Analysis Effect of rules changes on scores round 1 - 10

[Foley#41]

Anyone got the latest average score totals for this year?
666 doing wonders.

Good question. Yes I have, for the first 10 rounds.

There have been 90 games so far this year for a total of 180 scores.
To the end of round 10 last year there were 89 games, as Suns & Power had a bye in round 10 to allow time to clear customs in China.

The average score for all teams to round 10 this year has been 80.4 points.
The average score to round 10 last year was 83.1 points.

Now this might seem like a drop in scores, but a simple statistical analysis (2-tailed Student t-test for those who care) shows that these two numbers are not statistically different (p=0.28*). That is to say, statistically that difference in numbers could be just due to random variation and do not represent a meaningful change in either direction.
Conclusion: average score for all games is no different to last year. Not lower, but not higher either.

WTF you might say. Surely scoring has been lower this year.
Well not on average it hasn't. However....
If you look at the average score for winning teams there has been a drop.
2018 average winning score: 99.8
2019 average winning score: 93.9
That does represent a significant fall (p=0.040), so the average score for winning teams has fallen.

What about average score for losing teams? Well, that hasn't changed at all:
2018 average losing score: 66.3
2019 average losing score: 66.9
Not surprisingly, these are statically the same (p=0.83). Average losing scores are unchanged.

Now, since the winning score has fallen but the losing score is the same, that must mean the games are closer right?
Well actually it does. The average margin has fallen so far this year:
2018 average margin: 33.4
2019 average margin: 27.0
These are (just) statically significant (p=0.049), meaning that the margin this season is really lower than last season.
Want more proof that games are closer? Well, the number of games decided by less than two goals and less than one goal has also risen:
2018 margin < 12 points: 15 games —- 2018 margin < 6 points: 7 games
2019 margin < 12 points: 20 games —- 2019 margin < 6 points: 15 games

In summary:
For the first 10 rounds of the season, compared to the first 10 rounds of 2018, the average score has not changed but the games are closer.


Certainly the new rules have not increased scoring to this stage, but they probably haven't wrecked the game in the way that some people, including me, thought they would.
If you have any questions, or would like any further analysis, I'd be happy to have a look (within the limits of my basic statistical abilities).

*I've included the p values for those who are familiar with statistical analysis. Essentially, if the p value is < 0.05 it means that there is less than 1 in 20 chance that the difference in numbers is due to random variation, a commonly accepted cutoff for a real (rather than a random) difference. Please forgive the double-post, but I'm going to pop this on the MB as well since it took some time to calculate!

 

[Michaels]

No disrespect but when you say that the average score from last year to this year is no different, that is factually incorrect.

I've only got stats to round 8 but up till then only twice has the average score for the round of 2019 been better than 2018

Every other round its lower.

Further to that not once has the average round score been above the 9 season average round by round.

 

[Foley#41]

No disrespect but when you say that the average score from last year to this year is no different, that is factually incorrect.

I've only got stats to round 8 but up till then only twice has the average score for the round of 2019 been better than 2018

Every other round its lower.

Further to that not once has the average round score been above the 9 season average round by round.

View attachment 681212

Disrespect away! That's the great joy of statistics, is that you can use them different ways to make different points.
Tony asked about the average score for 2019. It's no different to the average score for 2018. That just means that if you average out the differences between rounds that you have indicated, the result is too close to be meaningful.

 

[Michaels]

Disrespect away! That's the great joy of statistics, is that you can use them different ways to make different points.
Tony asked about the average score for 2019. It's no different to the average score for 2018. That just means that if you average out the differences between rounds that you have indicated, the result is too close to be meaningful.

Fair play!

For me the bottom line is, the AFL heavily promoted that we would be absolutely seeing higher scoring games as a way to justify their rule changes.

And so far it's not delivering. That's not to say we aren't getting close games, but they had fed us bullshit with no accountability yet again.

 

[Foley#41]

I couldn't agree more.
Since you raise the question of total score for the games, which I guess the AFL would equate to 'entertainment value for a neutral observer':
2018 average total score for both sides: 166.1 points
2019 average total score for both sides: 160.8 points
Statistically, and I am sorry to keep using the term, unchanged (p=0.20) meaning the new rules have not increased scoring but it is not lower than last year either.

 

[Shazza_]

#gilandstevearewankas

 

[TJASTA]

 

[Mr Unbelievable]

Good question. Yes I have, for the first 10 rounds.

There have been 90 games so far this year for a total of 180 scores.
To the end of round 10 last year there were 89 games, as Suns & Power had a bye in round 10 to allow time to clear customs in China.

The average score for all teams to round 10 this year has been 80.4 points.
The average score to round 10 last year was 83.1 points.

Now this might seem like a drop in scores, but a simple statistical analysis (2-tailed Student t-test for those who care) shows that these two numbers are not statistically different (p=0.28*). That is to say, statistically that difference in numbers could be just due to random variation and do not represent a meaningful change in either direction.
Conclusion: average score for all games is no different to last year. Not lower, but not higher either.

WTF you might say. Surely scoring has been lower this year.
Well not on average it hasn't. However....
If you look at the average score for winning teams there has been a drop.
2018 average winning score: 99.8
2019 average winning score: 93.9
That does represent a significant fall (p=0.040), so the average score for winning teams has fallen.

What about average score for losing teams? Well, that hasn't changed at all:
2018 average losing score: 66.3
2019 average losing score: 66.9
Not surprisingly, these are statically the same (p=0.83). Average losing scores are unchanged.

Now, since the winning score has fallen but the losing score is the same, that must mean the games are closer right?
Well actually it does. The average margin has fallen so far this year:
2018 average margin: 33.4
2019 average margin: 27.0
These are (just) statically significant (p=0.049), meaning that the margin this season is really lower than last season.
Want more proof that games are closer? Well, the number of games decided by less than two goals and less than one goal has also risen:
2018 margin < 12 points: 15 games 2018 margin < 6 points: 7 games
2019 margin < 12 points: 20 games 2019 margin < 6 points: 15 games

In summary:
For the first 10 rounds of the season, compared to the first 10 rounds of 2018, the average score has not changed but the games are closer.


Certainly the new rules have not increased scoring to this stage, but they probably haven't wrecked the game in the way that some people, including me, thought they would.
If you have any questions, or would like any further analysis, I'd be happy to have a look (within the limits of my basic statistical abilities).

*I've included the p values for those who are familiar with statistical analysis. Essentially, if the p value is < 0.05 it means that there is less than 1 in 20 chance that the difference in numbers is due to random variation, a commonly accepted cutoff for a real (rather than a random) difference. Please forgive the double-post, but I'm going to pop this on the MB as well since it took some time to calculate!

Excellent work

 

[AlorraVex]

You mean to tell me that these new rules have arguably improved the game?

Despite the evidence suggesting such, I refuse to believe it for no reason other than my personal agenda of #gilandstevearewankas

 

[Cdog71]

Can it be as simple as there has been less blow outs than last year so it seems the margins are closer?

 

[JAKLAUGHING]

Less score blowouts!?!
Is that due to the New Rules or the way Free Kicks are awarded!?!

 

[Dustin Credible]

Less score blowouts!?!
Is that due to the New Rules or the way Free Kicks are awarded!?!

It may just be me.but when watching neutral games it sometimes feels like the umpires adjudicate with one eye on keeping the game close/a contest.

On POT-LX1 using BigFooty.com mobile app

 

[joelsyt]

It may just be me.but when watching neutral games it sometimes feels like the umpires adjudicate with one eye on keeping the game close/a contest.

On POT-LX1 using BigFooty.com mobile app

This is a lot of the reason I’ve lost interest.

I don’t care what hidden agenda Gil and Steve have given umpires just bloody adjudicate the game properly! Who cares if teams like Carlton get smashed on a weekly basis just adjudicate properly instead of trying to keep the game close! It’s frustrating and a lot of the reason Umpires are getting so much abuse this year.

 

[JAKLAUGHING]

This is a lot of the reason I’ve lost interest.

I don’t care what hidden agenda Gil and Steve have given umpires just bloody adjudicate the game properly! Who cares if teams like Carlton get smashed on a weekly basis just adjudicate properly instead of trying to keep the game close! It’s frustrating and a lot of the reason Umpires are getting so much abuse this year.

If the AFL allowed blowouts i reckon the TV rights would suffer...fancy watching cartooners being belted on Friday night TV by neutral fans?!?
Personally i could handle them being Fitzroy smashed by a 100 points +...
But the AFL won't allow it cos they can't handle the truth...

 

[JR82]

No disrespect but when you say that the average score from last year to this year is no different, that is factually incorrect.

I've only got stats to round 8 but up till then only twice has the average score for the round of 2019 been better than 2018

Every other round its lower.

Further to that not once has the average round score been above the 9 season average round by round.

View attachment 681212

Only got a fleeting mention on the "means test" this morning, rd 10 was the lowest for the yr, Whateley saying scoring is in a "trough". not sure what he means as a "trough" means that scoring has been up at some stage when it clearly hasn't....

 

[SavDog23]

Geelong are doing well so it doesn't matter what the scores are.

 

[RunningBounce]

Good question. Yes I have, for the first 10 rounds.

There have been 90 games so far this year for a total of 180 scores.
To the end of round 10 last year there were 89 games, as Suns & Power had a bye in round 10 to allow time to clear customs in China.

The average score for all teams to round 10 this year has been 80.4 points.
The average score to round 10 last year was 83.1 points.

Now this might seem like a drop in scores, but a simple statistical analysis (2-tailed Student t-test for those who care) shows that these two numbers are not statistically different (p=0.28*). That is to say, statistically that difference in numbers could be just due to random variation and do not represent a meaningful change in either direction.
Conclusion: average score for all games is no different to last year. Not lower, but not higher either.

WTF you might say. Surely scoring has been lower this year.
Well not on average it hasn't. However....
If you look at the average score for winning teams there has been a drop.
2018 average winning score: 99.8
2019 average winning score: 93.9
That does represent a significant fall (p=0.040), so the average score for winning teams has fallen.

What about average score for losing teams? Well, that hasn't changed at all:
2018 average losing score: 66.3
2019 average losing score: 66.9
Not surprisingly, these are statically the same (p=0.83). Average losing scores are unchanged.

Now, since the winning score has fallen but the losing score is the same, that must mean the games are closer right?
Well actually it does. The average margin has fallen so far this year:
2018 average margin: 33.4
2019 average margin: 27.0
These are (just) statically significant (p=0.049), meaning that the margin this season is really lower than last season.
Want more proof that games are closer? Well, the number of games decided by less than two goals and less than one goal has also risen:
2018 margin < 12 points: 15 games —- 2018 margin < 6 points: 7 games
2019 margin < 12 points: 20 games —- 2019 margin < 6 points: 15 games

In summary:
For the first 10 rounds of the season, compared to the first 10 rounds of 2018, the average score has not changed but the games are closer.


Certainly the new rules have not increased scoring to this stage, but they probably haven't wrecked the game in the way that some people, including me, thought they would.
If you have any questions, or would like any further analysis, I'd be happy to have a look (within the limits of my basic statistical abilities).

*I've included the p values for those who are familiar with statistical analysis. Essentially, if the p value is < 0.05 it means that there is less than 1 in 20 chance that the difference in numbers is due to random variation, a commonly accepted cutoff for a real (rather than a random) difference. Please forgive the double-post, but I'm going to pop this on the MB as well since it took some time to calculate!

Good analysis. Thanks.

I was no fan of the new rules either, so there's a certain amount of smug but grim satisfaction in seeing them have unintended effects.

And for mine, the game is wrecked. These rule changes on their own didn't do it, but they all add up to a situation where 1) no-one understands the rules, including media fatheads and 2) the rules are different at the elite level to suburban, junior and women's competitions.

One nerdly point, does the difference in average scores become significant if we use the one-tail T-test? I think we are testing whether scores have decreased, not just if they are different, making the one tailed version valid.

 

[RunningBounce]

No disrespect but when you say that the average score from last year to this year is no different, that is factually incorrect.

I've only got stats to round 8 but up till then only twice has the average score for the round of 2019 been better than 2018

Every other round its lower.

Further to that not once has the average round score been above the 9 season average round by round.

View attachment 681212

There's another type of statistical test that might be valid for checking the differences between the round by round average. If I get some time, I'll run that ruler over these numbers.

 

[JAKLAUGHING]

So what's the breakdown for individual teams compared to last year's totals?!?

 

[Foley#41]

So what's the breakdown for individual teams compared to last year's totals?!?

Well, Jak I'm glad you asked...

The average score change ladder below shows the average score for 2018, 2019 (up to round 10) and % change so you can see who's scoring more and who's scoring less in order. Of course, you will have to draw your own conclusions about the reasons for the change, whether it be the new rules or change in form and fortunes:

 

[JAKLAUGHING]

Have Geelong benefitted particularly from the rule changes, or are they just having a completely coincidental run of good form?

Well i am glad you finally asked that question Foley#41...and?!?!!?

 

[Foley#41]

Well i am glad you finally asked that question Foley#41...and?!?!!?

Sorry mate, I went and did what you asked after all - see edited post above.

 

[JAKLAUGHING]

Oh yeah, just another senior moment! ta...

 

[fleabitten]

I couldn't agree more.
Since you raise the question of total score for the games, which I guess the AFL would equate to 'entertainment value for a neutral observer':
2018 average total score for both sides: 166.1 points
2019 average total score for both sides: 160.8 points
Statistically, and I am sorry to keep using the term, unchanged (p=0.20) meaning the new rules have not increased scoring but it is not lower than last year either.

Well, it is lower. Just not by a statistically significant amount.

 

[Mr Unbelievable]

Fair play!

For me the bottom line is, the AFL heavily promoted that we would be absolutely seeing higher scoring games as a way to justify their rule changes.

And so far it's not delivering. That's not to say we aren't getting close games, but they had fed us ******** with no accountability yet again.

The AFL spin is full of bullshit ... but you already knew that