September 27, 2016

Mitre 10 Cup Predictions for Round 7

Team Ratings for Round 7

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 season.

Current Rating Rating at Season Start Difference
Canterbury 16.20 12.85 3.30
Taranaki 7.24 8.25 -1.00
Auckland 5.81 11.34 -5.50
Tasman 4.73 8.71 -4.00
Counties Manukau 4.26 2.45 1.80
Otago 2.79 0.54 2.30
Wellington 2.70 4.32 -1.60
Waikato -0.75 -4.31 3.60
Manawatu -4.24 -6.71 2.50
Bay of Plenty -4.63 -5.54 0.90
Hawke’s Bay -4.71 1.85 -6.60
North Harbour -5.44 -8.15 2.70
Southland -13.04 -9.71 -3.30
Northland -14.42 -19.37 5.00

 

Performance So Far

So far there have been 46 matches played, 36 of which were correctly predicted, a success rate of 78.3%.
Here are the predictions for last week’s games.

Game Date Score Prediction Correct
1 Southland vs. Bay of Plenty Sep 21 20 – 16 -4.90 FALSE
2 Northland vs. Wellington Sep 22 21 – 29 -14.20 TRUE
3 Counties Manukau vs. Waikato Sep 23 35 – 26 9.00 TRUE
4 Canterbury vs. Otago Sep 24 45 – 34 18.80 TRUE
5 Taranaki vs. Manawatu Sep 24 30 – 19 16.50 TRUE
6 Hawke’s Bay vs. Tasman Sep 24 29 – 36 -5.10 TRUE
7 North Harbour vs. Southland Sep 25 35 – 14 10.30 TRUE
8 Bay of Plenty vs. Auckland Sep 25 38 – 44 -5.70 TRUE

 

Predictions for Round 7

Here are the predictions for Round 7. The prediction is my estimated expected points difference with a positive margin being a win to the home team, and a negative margin a win to the away team.

Game Date Winner Prediction
1 Waikato vs. Canterbury Sep 28 Canterbury -13.00
2 Tasman vs. Counties Manukau Sep 29 Tasman 4.50
3 Wellington vs. Southland Sep 30 Wellington 19.70
4 North Harbour vs. Bay of Plenty Oct 01 North Harbour 3.20
5 Manawatu vs. Hawke’s Bay Oct 01 Manawatu 4.50
6 Auckland vs. Otago Oct 01 Auckland 7.00
7 Taranaki vs. Canterbury Oct 02 Canterbury -5.00
8 Northland vs. Waikato Oct 02 Waikato -9.70

 

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
    Mike Hanley

    Despite David’s clear prediction, I’m struggling to pick Canterbury over Waikato tonight. Waikato playing for the Shield at home, and their form has been improving.

    8 years ago

    • avatar
      david pettit

      They were lucky to beat Taranaki(stupid penalty off the ball cost Naki a try that would have sealed the match) and they were well beaten by Counties. So many injuries and D-Mac + TKB away with the ABs. There is less than half of their best starting XV left. Canterbury will punish them.

      8 years ago

      • avatar
        Graham Goldie

        I think Canterbury also. Hope we smash them

        8 years ago

  • avatar
    sam smith

    I have a hard time seeing Auckland beating Otago

    8 years ago

  • avatar
    Joseph Hijara

    I’m fairly sure most of these predictions are just the odds on favourite. I don’t need a fancy computer algorithm to tell me that people think the favourite will win… Counties Manukau to beat Tasman today

    8 years ago

    • avatar

      So how did your prediction work out?

      The advantage of an algorithm is that it avoids human judgement and hence some possible biases. I have some views as to who might win, but those views are irrelevant in the predictions I produce, and as to any odds that might be available, I have no idea whatsoever. If it does happen there is an odds-on favourite, I would hope my algorithm would suggest the favourite as the winner, and with a large predicted points difference.

      What you are right about is that the Mitre 10 Cup seems very predictable this year. Early on I was running at 84% correct for predicting the winner.

      8 years ago

      • avatar
        Megan Pledger

        Some of predicting well at the beginning could be the effect of the draw. Generally draws start with the hard teams playing the easy teams and end with the hard teams playing each other. It heightens the excitement at the end of the season.

        8 years ago

        • avatar

          Predicting at the start of the season can be difficult because of changes in personnel between seasons. I do some shrinkage of ratings between seasons to help with that. Some others allow their ratings to respond much more rapidly to recent results early in the season than later in the season to deal with the initial unpredictability.

          8 years ago

  • avatar
    forlan peter

    Hi David,

    I have been following your prediction for over two years. I find it very good.

    Understand that the accuracy of prediction vary on different tournament, from 65% to almost 80%.

    I do have one finding that would like to ask your opinion. For example Mitre 10, when you predict a game, if the spread is less than 13, usually it is very precise. However, if the spread goes to more than 20, and sometime more than 25 in Currie Cup, the prediction is not that precise. For example in Currie Cup, A vs B and the spread is almost 28, then A may most likely to win by less than 28 points.

    Is there a particular reason why?

    Thank you very much.

    Forlan

    8 years ago

    • avatar
      forlan peter

      I did a little research on the previous weeks’ prediction.

      Where the spread is over 20 points
      Round 5: 2
      Round 4: 2
      Round 3: 2
      Round 2: 1

      Out of these 7 games, only Canterburry 63-7 against Hawke’s Bay which can win more than the spread.

      I am just sharing these numbers. Again thank you so much for your prediction. I really like it.

      8 years ago

    • avatar

      A bit of theory. My predictions give the average points difference that I expect in the game. The actual result is distributed around that value (assuming I have predicted the average correctly) and from experience that distribution is heavy tailed meaning it can have quite extreme values compared to the normal distribution. You see that most readily in games where there is a real blowout in the margin. Auckland vs Wellington tonight is a case in point, but there are plenty of cases like that in the Currie Cup. The other side of that is when a team is expected to win by a very large margin but only scrapes in or even loses (Manawatu vs Canterbury).

      Sports prediction is a lesson in understanding variability. My predictions so far this year have been way better than I would normally hope. I can easily have a run of games where the predictions don’t work out. Hardest of all is predicting the NRL where I regularly see results which are hard to believe. I have always held suspicions that the NRL is not clean.

      8 years ago

  • avatar
    david pettit

    This algorithm has predicted numerous upsets(right or wrong)

    Off the top off my head it predicted Otago to beat Welly.

    The win percentage speaks for itself…it’s a great starting point. Of course the system will never predict an upset like Manawatu beating Canterbury but as a foundation it’s very good.

    8 years ago

  • avatar
    david pettit

    For the record Dave, your predictions are awfully close to the bookies lines quite often(highlights how good your methods are)

    Cantebury opened -9.5 favs but closed -12.5
    Tasman were -3.5 favs
    Wellington opened -17.5 favs closed -22.5 favs
    Harbour opened -5.5 favs now -6.5(bop horrid on the road historically)
    Manawatu opened -2.5 favs now -4.5
    Canterbury -5.5 Favs
    Waikato -10.5 favs(think this is very wrong when you factor in recent events)

    8 years ago

  • avatar
    Shayne Cochrane

    This algorithm has clearly stood up much better than the average picker. I’ve somehow lucked my way into a good lead at the top of the picks in the Superbru competition – average pickers are at 70% success – and I’m looking for a way to lazily and (if necessary) cynically hold my lead as we approach the finish line. Thanks, David!

    8 years ago