September 16, 2015

Rugby World Cup Predictions for Week 1

Team Ratings for Week 1

The basic method is described on my Department home page.

Here are the team ratings prior to this week’s games, along with the ratings at the start of the Rugby World Cup.

Current Rating Rating at RWC Start Difference
New Zealand 29.01 29.01 0.00
South Africa 22.73 22.73 0.00
Australia 20.36 20.36 0.00
England 18.51 18.51 0.00
Ireland 17.48 17.48 0.00
Wales 13.93 13.93 -0.00
France 11.70 11.70 0.00
Argentina 7.38 7.38 -0.00
Scotland 4.84 4.84 -0.00
Samoa -2.28 -2.28 0.00
Fiji -4.23 -4.23 -0.00
Italy -5.86 -5.86 0.00
Tonga -6.31 -6.31 0.00
Japan -11.18 -11.18 -0.00
USA -15.97 -15.97 -0.00
Georgia -17.48 -17.48 -0.00
Canada -18.06 -18.06 -0.00
Romania -21.20 -21.20 -0.00
Uruguay -31.04 -31.04 -0.00
Namibia -35.62 -35.62 -0.00

 

Predictions for Week 1

Here are the predictions for Week 1. The prediction is my estimated expected points difference with a positive margin being a win to the first-named team, and a negative margin a win to the second-named team.

Game Date Winner Prediction
1 England vs. Fiji Sep 18 England 22.70
2 Tonga vs. Georgia Sep 19 Tonga 17.70
3 Ireland vs. Canada Sep 19 Ireland 42.00
4 South Africa vs. Japan Sep 19 South Africa 40.40
5 France vs. Italy Sep 19 France 24.10
6 Samoa vs. USA Sep 20 Samoa 20.20
7 Wales vs. Uruguay Sep 20 Wales 45.00
8 New Zealand vs. Argentina Sep 20 New Zealand 28.10

 

avatar

David Scott obtained a BA and PhD from the Australian National University and then commenced his university teaching career at La Trobe University in 1972. He has taught at La Trobe University, the University of Sheffield, Bond University and Colorado State University, joining the University of Auckland, based at Tamaki Campus, in mid-1995. He has been Head of Department at La Trobe University, Acting Dean and Associate Dean (Academic) at Bond University, and Associate Director of the Centre for Quality Management and Data Analysis at Bond University with responsibility for Short Courses. He was Head of the Department of Statistics in 2000, and is a past President of the New Zealand Statistical Assocation. See all posts by David Scott »

Comments

  • avatar
    David Jenkins

    Are you weighting England and Wales for home advantage?

    9 years ago

    • avatar

      Yes, when they are playing in England and Wales respectively. (IIRC Wales play at Twickenham.)

      9 years ago

  • avatar
    Ed Do Cerro

    I was particularly impressed with your accuracy rates for the 2011 World Cup. I told my colleagues that stats can be used to predict match results. They were skeptical but your work is slowly converting them. I hope you’ll continue to post weekly predictions for the duration of the World Cup? Thumbs up so far.

    9 years ago

  • avatar
    Sam Warburton

    South Africa by 40!

    9 years ago

    • avatar
      Megan Pledger

      If sport were perfectly predictable there wouldn’t be much point.

      9 years ago

      • avatar

        I am perfectly happy with my prediction. I wouldn’t have any faith in a prediction system that predicted a Japanese win in that game. No prediction system is going to get things right always, and a failure in a particular game is not a failure of the system. What makes a prediction system good or bad is long term performance. As Megan says, sport is not predictable. Why would anyone play if the outcome was predictable?

        9 years ago

        • avatar
          Megan Pledger

          It looks like SA will get an easier run to the final if they come second in their pool – they’ll miss out on Australia/England in the quarter final and NZ in the semi-final.

          9 years ago

  • avatar
    Josh Crabbe

    Thoughts for A. England vs Wales? I think Wales are definitely in with a shot. And B. Scotland vs Japan??

    9 years ago

    • avatar

      I am afraid I don’t agree on Wales’ chances. I have a probability of 0.76 for an England win. That might be a bit high given that I have given England my standard home ground advantage for international games.

      9 years ago

  • avatar
    Remesh Kumar

    Hi there David Scott..
    I suppose in your rating calculations, you’d have an element for upset results..eg. South Africa v Japan
    SA=22.73
    J=-11.18
    Prediction Margin = 40.4

    What sort of weightage is assigned to these upset results? And.. would it require an adjustment factor if one see that the team has significantly showed vast amount of improvements in the game?

    Hence, updated ratings for SA & J would be ???

    Also, do you have any adjustment factor for major injury to main team players?

    Eg. NZAB = 29.01
    Assume that Carter, McCaw, A.Smith not included in squad… ratings change?

    Compare to USA = -15.97
    Assume Chris Wyles, AJ MacGinty, Samu Manoa & Chris Baumann out of squad.. ratings change?

    Thanks & Cheers.

    9 years ago

    • avatar

      I set up my data incorrectly for the first games which assigned home ground advantage wrongly. After correction my margin for South Africa was 34 points (just the difference between the ratings, no home ground advantage). As a result South Africa’s rating has gone down 1.6, Japan’s up by 1.6.

      I take no notice of team composition, just scores and home ground advantage. My approach is totally algorithmic and numeric, with no personal judgement involved other than in selection of parameters. It would be difficult to quantify the effect of the loss of different players. How much worse would the All Blacks be with Barrett, Cane and Perenara instead of the three you mention?

      9 years ago

      • avatar
        Remesh Kumar

        Hi David,

        Appreciate your comments above.

        I take note that your ratings are based on algorithmic & numeric and i respect that. My basic interest in your ratings is for comparison with the sports book odds and whilst i see similarities, there is differences between the 2 and I’m simply trying to figure our what contributes to these differences.

        I just figure that whilst they are based on historic data, there must be an adjustment factor for extraordinary items/events. Eg. Japan’s shock win over SA. Yes, we can say that it’s like a one-off thingy, but if one watch the game between to 2 teams, you got to admit that Japan have actually progressed vastly and their game display was not of a fluke but tactically proficient. Why I’m saying this is becoz, taking into adjusted ratings of SA 21.13 and Japan -9.58, that would mean if another game takes place between these 2 teams, the prediction would still be SA winner with +30pts. This is theoretical but I believe that it would be fairer to say that SA would still be favorite but with a lower ratings difference of +7pts to +10pts, simply becoz of an accelerated adjustment factor to take into account the present improved team of Japan.

        On a similar note, team composition also will contribute to the ratings (meaning requiring an adjustment factor). As mentioned earlier, you rightfully indicate that the ratings might not matter in the case of All Blacks should they lose Barrett, Cane and Perenara instead of the 3 players I mentioned. However, for a smaller team like USA, Namibia or Uruguay or Canada, loss of their major play makers should seriously impact the performance of the team and hence their ratings.

        Just trying to figure out a way to assign these adjustment factors if you could help, greatly appreciated.

        Thanks & Cheers !

        9 years ago

  • avatar
    Gavin Allwright

    David,

    This is great but I am not a stats guru and maybe I am missing something. Could you however explain to me how you came up with a PD for Tonga v Georgia. I am sure it would explain all the other PD’s you have. Additionally, if I understand it correctly, with the anomaly of the SA loss this weekend would it mean that SA has gone down to 21.13 and Japan up to -9.58

    9 years ago

    • avatar

      I have data back to 1995 on all international games between teams involved in the RWC. I start from 1995 with all teams rated 0, then update using game results to get my starting ratings for this RWC. Predictions are then rating of team1 minus rating of team2 + home ground advantage (if not neutral). For Tonga versus Georgia, they are playing on a neutral ground, so it is just the difference in ratings. I messed up the specification of neutral though in the first round of games. Corrected predictions will be included in my post for the second round of games.

      9 years ago

  • avatar
    Kevin Rogers

    Will you be releasing week 2 predictions before the games on Wednesday? (I note you said you’d be providing them each Wednesday).

    9 years ago

    • avatar
      PJ Hoon

      Yeah will you release the week 2 results on Wednesday?

      9 years ago

    • avatar

      Yes. Wednesday morning NZ time so well ahead of the actual game time.

      9 years ago

      • avatar
        Johan Nel

        OK thanks David. Trust it will be as accurate as round 1!!

        9 years ago

  • avatar
    Bob Xenophon

    Love your work David!

    Cold hard stats that’s seen me lead the office tipping comp despite the fact that I still don’t have a firm grasp on the difference between a scrum and a maw…

    Like the others eagerly waiting on your predications for week 2.

    9 years ago

  • avatar
    Dave Blakey

    Any picks for the matches tomorrow?

    9 years ago

  • avatar
    Frans Boot

    David
    Is South Africa’s shocking defeat going to influence your future predictions?

    9 years ago

  • avatar
    Wayne Henning

    Looking forward to trying out your predictions. Ill be waiting gor your updates. Thanks

    9 years ago

  • avatar
    Bruce Grandview

    Hi David,

    Thanks for your tips, I managed to get the Wales margin and take 1st place in my office comp!
    Can you please let me know if you will be posting this Sundays games before the weekend?
    Cheers.

    9 years ago

    • avatar

      Predictions up to the England-Wales game have been posted. I might have to give an update to include the games up to Ireland-Romania now that Australia and Scotland have played. I will have to do a bit of programming to sort that out.

      9 years ago

  • avatar
    Surette myburgh

    I want to follow your stats

    9 years ago

  • avatar
    Shreeni de Silva

    Please send me the new post if any

    8 years ago