were pretty s**t last year too.

 

[hk89]

Their predictions were pretty s**t last year too.

Well they got 6 of the top 8 right last year. They didn't predict Essendon or Richmond to play finals, which was probably fair enough at the time, and their model might not have factored in Essendon's returning drug suspensions (they had Bombers second last which seems harsh given who they had coming back into the side). While they had Hawthorn in 7th spot, they did give them a nearly 40% chance of not making finals which was a good read of their end of season form by their prediction model. Dogs was the other team they got wrong in their top 8, and I'm not sure many people would have predicted they'd drop out given how good their end of season form was. I'd say they were probably more accurate last year than most of the predictions you'll see in this thread will pan out to be.

They seem to have Melbourne a little too low this year, but perhaps they've got a 'mentally fragile' element factored in. Essendon seem perhaps a little low, but a harder draw could easily see them drop a few places. I'll be surprised if Richmond can finish top, but certainly not out of the question given their late season form. The rest of it looks pretty accurate based on form towards the end of 2017.

 

[Jatz]

Without reading through this huge thread, what the general consensus of the Dees playing finals this year?

I'm bullish on their chances but want to see if that's just me or whether there's a few that would expect them to make the 8?

 

[JRyan23]

1. Sydney
2. Adelaide
3. Port Adelaide
4. Geelong
5. Melbourne
6. GWS
7. Richmond
8. Essendon

9. Western Bulldogs
10. St Kilda
11. Brisbane
12. Gold Coast
13. Collingwood
14. Hawthorn
15. West Coast
16. Fremantle
17. North
18 Carlton

 

[Gene Simmons]

Top 4 (not in order):
GWS, Adelaide, Melbourne, PA

Bottom 4 (not in order):
GC, Freo, Lions, NM

Literally all other 10 teams in between can finish anywhere from 5th-14th

 

[Heisenberg_]

Without reading through this huge thread, what the general consensus of the Dees playing finals this year?

I'm bullish on their chances but want to see if that's just me or whether there's a few that would expect them to make the 8?

If they don’t make the 8 questions will be asked of Goodwin IMO

 

[Bazza97]

Without reading through this huge thread, what the general consensus of the Dees playing finals this year?

I'm bullish on their chances but want to see if that's just me or whether there's a few that would expect them to make the 8?

I think most people who have done their predictions have put Melbourne in the top 8. They've also been a pretty popular pick for top 4 from what I've seen. I reckon the general consensus on here is that Melbourne is the most likely team that finished outside the top 8 last year to make that jump into the finals this year.

 

[wolfpup]

Running super quickly through an excel spreadsheet thingy at work without really paying much attention got me this (obvious saints bias and it didn't calculate percentage). Think Essendon will probably be top 4 though have them at 13th here.

​

GWS Giants 17 68
Richmond 16 64
Adelaide 15 60
Port Adelaide 15 60
Sydney 15 60
Geelong 13 52
Melbourne 13 52
St Kilda 12 48
West Coast 11 44
Western Bulldogs 11 44
Brisbane Lions 10 40
Collingwood 10 40
Essendon 10 40
Hawthorn 10 40
Gold Coast 8 32
Fremantle 6 24
Carlton 3 12
North Melbourne 3 12​

 

[Blackhawk42]

Think people are still overrating GWS this season. There must be some mental scarring of the back to back prelim final losses. Coupled with the expectation that they should be a powerhouse winning flags right now. On top of that, Dawson Simpson looks like their number 1 ruckman and an ageing Heath Shaw will need to be the number 1 rebounding defender again with Williams out and Wilson gone. Have them finishing 9th and then regrouping again 2019.

 

[HTT]

99.8% of predictions here have bias e.g my biased opinion has the Crows winning 19 games. My unbiased opinion and a little itch in the back of my mind can see the wheels completely fall off and we struggle to make the 8. Yet middle bias says we win at least 13 games.

Your unbiased opinion says that you’ll win at least 13 games, and your negatively biased opinion says that the wheels will fall off. It’s all in the details . Unless you’re suggesting you honestly, deep down, believe you’ll miss completely.

 

[CholEstrol]

Without reading through this huge thread, what the general consensus of the Dees playing finals this year?

I'm bullish on their chances but want to see if that's just me or whether there's a few that would expect them to make the 8?

I think they will. List and age profile very good. Just need the other elements to click in injury luck, belief, discipline around the game plan and playing for each other.

 

[Roovolution]

If they don’t make the 8 questions will be asked of Goodwin IMO

Questions should be asked about the ticker of the team, not Goodwin.

 

[Roovolution]

Top 4 (not in order):
GWS, Adelaide, Melbourne, PA

Bottom 4 (not in order):
GC, Freo, Lions, NM

Literally all other 10 teams in between can finish anywhere from 5th-14th

Be quite surprised if Port are top 4.

 

[iGNITER]

Without reading through this huge thread, what the general consensus of the Dees playing finals this year?

I'm bullish on their chances but want to see if that's just me or whether there's a few that would expect them to make the 8?

9th

 

[Benno_900]

1. Adelaide
2. Richmond
3. Melbourne
4. Sydney
5. Geelong
6. Port Adelaide
7. GWS
8. Essendon

9. Hawthorn
10. Bulldogs
11. Collingwood
12. Brisbane
13. West Coast
14. Carlton
15. Fremantle
16. St Kilda
17. North Melbourne
18. Gold Coast

 

[rmcq]

Well they got 6 of the top 8 right last year. They didn't predict Essendon or Richmond to play finals, which was probably fair enough at the time, and their model might not have factored in Essendon's returning drug suspensions (they had Bombers second last which seems harsh given who they had coming back into the side). While they had Hawthorn in 7th spot, they did give them a nearly 40% chance of not making finals which was a good read of their end of season form by their prediction model. Dogs was the other team they got wrong in their top 8, and I'm not sure many people would have predicted they'd drop out given how good their end of season form was. I'd say they were probably more accurate last year than most of the predictions you'll see in this thread will pan out to be.

They seem to have Melbourne a little too low this year, but perhaps they've got a 'mentally fragile' element factored in. Essendon seem perhaps a little low, but a harder draw could easily see them drop a few places. I'll be surprised if Richmond can finish top, but certainly not out of the question given their late season form. The rest of it looks pretty accurate based on form towards the end of 2017.

A quick rundown on how Elo ratings like this work:

It’s team only, so individual players are irrelevant. Your rating changes based on the rating of the team you play against. At the start of each season, a team's rating is modified by 10 or 15% towards the mean (ie. good teams go down, bad teams go up).

So Richmond are in top because in the last 3 games they played, they thrashed other teams with high ratings. Sydney are second because they already had a high rating after spending months beating every other team, but then lost one game, the decrease in rating not enough to send their rating too far down.

There is no deeper analysis put into it, so when you have a team like Richmond in 2016 playing kids, it sends their rating through the floor, and it takes a long time to turn it around once they fixed their game plan and started playing a full strength side again.

 

[00Stinger]

Without reading through this huge thread, what the general consensus of the Dees playing finals this year?

I'm bullish on their chances but want to see if that's just me or whether there's a few that would expect them to make the 8?

Melbourne = Philly Eagles

Sent from my SM-G930F using Tapatalk

 

[hk89]

A quick rundown on how Elo ratings like this work:

It’s team only, so individual players are irrelevant. Your rating changes based on the rating of the team you play against. At the start of each season, a team's rating is modified by 10 or 15% towards the mean (ie. good teams go down, bad teams go up).

So Richmond are in top because in the last 3 games they played, they thrashed other teams with high ratings. Sydney are second because they already had a high rating after spending months beating every other team, but then lost one game, the decrease in rating not enough to send their rating too far down.

There is no deeper analysis put into it, so when you have a team like Richmond in 2016 playing kids, it sends their rating through the floor, and it takes a long time to turn it around once they fixed their game plan and started playing a full strength side again.

Yes, there are pros and cons to this approach. I think it produces outcomes that pass the 'eye test' better than systems that try to use the aggregation of the players' ratings in each team to predict outcomes because it captures how well the players operate as a team.

Elo can deal with the latter problem with a sensibly tuned "K-Factor" that controls how much single results will nudge the rating, but a poorly tuned K-Factor can either make the model too sensitive to recent results, or not sensitive enough. More modern versions of Elo-like systems like Glicko that have rating deviation components probably allow for a more elegant handling of that issue, and the Glicko RD also provides a convenient way of calculating confidence intervals on the models' stated ratings.

RD does a good job of degrading the certainty of ratings in the off season, and so helps accelerate ratings to their 'true' levels as the season progresses, but it doesn't do anything to deal with the problem that your ratings are purely team performance based and that until they start playing again, team composition changes are not well reflected (or as you point out, reflected at all). So these Elo-like systems are a pretty good indication of most recent form, and do a good job of factoring out things like draw difficulty differences, because the opponent rating is taken into account when adjusting your rating, but without having a more sophisticated approach than moving towards the mean in the off-season, it definitely pays to be aware they are not going to factor in recruits or departures until the season starts, so your point is well made.

From that point of view, it is not surprising that it gets Essendon so wrong in 2017, but somewhat surprising that it got Bulldogs so wrong. Because of the impact of your opponents rating on your own rating, Essendon's bottom team draw from last year might be part of the reason why they are starting further down the predicted ranks than their fans might like, as Elo will have been less impressed with their 2017 performance than a ranking system based on pure wins is.

Elo also explains why Hawthorn is rated 7th where a lot of people see them as bottom 8 material. 7th is basically our 18-round form ladder ranking, and Elo cares a lot more about the last 18 games than it does about the 4 we played at the start of 2017. Basically Elo produces a form ladder corrected for draw difficulty.

 

[Kane McGoodwin]

So if we take an average of ARC & Champion Data predictions...

1 Sydney (2,2) +%
2 Adelaide (3,1)
3 Richmond (1,5)
4. Port (6,3) +%
5.GWS (5,4)
6 Geelong (4,6)
7. Collingwood (9,7)
8. Melbourne (11,8) ++%
----
9. Western Bulldogs (10,9) +%
10. West Coast (7,12)
11. Hawthorn (8,13)
12. St Kilda (12,11) +%
13. Essendon (13,10)
14. Carlton (15,15)
15. Gold Coast (18,14) ++%
16. North (14,18) +%
17. Fremantle (16,16)
18. Brisbane (17,17)

 

[LeverPuller]

Last Chance Prediction - End of JLT/Start of Season Edition

1. Richmond
2. Sydney
3. GWS
4. Melbourne
5. Port
6. Geelong
7. Adelaide
8. Fremantle
-----------------------------------
9. Western Bulldogs
10. Essendon
11. Hawthorn
12. St Kilda
13. Collingwood
14. West Coast
15. Carlton
16. Gold Coast
17. Brisbane
18. North Melbourne

 

[XYZ123]

Come on mate if we beat Sydney,Brisbane,Carlton x2,Fremantle x2 would have made top 4 easily last year

If it was any other poster I'd think surely this is a piss take post.. but you're actually using this as a legitmate argument and example arent you?

Jeepers.

 

[CaptainHowdy87]

1. Adelaide
2. Richmond
3. Melbourne
4. Sydney
5. Geelong
6. Port Adelaide
7. GWS
8. Essendon

9. Hawthorn
10. Bulldogs
11. Collingwood
12. Brisbane
13. West Coast
14. Carlton
15. Fremantle
16. St Kilda
17. North Melbourne
18. Gold Coast

Was in full agreement with this until I saw St Kilda 3rd from the bottom. Why don't you rate them? You really see West Coast finishing several spots in front of them?

 

[BOMBERS 4 LIFE!!!!]

If it was any other poster I'd think surely this is a piss take post.. but you're actually using this as a legitmate argument and example arent you?

Jeepers.

The only piss take post is you mate wtf would you know you don't watch Essendon play.

 

[XYZ123]

The only piss take post is you mate wtf would you know you don't watch Essendon play.

Your response doesnt even make sense and has nothing to do with what i posted.

Of course if any team had an additional six wins instead of losses they'd finish far far higher up the ladder? Absolute bizarre reasoning.

Your fellow posters must absolutely cringe when they see a new post with your name on it.

 

[Lsta062]

Final Ladder after JLT Series:

1. Richmond
2. Sydney
3. GWS
4. Melbourne
5. Port Adelaide
6. Adelaide
7. Essendon
8. Geelong
----------------------------
9. Collingwood
10.Western Bulldogs
11. Hawthorn
12. Carlton
13. Fremantle
14. St Kilda
15. Gold Coast
16. West Coast
17. Brisbane
18. North Melbourne