The Statistics Thread

[dogwatch]

Lot's of talk about our goal accuracy this week so I thought I'd compile a summary of how we are doing across the board from shot type, distance and angle, along with a player table.

The following is our shot at goal accuracy percentage along with our competition ranking (competition average is in brackets)...

Overall
48% - 9th (48%)

Shot Type

Set Shot
56% - 6th (53%)

On The Run
46% - 6th (43%)

Snap
35% - 16th (40%)

Mark Play On
56% - 11th (56%)

Ground
25% - 15th (40%)

Distance

0m-15m
56% - 18th (77%)

15m-30m
53% - 13th (56%)

31m-40m
45% - 8th (45%)

41m-50m
45% - 5th (41%)

50m+
42% - 1st (27%)

Angle

Left Acute
8% - 17th (32%)

Left Angle
41% - 13th (44%)

Corridor
49% - 10th (52%)

Right Angle
34% - 15th (41%)

Right Acute
33% - 5th (24%)

Now below is our player table for the season with goal accuracy percentage along with expected goal accuracy percentage. For those unfamiliar with how expected score/accuracy is calculated click here to watch a simple but informative explanation.

View attachment 1160903

Bruce is ranked 1st in the competition for goal accuracy for any player with 40+ shots and is over 12% ahead of his expected average which means he is nailing difficult shots. The same can't be said for Naughton who is almost 15% below his expected accuracy, just think how good he will be when he fixes this up. Bontempelli is ahead of his expected accuracy but that's due to him taking a lot of shots on the run from 40m-50m+ which he is nailing most of the time, however his accuracy from 0m-30m is 40% which is certainly not good enough.

Hannan, McNeil and B. Smith are approximately 12-14% below their expected accuracy, while Hunter is around 4% below and Weightman and English are around their expected accuracy.

Excellent post. Thanks OG.

The conclusions I draw from that are:

• Our biggest problem seems to be a horrible rate of conversion from really easy shots (56% within 15m, against comp avge of 77%). That is consistent with the eye test.
• I don't care about players who are below conversion average if they have only had a handful of shots. Let's focus on the ones who are regularly getting the ball in scoring range. Say at least once a game on average. The main culprits here are Naughton, Hannan, Vandermeer, B Smith, Scott, McNeil and Dunkley.
• Also, not a criticism, but the table of players doesn't drill down far enough to reveal some player anomalies. The most obvious example is Bont's figures which look good because he is an excellent shot on the run and from a distance (well above comp or expected average I'd guess) but he is deplorable from easy set shots close to goal. He should be very much under the microscope for that and should not get off lightly because of his good long range goals. It may well have cost us the win in an "8-point game" last Friday. It just seems such an easily fixable problem. My guess is he has about a 95% conversion rate on those at training. English might be another I suspect, but not as glaring. And Naughton definitely, but his averages are already sub-par overall.
• The other thing that makes it hard to draw conclusions is that we only have percentages and AFL ranking not absolute numbers in each of the team-wide categories. Again let's not sweat the stuff where the total numbers are small and a 15-20% improvement might equate to only two or three more goals over the season so far. Let's look instead at where we create our greatest number of opportunities but fall well short of the competition average or expected score.

 

[Fossie 32]

I guess expected score is based on competition average which includes some innacurate kickers so we would hope / aim for our players to be better than that?
It looks like of those who have had a decent number of shots, 5 are above av, 5 about av and 7 under. Wonder how many shots we have had from outside 50

 

[threenewpadlocks]

I guess expected score is based on competition average which includes some innacurate kickers so we would hope / aim for our players to be better than that?
It looks like of those who have had a decent number of shots, 5 are above av, 5 about av and 7 under. Wonder how many shots we have had from outside 50

I'm not so sure that there should be an "expectation" that our team as a whole is better than average, but it should just be average in accuracy. There's very few teams in the 10 or so years that we've had these expected accuracy stats that you can say with statistical certainty beyond just randomness are a better than average goalkicking team: Hawthorn's premiership teams in part because their small forwards were so good at kicking goals whilst winning contested possessions and under pressure (think the best version of Puopolo and Rioli) and West Coast of the last few years mainly because Kennedy is such a reliable shot for goal.

Apart from that - of 180 instances of team seasons over 10 years - virtually none of them have kicked more or less accurately than what can simply be explained by the statistical equivalent of flipping a coin 100 times and getting heads 55 of them (ie that doesn't the coin biased to heads, just like kicking a goal 55% of the time when the league average is 50% doesn't necessarily make you more accurate).

 

[Fossie 32]

I'm not so sure that there should be an "expectation" that our team as a whole is better than average, but it should just be average in accuracy. There's very few teams in the 10 or so years that we've had these expected accuracy stats that you can say with statistical certainty beyond just randomness are a better than average goalkicking team: Hawthorn's premiership teams in part because their small forwards were so good at kicking goals whilst winning contested possessions and under pressure (think the best version of Puopolo and Rioli) and West Coast of the last few years mainly because Kennedy is such a reliable shot for goal.

Apart from that - of 180 instances of team seasons over 10 years - virtually none of them have kicked more or less accurately than what can simply be explained by the statistical equivalent of flipping a coin 100 times and getting heads 55 of them (ie that doesn't the coin biased to heads, just like kicking a goal 55% of the time when the league average is 50% doesn't necessarily make you more accurate).

We could be trendsetters then!
We try to be better than average in everything else so we can win the flag, so why not this?
It seems quite a few games are won by one team being more accurate than the other - eg Dogs v Cats last weekend
Just can't see that it wouldn't be better if we improved in this area.

 

[Oliver G]

Naughton getting some great looks, 58% EX on that amount of shots is crazy.

Would love to see how his EX numbers compare to the likes of McKay, Hawkins and Walker.

McKay - 55.4%
Hawkins - 47.7%
Walker - 45.7%

I guess expected score is based on competition average which includes some innacurate kickers so we would hope / aim for our players to be better than that?
It looks like of those who have had a decent number of shots, 5 are above av, 5 about av and 7 under. Wonder how many shots we have had from outside 50

We've had 36 shots from outside 50 the equal 9th most. As for expected score each shot at goal is given a percentage chance based on distance, angle, kicking foot and pressure. For example a shot at goal from the top of the goal square has a 95% chance of being a goal based on historical data going back to 2002 so therefore it's given an expected score of 5.7 (95% of 6). Each shot is goal is put through this formula and that's the total expected score.

Excellent post. Thanks OG.

The conclusions I draw from that are:

• Our biggest problem seems to be a horrible rate of conversion from really easy shots (56% within 15m, against comp avge of 77%). That is consistent with the eye test.
• I don't care about players who are below conversion average if they have only had a handful of shots. Let's focus on the ones who are regularly getting the ball in scoring range. Say at least once a game on average. The main culprits here are Naughton, Hannan, Vandermeer, B Smith, Scott, McNeil and Dunkley.
• Also, not a criticism, but the table of players doesn't drill down far enough to reveal some player anomalies. The most obvious example is Bont's figures which look good because he is an excellent shot on the run and from a distance (well above comp or expected average I'd guess) but he is deplorable from easy set shots close to goal. He should be very much under the microscope for that and should not get off lightly because of his good long range goals. It may well have cost us the win in an "8-point game" last Friday. It just seems such an easily fixable problem. My guess is he has about a 95% conversion rate on those at training. English might be another I suspect, but not as glaring. And Naughton definitely, but his averages are already sub-par overall.
• The other thing that makes it hard to draw conclusions is that we only have percentages and AFL ranking not absolute numbers in each of the team-wide categories. Again let's not sweat the stuff where the total numbers are small and a 15-20% improvement might equate to only two or three more goals over the season so far. Let's look instead at where we create our greatest number of opportunities but fall well short of the competition average or expected score.

Agree, Bont's biggest issue is his kicking from 15m-40m where he has kicked just 6 goals from 16 shots of which the majority of have been in the corridor.

 

[Fossie 32]

McKay - 55.4%
Hawkins - 47.7%
Walker - 45.7%



We've had 36 shots from outside 50 the equal 9th most. As for expected score each shot at goal is given a percentage chance based on distance, angle, kicking foot and pressure. For example a shot at goal from the top of the goal square has a 95% chance of being a goal based on historical data going back to 2002 so therefore it's given an expected score of 5.7 (95% of 6). Each shot is goal is put through this formula and that's the total expected score.



Agree, Bont's biggest issue is his kicking from 15m-40m where he has kicked just 6 goals from 16 shots of which the majority of have been in the corridor.

We've had 36 shots from outside 50 the equal 9th most. As for expected score each shot at goal is given a percentage chance based on distance, angle, kicking foot and pressure. For example a shot at goal from the top of the goal square has a 95% chance of being a goal based on historical data going back to 2002 so therefore it's given an expected score of 5.7 (95% of 6). Each shot is goal is put through this formula and that's the total expected score.
.

Hmmm being rated first for outside 50 is great given we have had a substantial amount of shots from that range then. Maybe it suggests we should do it more often rather than passing off though need to get it into the right hands first.
I get how the expected score is calculated [did watch the video!] but again it would be based on AFL average, so we should be better than average there as well as in other areas.

 

[Oliver G]

Another nice stat for Jacko is that he is now equal 1st in Kicks Inside 50 - Mark Rate (65+ kicks). Essentially meaning that 21% of his kicks inside 50 results in a mark. Jacko and Bonti's rates are almost double that of Petracca.



Now for an update on our list's accuracy and scoreboard impact. IP is Impact Points which is the player's score plus the score they've assisted. I50 - T is Inside 50 Target.

 

[maddog37]

Petracca would be taking a lot more long shots at goal than our boys wouldn’t he though?

 

[Fossie 32]

Another nice stat for Jacko is that he is now equal 1st in Kicks Inside 50 - Mark Rate (65+ kicks). Essentially meaning that 21% of his kicks inside 50 results in a mark. Jacko and Bonti's rates are almost double that of Petracca.

Got anything on most Bulldogs goals in first 16 minutes of a game?

 

[Footscrazy]

Jacko displaces Crossy to move into the top 10 disposals all time for the club. Should get Granty in a couple of weeks too.

 

[the dog's kennel]

Interestingly on On The Couch last night we were 2nd on the ladder for expected points, with Melbourne 1st. Which suggests that Melbourne are having more accuracy issues than we are. They also noted that Freo were 6th on the ladder for expected points (so their accuracy is a big problem), and Essendon and West Coast were right at the bottom of the ladder (16th and 17th from memory) so they are basically only being kept in it by their great accuracy (which we certainly saw examples of when we played the Eagles last week).

Of course we shouldn't get too excited about that analysis cause defence has something to do with ladder position as well!

 

[Boristown]

Interestingly on On The Couch last night we were 2nd on the ladder for expected points, with Melbourne 1st. Which suggests that Melbourne are having more accuracy issues than we are. They also noted that Freo were 6th on the ladder for expected points (so their accuracy is a big problem), and Essendon and West Coast were right at the bottom of the ladder (16th and 17th from memory) so they are basically only being kept in it by their great accuracy (which we certainly saw examples of when we played the Eagles last week).

Of course we shouldn't get too excited about that analysis cause defence has something to do with ladder position as well!

Nah, it just means they should have won the Adelaide game based on expected score and been 1 game ahead.

 

[Oliver G]

Interestingly on On The Couch last night we were 2nd on the ladder for expected points, with Melbourne 1st. Which suggests that Melbourne are having more accuracy issues than we are. They also noted that Freo were 6th on the ladder for expected points (so their accuracy is a big problem), and Essendon and West Coast were right at the bottom of the ladder (16th and 17th from memory) so they are basically only being kept in it by their great accuracy (which we certainly saw examples of when we played the Eagles last week).

Of course we shouldn't get too excited about that analysis cause defence has something to do with ladder position as well!

Here is each clubs average score, expected score, accuracy and expected accuracy.

 

[the dog's kennel]

Here is each clubs average score, expected score, accuracy and expected accuracy.

View attachment 1172059

I don't understand why your ladder of expected scores is so different to the ladder from On The Couch last night (which I have attached). I have no problem believing your stats over theirs, Oliver. But there is something funny going on somewhere. For example - look at Essendon - they have them at 17th. The table you quoted has them at 7th! The problem is their description is so vague and with so little info it's hard to know where the difference lies.

 

[stevieg1969]

I don't understand why your ladder of expected scores is so different to the ladder from On The Couch last night (which I have attached). I have no problem believing your stats over theirs, Oliver. But there is something funny going on somewhere. For example - look at Essendon - they have them at 17th. The table you quoted has them at 7th! The problem is their description is so vague and with so little info it's hard to know where the difference lies.

I’d say the On The Couch ladder is based on a team’s expected score v. the expected score of the team they played in a particular round - and then the associated wins/losses make up the ladder.

 

[Oliver G]

I don't understand why your ladder of expected scores is so different to the ladder from On The Couch last night (which I have attached). I have no problem believing your stats over theirs, Oliver. But there is something funny going on somewhere. For example - look at Essendon - they have them at 17th. The table you quoted has them at 7th! The problem is their description is so vague and with so little info it's hard to know where the difference lies.

The Expected Ladder which is the graphic shown on Fox Footy last night is how the ladder would be if each game result used Expected Scores.

So for example in R1 Adelaide beat Geelong 103-91. The Expected Score that day was Geelong winning 101-88 so that result is reversed on the Expected Ladder.

The table I posted earlier was each teams average score and average expected score.

 

[X_box_X]

I'm not sure if this is our record, but we have won our past ten matches against interstate sides.

 

[Pugz89]

I'm not sure if this is our record, but we have won our past ten matches against interstate sides.

Here is a quirky one, we have been ahead at the final siren at every ground we have played at this year.
MCG, Marvel, MARS Stadium, Manuka Oval, Adelaide Oval, Optus Stadium, GMHBA.

 

[worbod]

Something else quirky: No player with a surname beginning with U has played so far this season. After Marra Ugle-Hagan makes his debut on Sunday, this season will have seen players take the field whose surnames begin with every letter of the alphabet. Now that Tristan Xerri

38

Tristan Xerri

Age

25

Ht

201cm

Wt

99kg

Pos.

Ruck

Career

Season

Last 5

D

5.5

1star

K

3.0

1star

HB

2.5

2star

CL

1.0

3star

HO

2.5

5star

D

5.5

1star

K

3.0

1star

HB

2.5

2star

CL

1.0

3star

HO

2.5

5star

D

5.5

2star

K

3.0

1star

HB

2.5

3star

CL

1.0

3star

HO

2.5

5star

of North Melbourne has provided the league with its first ever player with an X surname, this means that 2021 is the first ever season where this could possibly have happened, as there were no players with U surnames in any team squads last year.

 

[Oliver G]

Time for an update on us giving up scores in the dying stages of quarters...

We've now given up 18.12 in the dying stages of quarters this year. As always I've defined dying stages as the last 3 minutes of total time in each quarter (i.e 30 min quarter, scores from 27:00 to 30:00).

Opposition Scoring - Quarter by Quarter (Dying Stages)


On the other side of the coin we've scored 18.16 in the dying stages.

Western Bulldogs Scoring - Quarter by Quarter (Dying Stages)


We've scored just 2 times in our last 12 quarters compared to the opposition scoring 10 times and since the bye we've just scored 2.1 to 7.5.

 

[X_box_X]

Time for an update on us giving up scores in the dying stages of quarters...

We've now given up 18.12 in the dying stages of quarters this year. As always I've defined dying stages as the last 3 minutes of total time in each quarter (i.e 30 min quarter, scores from 27:00 to 30:00).

Opposition Scoring - Quarter by Quarter (Dying Stages)
View attachment 1176198

On the other side of the coin we've scored 18.16 in the dying stages.

Western Bulldogs Scoring - Quarter by Quarter (Dying Stages)
View attachment 1176201

We've scored just 2 times in our last 12 quarters compared to the opposition scoring 10 times.

Wow, that's good. I love finding weaknesses because I take comfort knowing there's room for improvement to an already well-oiled machine.

 

[Oliver G]

Both West Coast and North Melbourne in the wet finished with better kicking efficiency than us yesterday.

 

[dogwatch]

Time for an update on us giving up scores in the dying stages of quarters...

We've now given up 18.12 in the dying stages of quarters this year. As always I've defined dying stages as the last 3 minutes of total time in each quarter (i.e 30 min quarter, scores from 27:00 to 30:00).

Opposition Scoring - Quarter by Quarter (Dying Stages)
View attachment 1176198

On the other side of the coin we've scored 18.16 in the dying stages.

Western Bulldogs Scoring - Quarter by Quarter (Dying Stages)
View attachment 1176201

We've scored just 2 times in our last 12 quarters compared to the opposition scoring 10 times and since the bye we've just scored 2.1 to 7.5.

So across the season it's not especially a weakness, even if the last three weeks have been worse (barely statistically significant). Like pretty much every club I'd suggest.

It's like free kicks. You notice them and amplify them when they go against you but when they go your way you just think "wow how good are we"?

 

[X_box_X]

Under Bevo, whenever our last Home and Away match is against an interstate side, we go on to play finals. Whenever our last Home and Away match is against a Victorian team, we miss finals.

Looking ahead at the fixture, I see we play an interstate team in the final round.

I am therefore confident we will play finals. It's written in the stars.

Can we win none of our remaining six matches to secure a finals birth?

 

[Bont2Bruce]

Can we win none of our remaining six matches to secure a finals birth?

Wouldn't be guaranteed of a finals spot.