From inside the world of Champion Data

Discussion in 'Blog' started by ppym, Jun 15, 2011.

By ppym on Jun 15, 2011 at 10:00 AM
  1. ppym

    ppym New Member

    Joined:
    Jan 3, 2016
    Messages:
    1
    Likes Received:
    0
    A few weeks ago, I was lucky to get the opportunity to ask Karl Jackson, one of the statisticians at Champion Data a few questions about some of the finer points of this game that we all love.

    So tune in over the break for a little glimpse behind the CD Curtains...

    <!--more-->

    We all know that a game-winning goal kicked at the death of a close game is worth a premium, but how does the weighting work in other situations? For instance, in a blowout (Think Ess v GC), do the actions from the first quarter receive a premium? What about a game that is tight until half time that blows out in the third?

    KJ: The weighting is applied to everything that happens within a game. To put it basically, the more likely it is that the result of the match is going to change, the higher the weighting is. This works in both ways – as time goes on in tight games, a player’s actions become more important because the game is up for grabs, but as time goes on in blowout games, events are weighted down because the game is moving closer and closer to “junk time”.

    We follow along with the quarter time scores posted on the HUN and know well about the 3300 rule, but why is it that it appears that 1/3 of the 3300 seems to be allocated at the quarter time break?

    KJ: There’s no hard and fast rule about how many points to allocate at the end of the first quarter. There is so long left in the game that it would be folly to guess how it will be played out. From half time onwards we’ve got a better idea of how the game has/will be played out so that’s when the standardised scores are used.

    If a handball or kick is received on the bounce by a teammate, is it still effective?

    KJ: If the teammate was put under unnecessary pressure because of the disposal dropping short it will be called ineffective, but if it hasn’t put him under pressure it will still be called effective.

    What about a dropped mark? Does the kicker still get an effective stat? Or is it completely dependent on the receiver not messing up?

    KJ: The actions of the receiver don’t impact the decision on whether a disposal was effective. If the disposal player did everything he could to lay it on a platter for the receiver it will be effective even if it’s a blatant error at the other end.

    Is a player whos kicking in from an opposition behind awarded points for that kick? or only when its effective and long?

    KJ: Kick-ins are judged the same as any other kick. Effective kick-ins get a positive score (though not as high as a general play kick) and clanger kick-ins get a negative score. As with general play kicks, the most valuable outcome for a player is kicking long and finding an uncontested teammate.

    We noticed in the off-season that a group of new stats were created, are these changes applicable this year and how exactly does a mark on the lead work?

    KJ: Nearly every year there are new stats recorded, but not all make their way into the rankings formula. Marks on lead have been around for a few years. They are graded higher than a standard uncontested mark but lower than a contested mark.

    How many people do you guys have monitoring each game? And how do you determine the outcome of a contentious decision?

    KJ: In total there are 10 people working on every game but not all of those have an impact on the standard stats capture. There is a core team of four that are involved in that process – one main caller (based at the ground), one keyboarder (entering the data live as it happens), one back-up caller (watching a clean vision feed stripped of advertisements) and an at-ground support role (assisting with free kicks and off-ball incidents). The main caller and the back-up caller are the two with the most influence and in the event of a contentious decision it will be reviewed at the quarter break. As a result you may see the efficiency, disposal count, tackle count, etc, etc vary from the end of quarter to the start of the next.

    The other six we have working on games are capturing other information that doesn’t affect fantasy, such as interchange moves, player match-ups, pressure on opposition disposals and high detail information about every kick and the location of every possession.

    Now, as a particular case study - it’s been asked, what exactly did Rioli do in round 9 to secure 100 points? 13 touches at a 77% efficiency rating including a goal and two goal assists in a 56 point win seems a little much, particularly when you have someone like Hodge, who scored 99 points off 10 more touches and 2 goals. Is there anything you can say that could shed some light on this example?

    KJ: As for Rioli v Hodge – Rioli’s goal and two assists all came in the third quarter, which is when Hawthorn ran away with the game. With the margin at 18 points with four minutes to go in the third quarter, Cyril made a tackle, won the free kick and kicked the goal – a 30 second block that saw him get 19 points. Hawthorn went on to kick another four unanswered goals after Rioli’s goal so that was when the game was most in the balance. Luke Hodge had six touches in the first 15 minutes of the third, but just 50% were effective. When Hawthorn were on their streak of five goals in a row, Hodge only had one touch and that was after they’d stretched the lead to 42 points, so he missed out on getting bumped up.

    On a side note, I’ll just highlight the fact that you can’t compare the outputs of players in different games and expect their rankings to follow suit. Because of the 3300 point rule, a 25 disposal game in a tight, low-possession game could be worth the same as a 40 disposal game in a free-flowing contest. Likewise, a 10 goal game in a high scoring contest may be less valuable than a four goal return in a low scoring game. It avoids any bias towards a particular game style or weather conditions.

    The other example thats being asked about is Goddard from round 10- injured in the first quarter and yet his quarter-by-quarter scores went: 56 SuperCoach at qtr time 44 at half time and 51 at 3 qtr time and finally 57 at full time- Im assuming it weighting his first quarter efforts so highly because the game blew out so far in the 1st, but an official word would be most appreciated!

    KJ: The quarter time updates on the Herald Sun website take into account the 3300 point rule, but because of the fluctuations that happen within the game, they’re always going to change. As the game blows out more and more, the first quarter gets pumped up because of the reduced value of actions later in the game. In a stock standard game you’d expect 25% of the ranking points to come from each quarter, but as an example, in that Fremantle v St Kilda game in Round 10, 39% of the points were won in the first quarter. This is what helped Goddard maintain his score. If it was tight all the way to the siren, his score would have dropped as the game went on.

    So there you go, some solid answers for some of the questions that weve been wondering about, Im hoping that it sheds a little light not only on what makes this game a joy, but also on those seemingly annoying quirks that we wonder about.

    And as for the question of why Goddard scored so well earlier in the year despite what looked like poor form- its because when he did get into the game, it was when the game was still there to be won. (And not the gag response of- because they had him locked in as their perma-captain early.)
     

Comments

Discussion in 'Blog' started by ppym, Jun 15, 2011.

Share This Page