Analysis SuperCoach Scoring Explained, Observations & Complaints On Scoring
- Thread starter I am Cursed :(
- Start date
I'm loving Rance's 14.00 DT : SC ratio. It's beautiful even for a non-owner. Although, hope he gets his 75 SC for the bingo 
I noticed last night how Rance is constantly drawing and shepherding the opposition. He might get an uncontested possession but because he draws the opponent before he disposes of the ball the disposal is contested. He won't just handpass to a team mate. He will turn his backside into the opponent and then handpass to a team mate. He is constantly looking for body contact.
Edit: Yes, I'm having a quiet day watching the rain.
- Joined
- 6 Jul 2012
- Messages
- 2,301
- Likes
- 1,721
- AFL Club
- North Melb.
See the comments of Stephen and Darkie, I think they're absolutely right. In the past, CD were lagging behind a bit in terms of updating the "during the game weighting" based on closeness of the match, so it happened a bit after the game, and was therefore a correction of an error.
Bare in mind also, that if every player goes up by the same % in post game scaling, players that have scored more will go up more. Eg. Say Parfitt scores 50 and Danger 120, and both players are scaled up by 4 %. Parfitt will only go up by 2, while Danger will go up by 4.8 (rounded to 5).
This shouldn't affect whether it's a contested possession or not. Loose ball gets and hard ball gets both count as contested possessions, because they are winning the ball in dispute, and are consequently worth the same points.
Bare in mind also, that if every player goes up by the same % in post game scaling, players that have scored more will go up more. Eg. Say Parfitt scores 50 and Danger 120, and both players are scaled up by 4 %. Parfitt will only go up by 2, while Danger will go up by 4.8 (rounded to 5).
Under the SCS ranking system Rance had 12 possessions with 6 of them being contested.
I noticed last night how Rance is constantly drawing and shepherding the opposition. He might get an uncontested possession but because he draws the opponent before he disposes of the ball the disposal is contested. He won't just handpass to a team mate. He will turn his backside into the opponent and then handpass to a team mate. He is constantly looking for body contact.
This shouldn't affect whether it's a contested possession or not. Loose ball gets and hard ball gets (both getting the ball when in dispute), both count as contested possessions and are worth the same.
Edit: Yes, I'm having a quiet day watching the rain.
I noticed last night how Rance is constantly drawing and shepherding the opposition. He might get an uncontested possession but because he draws the opponent before he disposes of the ball the disposal is contested. He won't just handpass to a team mate. He will turn his backside into the opponent and then handpass to a team mate. He is constantly looking for body contact.
This shouldn't affect whether it's a contested possession or not. Loose ball gets and hard ball gets (both getting the ball when in dispute), both count as contested possessions and are worth the same.
Edit: Yes, I'm having a quiet day watching the rain.
- Joined
- 6 Jul 2012
- Messages
- 2,301
- Likes
- 1,721
- AFL Club
- North Melb.
- Joined
- 6 Jul 2012
- Messages
- 2,301
- Likes
- 1,721
- AFL Club
- North Melb.
For those wondering about Bontempelli's game tonight:
1 long kick to advantage = 5
appoximately 7 other effective kicks = 21
approx 10 effect handballs = 10
1 contested mark from teammate in forward 50 = 5
1 contested mark from teammate in middle = 4
13 other CP = approx 42-48 depending on field position
7 UP = approx 7
7 tackles = 21
3 one %s = 3
3 goals = 18
1 assist = 3
2 clangers = -6
TOTAL = 133-138
His worst quarter was the least important first quarter, so probably added about 3-4 to his score.
EDIT: Erich's observations tell him that a free against in a contested situation is only -1, rather than than -3. Personally I never noticed that, but could easily have missed it. Bont's two clangers were frees against, so could put his score even higher.
1 long kick to advantage = 5
appoximately 7 other effective kicks = 21
approx 10 effect handballs = 10
1 contested mark from teammate in forward 50 = 5
1 contested mark from teammate in middle = 4
13 other CP = approx 42-48 depending on field position
7 UP = approx 7
7 tackles = 21
3 one %s = 3
3 goals = 18
1 assist = 3
2 clangers = -6
TOTAL = 133-138
His worst quarter was the least important first quarter, so probably added about 3-4 to his score.
EDIT: Erich's observations tell him that a free against in a contested situation is only -1, rather than than -3. Personally I never noticed that, but could easily have missed it. Bont's two clangers were frees against, so could put his score even higher.
Last edited:
That's from the confirmed ways to score by CD, what about the other 100 they don't list? This whinging about scoring happens every year (and yes i've been part of it from time to time), but people just need to get over it. If you want a stat by stat and an open list of ways players score, Supercoach isn't for you. I've finally come to just accept how CD score players and enjoy the horribly umpired games.
- Joined
- 6 Jul 2012
- Messages
- 2,301
- Likes
- 1,721
- AFL Club
- North Melb.
I wouldn't take such a negative view of either CD or umpiring. I don't think the scoring system is as mysterious as you make out. The scores I've worked out are based mainly on observation, thanks to field view on the herald sun website last year (which they've unfortunately now taken away, but luckily I wrote down what everything was worth). What makes you think there are 100 other stats? Players scores don't mysteriously go up or down, they follow a surprisingly consistent pattern.
As for the umpiring, given the AFL's ridiculous rule changing, they're doing a pretty damn good job!
As for the umpiring, given the AFL's ridiculous rule changing, they're doing a pretty damn good job!
Just on Bont, he scored 52 in the last quarter. From the words of @championdata on twitter, in the last quarter these were his stats:
4 contested possessions (most was 5), including a contested mark in F50
4 tackles (most of any player)
3 disposals all effective
1 goal
Agree or disagree, there it is. Is it justified? Seems a matter of opinion.
4 contested possessions (most was 5), including a contested mark in F50
4 tackles (most of any player)
3 disposals all effective
1 goal
Agree or disagree, there it is. Is it justified? Seems a matter of opinion.
This is one that goes under the radar for sure. I see a lot of people go "[insert name of 'CD favourite' here] just gave away a free and I didn't even notice their score go down!". Easy to miss a -1, especially if they've gained other points in the passage of play. And then a player they have on their team gives away a free in an uncontested situation and gets -3 on their score and it seems to set them off on a little rant about 'CD favourites', 'bias towards certain players', 'golden tickets' etc...
Someone might go, "this bloke has 5 clangers, he must have butchered the ball, how is his SC score so high?!" when in fact, the player may have given away 3 or 4 frees against in contested situations, which have little negative affect on scoring.
Someone might go, "this bloke has 5 clangers, he must have butchered the ball, how is his SC score so high?!" when in fact, the player may have given away 3 or 4 frees against in contested situations, which have little negative affect on scoring.
That's fine, you don't have to be negative. I wouldn't say i'm negative per-se, but more of a realist.
In my opinion, they can be consistent without being transparent and transparency equals clarity. The scoring system is very mysterious, especially when the Herald Sun/Champion Data don't provide every way to score in the FAQ on the Supercoach website. Also, i'm 99% sure, they have said in direct terms, that there are over 100 ways for players to score/lose points and considering they only show 20 on the website i took a leap. They keep a lot of ways to score under wraps so other statistical teams don't use their methods. It's what keeps them ahead of the game and why the AFL/clubs use their data. You've mentioned 1%ers, yet in the FAQ page this is not noted. I'm fairly certain CD reward carrying the ball and bouncing it, this is also not listed as a scoring possibility. How about pressure acts, shepherds, spoils or even giving away a 50m penalty?
This is noted in the FAQ page:
"Some scores involve an element of judgement by Champion Data. No correspondence will be entered into regarding the allocation of scores."
http://supercoach.heraldsun.com.au/...elp/maintablemiddle/contactus?hidemodal=false
To me, that says enough for them to be as vague as they want without answering any questions.
I 100% agree with the ridiculous rule changing making it more difficult for umpires and yes they do have a very tough job. Isn't that partly why the AFL added more umpires for each game? However, do you not think the umpires/umpiring committee have a voice in what is logical and what moves forward with the new rules? Don't they have meetings on how the new (and old) rules should be interpreted? Of course they do. In round 5 we had a free kick range of 18-46. How can there be that much of a disparity game to game? 23, 34, 32, 42, 45, 46, 26, 44 and 18. The AFL and umpires need to make up their mind if they are going to pay every single little infringement or have a more lenient interpretation. I'm sure it confuses the players as much as it does the fans, yet they can't show any criticism as they will be fined for pretty much any negative remark.
I didn't want to be drawn into this type of thread again and it's why i said people should just accept how CD score and be done with it.
In my opinion, they can be consistent without being transparent and transparency equals clarity. The scoring system is very mysterious, especially when the Herald Sun/Champion Data don't provide every way to score in the FAQ on the Supercoach website. Also, i'm 99% sure, they have said in direct terms, that there are over 100 ways for players to score/lose points and considering they only show 20 on the website i took a leap. They keep a lot of ways to score under wraps so other statistical teams don't use their methods. It's what keeps them ahead of the game and why the AFL/clubs use their data. You've mentioned 1%ers, yet in the FAQ page this is not noted. I'm fairly certain CD reward carrying the ball and bouncing it, this is also not listed as a scoring possibility. How about pressure acts, shepherds, spoils or even giving away a 50m penalty?
This is noted in the FAQ page:
"Some scores involve an element of judgement by Champion Data. No correspondence will be entered into regarding the allocation of scores."
http://supercoach.heraldsun.com.au/...elp/maintablemiddle/contactus?hidemodal=false
To me, that says enough for them to be as vague as they want without answering any questions.
I 100% agree with the ridiculous rule changing making it more difficult for umpires and yes they do have a very tough job. Isn't that partly why the AFL added more umpires for each game? However, do you not think the umpires/umpiring committee have a voice in what is logical and what moves forward with the new rules? Don't they have meetings on how the new (and old) rules should be interpreted? Of course they do. In round 5 we had a free kick range of 18-46. How can there be that much of a disparity game to game? 23, 34, 32, 42, 45, 46, 26, 44 and 18. The AFL and umpires need to make up their mind if they are going to pay every single little infringement or have a more lenient interpretation. I'm sure it confuses the players as much as it does the fans, yet they can't show any criticism as they will be fined for pretty much any negative remark.
I didn't want to be drawn into this type of thread again and it's why i said people should just accept how CD score and be done with it.
Interesting conversation between a twitter user and Champion Data:
User:
Still no explanation into why Danger gets 30pts for a winning goal and Roughie only gets 9pts.
CD:
Danger got 21. Roughead got 18.
Dangerfield won a hardball get (4.5 points). Roughead an uncontested mark (2 points).
Multipliers for state of game (margin & time left):
Dangerfield: 1.95
Roughead: 1.96
3300 Point multiplier:
Danger: 0.62
Rough: 0.60
User:
Did Roughead get 18pts though? Are these scores and times incorrect?
27:30 71
28:05 76
28:34 87
Full Time 85
+ late update 86
CD:
Live scores are only provisional. Every time something happens in a game the value of every previous event is changed.
When it's live, the algorithm "guesses" the total points that will be scored pre-3300. As game goes on, that guess is closer to final points.
At the full time siren there is no need to extrapolate because it's known.
Pretty interesting stuff and nice to know, just a shame someone has to nag away at Champion Data on twitter for a few days to get this kind of info out of them.
User:
Still no explanation into why Danger gets 30pts for a winning goal and Roughie only gets 9pts.
CD:
Danger got 21. Roughead got 18.
Dangerfield won a hardball get (4.5 points). Roughead an uncontested mark (2 points).
Multipliers for state of game (margin & time left):
Dangerfield: 1.95
Roughead: 1.96
3300 Point multiplier:
Danger: 0.62
Rough: 0.60
User:
Did Roughead get 18pts though? Are these scores and times incorrect?
27:30 71
28:05 76
28:34 87
Full Time 85
+ late update 86
CD:
Live scores are only provisional. Every time something happens in a game the value of every previous event is changed.
When it's live, the algorithm "guesses" the total points that will be scored pre-3300. As game goes on, that guess is closer to final points.
At the full time siren there is no need to extrapolate because it's known.
Pretty interesting stuff and nice to know, just a shame someone has to nag away at Champion Data on twitter for a few days to get this kind of info out of them.
Last edited:
You probably already are aware that in each game the sum of all 44 players SC scores adds to 3,300 give or take a couple of points. However if one totals all of the pure points that players are rewarded for each action they are involved in it will often be miles off the 3,300 figure. So in the Roughead game above the total for Swans and Hawks players must have come to around 5,500. Hence the '3300 point multiplier' in this case was 0.60
What is interesting is that the algorithm is constantly guessing throughout the game what the final total points will likely be, hence the '3300 multplier' factor is being shown immediately and constantly adjusting. This likely explains the reasoning why in Bontempele and Erich's observations, in their great knowledge of the scoring system, say that most of the points allocated for an individual act are a bit less than what the starting point allocation in the tables earlier in this thread that CD have stated. As the algorithm probably starts around the 0.7 figure.
Further its is one of many explanations in the wide variance between players scores in different games. Let's assume in the very rare instance that Player A in game A has the exact same number and type of acts in a game as Player B in game B. So exactly the same number of hard ball gets, effective and ineffective long & short kicks and handballs, tackles, 1%ers and all of the others measures that CD don't even tell us about. So both have and identical raw score of 200. Players A's '3300 multiplier' is 0.5 meaning that he scores 100. Player B's '3300 multiplier' is 0.70 meaning he scores 140.
That's a massive 40 point difference for identical statistical performances!! The only difference being the number of pure SC points the other 43 players in both games achieved. Then if one was to factor in the 'state of the game' multiplier, so say the majority of Player B's points occurred when the game was on the line, whereas Player A had a lot of his points come in junk time; that 40 point difference would blow out to be even larger!
Yet we punters like to look at just a few key stats and wonder why the scores are so different, when we are barely even scratching the surface.
What is interesting is that the algorithm is constantly guessing throughout the game what the final total points will likely be, hence the '3300 multplier' factor is being shown immediately and constantly adjusting. This likely explains the reasoning why in Bontempele and Erich's observations, in their great knowledge of the scoring system, say that most of the points allocated for an individual act are a bit less than what the starting point allocation in the tables earlier in this thread that CD have stated. As the algorithm probably starts around the 0.7 figure.
Further its is one of many explanations in the wide variance between players scores in different games. Let's assume in the very rare instance that Player A in game A has the exact same number and type of acts in a game as Player B in game B. So exactly the same number of hard ball gets, effective and ineffective long & short kicks and handballs, tackles, 1%ers and all of the others measures that CD don't even tell us about. So both have and identical raw score of 200. Players A's '3300 multiplier' is 0.5 meaning that he scores 100. Player B's '3300 multiplier' is 0.70 meaning he scores 140.
That's a massive 40 point difference for identical statistical performances!! The only difference being the number of pure SC points the other 43 players in both games achieved. Then if one was to factor in the 'state of the game' multiplier, so say the majority of Player B's points occurred when the game was on the line, whereas Player A had a lot of his points come in junk time; that 40 point difference would blow out to be even larger!
Yet we punters like to look at just a few key stats and wonder why the scores are so different, when we are barely even scratching the surface.
Great explanation Phizsay.
So on that basis, if the siren sounded to start the game and one player got an effective clearance to another player and then immediately all players went on STRIKE and sat down for the rest of the match, then those two players would get all the 3300 points...(despite having a very ordinary game)
Now the SC challenge is to work out when that strike action is going to start and who those two players are going to be. If you nail this, you win the SC prize. Simple game...
So on that basis, if the siren sounded to start the game and one player got an effective clearance to another player and then immediately all players went on STRIKE and sat down for the rest of the match, then those two players would get all the 3300 points...(despite having a very ordinary game)
Now the SC challenge is to work out when that strike action is going to start and who those two players are going to be. If you nail this, you win the SC prize. Simple game...