Is it time for a new finals system?

[Powerstufff]

I reckon Chiz has put good work into the post that opened this thread but I'm happy with the current system. The top 4 earn their spot, and their double-chance forward. 4 more teams get an elephant stamp. And using this season as a typical example even with 3 weeks to go for the H&A we have several teams outside the 8 that still have hope.

.... If a team is good enough they will eventually do it .....

That's a good summary.

 

[Chiz]

Sounds all sparkly and happy-making in theory, but the fact is if you want to change something you have do the convincing. Those you're trying to convince have absolutely no obligation to respond "reasonably" if you haven't made enough of an impact to change their minds.

I argued my case with statistical evidence. Your response was 'you can use statistics to make any point' and 'you're only arguing this as your team is going to finish 5-8'. I understand I'm never going to convince everybody, but your contribution to this discussion has been zero.

When the Mormons come to my front door, I don't bother giving reasoned arguments as to why they should depart the premises forthwith.
They, on the other hand, would probably very disappointed that so few are willing to give them the time of day, as they had all manner of reasonable arguments to present.

Sounds like a straw man argument to me.

Now, if you can come up with a reasoned solution to the more pressing problem of interstate travel, I might be more amenable.

Whatever this problem is, it certainly hasn't seemed to result in interstate teams winning any less than Victorian teams since the AFL was formed. 11 out of 23 premierships (48%) when on average only 30% of teams have been non-Victorian (but statistics can be used to make any point...).

 

[Chiz]

I reckon Chiz has put good work into the post that opened this thread but I'm happy with the current system. The top 4 earn their spot, and their double-chance forward. 4 more teams get an elephant stamp. And using this season as a typical example even with 3 weeks to go for the H&A we have several teams outside the 8 that still have hope.

I think as long as most fans are happy with the current system the AFL will not look to change it. However, I also think that if results continue to go like they have in recent years (i.e. no upsets in week 2 and 3) then more and more fans will become discontented.

 

[Dan26]

Here's an idea for a final-7 THIS year.

Suppose Essendon lose all our remaining games (which we will) and we finish 8th with a 13-9 record.

And we get stripped of points.

That means 5th would have normally played us, right?

Well, 5th could have a bye. It would mean they have a mathematically better chance than 6th, and 7th, but not as good as 1,2,3,4.

The current probabilities for the top-8 are:

1 - 18.75%
2 - 18.75%
3 - 18.75%
4 - 18.75%
5 - 6.25%
6 - 6.25%
7 - 6.25%
8 - 6.25%

But, under the situation I have listed, the probabilities are:

1 - 18.75%
2 - 18.75%
3 - 18.75%
4 - 18.75%
5 - 12.5%
6 - 6.25%
7 - 6.25%

It actually makes the finals (which would be a final-7) fairer.

 

[Chiz]

Here's an idea for a final-7 THIS year.

Suppose Essendon lose all our remaining games (which we will) and we finish 8th with a 13-9 record.

And we get stripped of points.

That means 5th would have normally played us, right?

Well, 5th could have a bye. It would mean they have a mathematically better chance than 6th, and 7th, but not as good as 1,2,3,4.

The current probabilities for the top-8 are:

1 - 18.75%
2 - 18.75%
3 - 18.75%
4 - 18.75%
5 - 6.25%
6 - 6.25%
7 - 6.25%
8 - 6.25%

But, under the situation I have listed, the probabilities are:

1 - 18.75%
2 - 18.75%
3 - 18.75%
4 - 18.75%
5 - 12.5%
6 - 6.25%
7 - 6.25%

It actually makes the finals (which would be a final-7) fairer.

An interesting suggestion. My issue with it is that although 5th has a lower probability of winning than 4th, in week 2 they have an unquantifiable advantage (coming off the bye) over their opponent who will be seeded higher.

 

[Proud SJW]

I argued my case with statistical evidence. Your response was 'you can use statistics to make any point' and 'you're only arguing this as your team is going to finish 5-8'. I understand I'm never going to convince everybody, but your contribution to this discussion has been zero.

You're on a football forum. Expect orneriness.
Ad-hominem passes for enlightened discussion around here. Nod nod, wink wink.

Sounds like a straw man argument to me.

Perhaps you should revisit your logical fallacies dictionary then.
A straw man would be a situation in which I tried to represent your argument as something it was not.
The example above is drawing a parallel. You've already admitted you were you weren't going to reach many when you started out, are you reversing that thought now?

Whatever this problem is, it certainly hasn't seemed to result in interstate teams winning any less than Victorian teams since the AFL was formed. 11 out of 23 premierships (48%) when on average only 30% of teams have been non-Victorian (but statistics can be used to make any point...).

Ever consider that there might be other factors contributing to that?
(In other words: Yes, you can).

 

[Tigerman20]

2. To win from 1-4, a team must defeat at least two other top four teams.

Why do you have to defeat 2 top 4 sides to win the premiership?

I guess you meant the likely scenario taking into account the historical evidence as I see from this thread, don't mean to be a smartarse.

But factually

You finish 1st, you win the first match thats one top four side. After that you could play 5th or 8th, then In GF play 6th or 7th.

 

[Tigerman20]

edit above,

I think its actually 6th or 7th then 5th or 8th but you get the idea

 

[melanChronic]

Best one I can think of is a final 7.
Week 1: 2v7, 3v6, 4v5 (1st is rewarded with a week off)
Week 2: 1v4, 2v3 (predicted results)
Week 3: 1v2

39% of teams make the finals, every game is knockout, every position has significant/reasonable advantage over the one below it.

 

[TheFVK]

Well considered OP, but this:

Finding it hard to see how you could come up with a system that retains those three positive points but eliminates the negatives. The current system is most likely the best you're going to get.

 

[Chiz]

Ever consider that there might be other factors contributing to that?
(In other words: Yes, you can).

Of course there are other factors. If these other factors are randomly occurring, then over time we'd expect the results to balance out. However, if the other factors give a net advantage to non-Victorian sides (as the results suggest), then you can hardly cry foul on the problem of interstate travel.

Why do you have to defeat 2 top 4 sides to win the premiership?

OK, technically you could get lucky, but the odds of that happening are very low (5th has more chance of winning the GF!).

Well considered OP, but this:

Finding it hard to see how you could come up with a system that retains those three positive points but eliminates the negatives. The current system is most likely the best you're going to get.

In my view, this one is better:

Best one I can think of is a final 7.
Week 1: 2v7, 3v6, 4v5 (1st is rewarded with a week off)
Week 2: 1v4, 2v3 (predicted results)
Week 3: 1v2

39% of teams make the finals, every game is knockout, every position has significant/reasonable advantage over the one below it.

Big advantage to finishing top (possibly too big in that it may make the 1v4 final a bit one-sided), and while 2-7 have the same mathematical probability of winning, it is much harder to win from 7th, given that you'd have to defeat 2nd, 1st (both away), then probably 3rd in the GF. In contrast, 2nd plays 7th, 3rd or below (both at home), then probably 1st in the GF.

However, I still think the knock-out final 8 with wildcard playoffs for the final two spots is the best system.

 

[Powerstufff]

...I still think the knock-out final 8 with wildcard playoffs for the final two spots is the best system.

Wildcard? Surely not from the whole 12 teams not already in the top 6?

 

[Chiz]

Wildcard? Surely not from the whole 12 teams not already in the top 6?

No, it'd be something like 7v10, 8v9 (there are systems that include more wildcard games, but I think this would be plenty). Basically if you make the top 6 you've got a shot at the flag (obviously it becomes harder the further down you are), and if you finish 7-10, you have a chance to play a final (but winning the premiership would be almost impossible)

 

[Pie eyed]

Why isn't there a poll with a simple NO option?

 

[Chiz]

For interests sake, just to reiterate some points, imagine that there is no difference in ability between any team in the finals (e.g. in a neutral match, 8 has a 50% chance of beating 1).
If we assume that a home ground advantage increases the probability of the home team winning by 10%, and that following up from an elimination final decreases the probability of winning by 10% (i.e. a further advantage in weeks 2 and 3), assuming no advantage to either side in the GF, then the probability of each team winning the GF looks like this:



Again, the only advantage 1 and 2 have over 3 and 4 is a home ground in week 1 - as can be seen, winning from 4th is only slightly more difficult than winning from 1st. But winning from 5th or below is almost impossible (and remember, this assumes that each team is equal). If you assume some difference in the quality of each team (e.g. 1st is slightly better than 2nd, 2nd is slightly better than 3rd and so on), then this skews the distribution more in favour of the top teams (although it doesn't advantage 1st over 2nd due to the preliminary final - if having a harder prelim final was factored in, then 2nd might have even more chance of winning than 1st).

In contrast, the knockout final 8 system, factoring in that 7-10 have to win a wildcard match to play in the finals, distributes the probabilities like this (same assumptions as above - i.e. all teams are just as good as each other):



The probability of winning from 1st or 2nd is slightly reduced, but this system compensates for that by giving 1 and 2 an easier semi final. The chances of winning from 7-10 are slightly higher than winning from 5-8 under the current system, but in reality it is probably more difficult when you factor in the likely opponents.

The key difference in the distribution is that some of 3rd and 4th's advantage is given to 5th and 6th, but not to the point at which 5th and 6th are advantaged over 3rd and 4th. For each set of two ladder spots there is a clear advantage, but there is no position at which this advantage drops anywhere near as much as it does from 4th to 5th under the current system (the biggest drop is 6th to 7th).

 

[Father Jack]

With the compromised fixture the fairest way to do it imo would be to separate the competition into two halves, so that each team is vying against teams with the same or similar fixture for finals positions. You could then have two 'final four' style systems running concurrently, with the respective winners playing off for the premiership. The two 'conferences' or whatever you want to call them don't have to be the same year after year if you're worried about two particular teams never being able to play off in the Grand Final.

 

[Proud SJW]

Of course there are other factors. If these other factors are randomly occurring, then over time we'd expect the results to balance out. However, if the other factors give a net advantage to non-Victorian sides (as the results suggest), then you can hardly cry foul on the problem of interstate travel.

Of course I can. If one unbalancing issue is seemingly balanced by another, it does not negate either being an issue to begin to begin with. Given the sample size of seasons and teams played so far, you still can't rule out good recruiting and pure luck.

I was, however, using it as an example that you can use statistics to prove any point. Your reply was that seeing as interstate teams seem to appear in finals as much as Victorian teams do, this must indicate that travel is not an issue in the AFL.
Clearly, that's an incorrect conclusion.

 

[Dan26]

That's pretty much the system I suggested during the thread. The only difference was I said had 7v10 and 8v9 as wildcard entry games to get into the knockout final 8 (but the effect is the same).

The finals system you suggested has double chances.

Double chances need to be eradicated. They are a cancer on the finals.

While I'm a fan of finals being about performing on the day, I'm not completely opposed to double chances, in and of themselves. But it's the double chances that are the primary cause of all the issues with the current system (and previous systems) - we can still play plenty of finals without double chances, so I'm totally in favour of getting rid of them.

I've explained before why double chances are not needed. And I can explain it mathematically.

Under the current system, the top-4 don't receive a guaranteed double chance. They receive a 50% chance of using a double chance, and a 50% chance of having a week off. ONE OR THE OTHER.

All you need to do, is take that 50% chance of "using a double chance" and add it to the 50% chance of having a week off and you get a 100% chance of having a week off. This, therefore replaces the double chance mathematically.

The NFL do this. The top 2 teams in each conference have a 100% guaranteed week off, but NO double chance.

What we do in the AFL, is have a 50% chance of having a week off and a 50% chance of using a double chance. I don't like this, as it allows teams to 'not perform on the day" yet still win the premiership. The 100% guaranteed week off replaces the double chance. My final-10 system solves that problem.

The PF final results since 2000 have shown that playing after a bye (against a team coming off a must-win final) is a huge advantage. Thus, 1 and 2 are clearly advantaged in this system, which I like.

Under the final-10 the advantages are as follows:

1st and 2nd
Get a week off, and get to play the two winners from the 7v10 and 8v9 matches. So a rested team plays a low-seeded team who played the week before

3rd and 4th
have a home game against y6th and 5th respectively. Have a week off, but the week off isn't an advantage (like it is for 1st and 2nd) because their opponent has the week off also.

5th and 6th
Have an away game against 4th and 3rd respectively. Have a week off, but the week off isn't an advantage because their opponent has the week off also

7th and 8th
Have a home game versus 10th and 9th respectively. Then are away from hoem for the rest of the finals.

9th and 10th
away from home for the entire finals series.


The top-4 all have a home game to start the finals (unlike the current system) and there is a clear advantage in finishing 1st or 2nd. The prospect of playing a team from 7,8,9,10 who PLAYED in week one is clearly better than finishing 3rd and playing 6th who also had a week off.

 

[JoondalupJ]

I've been well aware of these stats for the past few years, though each year there are no upsets in weeks 2 and 3 (five years in a row so far), the figures look worse and worse.



It's never been about making it 'easier' for lower teams to win finals, it's about distributing the benefits across finals teams so that it becomes a bit more difficult for each position down the ladder (currently it's probably easier to win the GF from 2nd than 1st, which is ridiculous) - e.g. it should be a lot harder to win from 5th than from 1st, but it shouldn't be so much harder to win from 5th than from 4th.




Certainly a sample of 13 would be too small, but 52 games have been played in weeks 2 and 3, which is definitely a large enough sample to draw some conclusions about the likelihood of upsets. Incidentally, there has also been 52 games played in week 1 (in which there are also home ground advantages). I think it's fair to say that most of the difference between the likelihood of an upset in week 1 compared to weeks 2 and 3 can be attributed to the additional benefits the home team gets in these games. Upsets have occurred in about 30% of week 1 games, but in only 10% of weeks 2 and 3 games.

To put it a different way, over six rounds of footy tipping, what are the odds that you would only get five tips wrong? It would be pretty unlikely, in my opinion, as upsets tend to occur regularly (for instance, in my best six consecutive six rounds this season I still got 10 wrong, and Melbourne and GWS played in 12 of those games!). We might expect upsets to occur even more regularly in finals, since the teams are more evenly balanced on the whole. But in weeks 2 and 3, they don't...



Well, I don't think anyone has proposed a roulette system. If we want the best prelim finals, the important thing is that the better teams win the semi-finals, not the higher placed teams (on average, the higher placed teams are the better teams, but not always).



Despite having had some good prelims and semis, I don't think you can deny that if the advantage to the home side was reduced, we'd have even more good semis and prelims.



It's about balance. Conversely, if every grand final was won by the top team, we would see that as a negative on the system. Over a long period of time, say 100 years, we'd want maybe 5-10 grand finals to have been won by teams 5-8 (i.e. it needs to be an unlikely, but not unrealistic outcome).




I think a lot has to go right for any team to win the premiership, but to win from 5th to 8th, you'd probably still have to be one of the best two teams (i.e. if you are the 5th best team and you finish 5th, you are not going to win, but if you're the best team and you finish 5th, you have a chance, albeit small). So if/when it finally happens, I don't think we'll be calling it a fluke (miracle is fair enough), it'll just be a case of a team finishing below where they belong, and doing it the hard way in the finals.



You're probably right, a lot will have to go right for Freo to win, but given they are in the top four they are in with a chance - if they were 5th, it'd be almost impossible.



Which is probably the main reason I'm opposed to the idea of reducing the number of teams in the finals, even though the flag is only ever won from the top 4.




I've responded to a similar point already. While 4th has a hard game first up, if they do manage to win, they suddenly have the easiest path to the grand final (easier than the winner of 2v3, since they will probably have to beat 1st in the prelim). As for the 'difficult' semi final against 5th if they lose, the fans continue to fall for this every year - 4th has never lost to 5th in a semi final, even though 5th is often billed as the favourite (in which case, head to the bookies).




True, that is a positive, and I'm quite happy to have a reasoned debate with anyone who gives a reasoned argument. Of course, there are always a few who don't...

Yep your right . Keep the eight , now we have 18 teams .

 

[seysearles]

Is it time for a new finals system?

 

[Chiz]

Of course I can. If one unbalancing issue is seemingly balanced by another, it does not negate either being an issue to begin to begin with. Given the sample size of seasons and teams played so far, you still can't rule out good recruiting and pure luck.

My point is that on balance, there is no evidence that non-Victorian teams face a net disadvantage.

I was, however, using it as an example that you can use statistics to prove any point.

The statistics I quoted in regards to weeks 2 and 3 of the finals are pretty damning evidence that these games are heavily one-sided. If you think they are wrong you need to explain why - simply saying that statistics can be used to prove anything isn't a sufficient argument.

Your reply was that seeing as interstate teams seem to appear in finals as much as Victorian teams do, this must indicate that travel is not an issue in the AFL.
Clearly, that's an incorrect conclusion.

No, my conclusion was that if travel for non-Vic teams was such a big issue, then we'd expect to see this reflected in the results - the reality is that non-Vic teams have far outperformed Vic teams.

 

[Chiz]

The finals system you suggested has double chances.

Go back and read it again - the system I've proposed is, for all sakes and purposes, the same as yours. The only difference is that I would brand the 7v10 and 8v9 games as 'wildcard playoffs' rather than finals.

Double chances need to be eradicated. They are a cancer on the finals.

My point was more that I don't completely object to double chances on principle, like you do, but I object to them because they are the primary cause of most of the issues I outlined in the OP (e.g. they are the reason that 2nd gets an easier prelim than 1st, 4th has it 5 times better than 5th, etc.). That alone should be enough to eradicate them from finals.

I've explained before why double chances are not needed. And I can explain it mathematically.

Under the current system, the top-4 don't receive a guaranteed double chance. They receive a 50% chance of using a double chance, and a 50% chance of having a week off. ONE OR THE OTHER.

All you need to do, is take that 50% chance of "using a double chance" and add it to the 50% chance of having a week off and you get a 100% chance of having a week off. This, therefore replaces the double chance mathematically.

The NFL do this. The top 2 teams in each conference have a 100% guaranteed week off, but NO double chance.

What we do in the AFL, is have a 50% chance of having a week off and a 50% chance of using a double chance. I don't like this, as it allows teams to 'not perform on the day" yet still win the premiership. The 100% guaranteed week off replaces the double chance. My final-10 system solves that problem.

Technically it does reduce the probability of winning the GF, but a top team still needs to win three consecutive finals to win the premiership, which is the same as now - the loss of the double chance is compensated for with an easier final first up. In any case, I dislike the idea that a team can lose a final yet still win the premiership.

Under the final-10 the advantages are as follows:

1st and 2nd
Get a week off, and get to play the two winners from the 7v10 and 8v9 matches. So a rested team plays a low-seeded team who played the week before

3rd and 4th
have a home game against y6th and 5th respectively. Have a week off, but the week off isn't an advantage (like it is for 1st and 2nd) because their opponent has the week off also.

5th and 6th
Have an away game against 4th and 3rd respectively. Have a week off, but the week off isn't an advantage because their opponent has the week off also

7th and 8th
Have a home game versus 10th and 9th respectively. Then are away from hoem for the rest of the finals.

9th and 10th
away from home for the entire finals series.

The top-4 all have a home game to start the finals (unlike the current system) and there is a clear advantage in finishing 1st or 2nd. The prospect of playing a team from 7,8,9,10 who PLAYED in week one is clearly better than finishing 3rd and playing 6th who also had a week off.

This gives a much better spread of benefits than the current system. 1st and 2nd have more of an advantage over 3rd and 4th than present, but 3rd and 4th, while still maintaining an advantage over 5th and 6th, have less of an advantage. What it means is that instead of having 4 teams who can realistically win the premiership, we have 6 teams (one third of the league - this seems reasonable to me), even though winning from 5th/6th would still be very difficult.

Have a look at my post above (#190), in which I graphically compare the distribution of benefits in the current system to this one - there's no comparison as to which is the better system, and that's before you even consider some of the other issues, like the second seed having the easier prelim.

 

[Spider Eater]

The finals system you suggested has double chances.

Double chances need to be eradicated. They are a cancer on the finals.

Under the final-10 the advantages are as follows:

1st and 2nd
Get a week off, and get to play the two winners from the 7v10 and 8v9 matches. So a rested team plays a low-seeded team who played the week before

3rd and 4th
have a home game against y6th and 5th respectively. Have a week off, but the week off isn't an advantage (like it is for 1st and 2nd) because their opponent has the week off also.

5th and 6th
Have an away game against 4th and 3rd respectively. Have a week off, but the week off isn't an advantage because their opponent has the week off also

7th and 8th
Have a home game versus 10th and 9th respectively. Then are away from hoem for the rest of the finals.

9th and 10th
away from home for the entire finals series.

The top-4 all have a home game to start the finals (unlike the current system) and there is a clear advantage in finishing 1st or 2nd. The prospect of playing a team from 7,8,9,10 who PLAYED in week one is clearly better than finishing 3rd and playing 6th who also had a week off.

I'm against giving teams outside the top 50% the chance to win the premiership but in trying to understand your system and liking the NFL playoff system myself I made this graph... everything looks better in graph form!
And I actually like your system Dan... same number of finals and weeks also.

 

[santa claws]

it sure is time.
time to stop rewarding mediocrity and go back to a final 5. there is no fairer system

 

[bluespooner]

Should have no finals and just award the premiership to the side that finishes top. The top side should then go on to be awarded the first pick in the national draft.