The Curling News Blog

Bad math: Curling statistics are wrong

[Håvard Vad Petersson is the longtime lead for Team Norway and skip Thomas Ulsrud, currently battling at the 2013 world men’s championship in Victoria, B.C. This opinion column appeared in the February 2013 issue of The Curling News, and has been republished here – by popular demand – for all to read, not just TCN subscribers]

The author in action at ECC 2012

By Håvard Vad Petersson

For ’spiel after ’spiel, and championship after championship, the CurlingZone or Curl IT or Canadian Curling Association or CCT/WCT crews are doing an amazing job keeping fans updated on the results of curling competitions across the globe.

At some of these events, they also provide statistics telling fans and the media how each player is performing.

After being in the business for quite a while now, I would like to tell the curling world something: these numbers are more of a lottery than a fact.

I’ll admit that the numbers have been close to fair a couple of times in my six years of championship curling, but at least 90 per cent of the time they are way off.

I’ve seen teammates outcurl their opposition, but the stats have suggested otherwise.

I’ve seen teammates who were outcurled by their opposition, but those stats have suggested otherwise.

I’ve received texts from friends and family congratulating me on a flawless game, when the truth is that I was disappointed with my performance.

I’ve come off the ice expecting numbers in the 90s, only to be surprised by a 67.

I’ve seen our national coach’s numbers differ from the official stats by up to 25 per cent! Who is right?

Most fans don’t understand “Efficiency” stats anyway: do you?

I stopped looking at these numbers ages ago. If my team is playing well, that’s all I really need to know. But when I see games on television and these numbers show up on the screen in the sixth end, and the commentators use them to tell the viewers how the players are performing, I can’t help getting upset… because the numbers are most likely wrong.

Has anyone heard stories about young curlers getting upset and wanting to quit the game because their stats are low? I have.

My Canadian friends tell me that some provinces – like Ontario – dropped all statistics services for their major championships years ago due to “negative player feedback.”

Here’s one clear example of how the stats volunteers – and they are mostly volunteers, bless their hearts – are missing the boat completely.

On swingy ice, more and more teams (European teams, at least) are drawing their first stone to the tee line or even to the back of the four-foot when leading by three points or more. This is to avoid the come-around behind a rock in the front of the tee line. No one is afraid of a freeze anymore.

Will the statisticians ever pick this up? I’m starting to think the answer is no.

Luckily for us, the games are still being decided by the team that gets the most points on the board. Thank God for that.

Because the percentages are most likely wrong.

[WCF photo (top) copyright ® by Richard Gray – click on images to increase viewing size]

[If you like this story, subscribe to The Curling News – each print issue is packed with interesting and informative curling news and features, just like this one!]

VN:F [1.9.20_1166]
Rating: 4.8/5 (15 votes cast)
Bad math: Curling statistics are wrong, 4.8 out of 5 based on 15 ratings
Be Sociable, Share!

7 Comments »

  1. Adam Said,

    April 3, 2013 @ 12:52 pm

    Ok Geniuses. Go back and rewatch this game and score it yourself and tell us how you would have marked it.

    VA:F [1.9.20_1166]
    Rating: 1.7/5 (6 votes cast)
  2. John Murphy Said,

    April 3, 2013 @ 2:41 pm

    I suggest that experienced stats people are quite capable of determining the called shot. Not only is the actual call a clue, but the amount of sweeping and reaction of the players is as much of a clue as the actual call. I have used statistics as a coaching tool for 42 years and “player feedback” is irrelavant.

    VA:F [1.9.20_1166]
    Rating: 1.8/5 (5 votes cast)
  3. Harold Said,

    April 3, 2013 @ 3:21 pm

    Of course.. Who would admit they have been wrong for 42 years?
    “Player feedback” is irrelevant.. A bit arrogant I would say.

    VA:F [1.9.20_1166]
    Rating: 4.0/5 (5 votes cast)
  4. Craig Said,

    April 3, 2013 @ 4:44 pm

    The stats system used is horribly biased. Also, possibly greatly inflated due to the fact that there is a bonus point but no negative point to use as a penalty for a shot that hurts a team.

    VA:F [1.9.20_1166]
    Rating: 0.0/5 (0 votes cast)
  5. Tom Said,

    April 3, 2013 @ 10:38 pm

    One of the problems with the stats is that the scorers don’t really understand the game. In 1993 in Geneva at the World Championships, we were playing David Smith of Scotland. In the 6th end we were slightly ahead but in trouble. I made an around-the-horn triple and rolled out the front of the house. That shot salvaged the end and eventually the game. After the game, looking at the percentages, I found I was given 2 out of 4 for that shot. Three years later I saw David in Hamilton and asked him if he remembered me. His reply was that he didn’t forget people who made triples against him. Now I ask you, how can a shot that he remembers three years later only rate 50%?
    So you can see that the percentages are really subjective.

    VA:F [1.9.20_1166]
    Rating: 4.3/5 (3 votes cast)
  6. Jimbo Said,

    April 12, 2013 @ 9:47 pm

    I know that one statistic that needs to be updated is the “tick” as it affects leads.

    VA:F [1.9.20_1166]
    Rating: 0.0/5 (0 votes cast)
  7. Tyler Said,

    September 24, 2013 @ 11:40 am

    Tom, while your story gives a great example of when stats are misapplied, I think we should look at when and where it was: Geneva, Switzerland in the early 90’s. Despite the average to decent results of the Swiss leading into those world’s, curling was a relatively obscure sport to the rest of the world at the time and to find people willing to volunteer their time to score curling was probably quite difficult, never mind finding competent people who have a good understanding of the game.

    It has completely changed since then. My mother-in-law has recently been named to do stats at the upcoming Sochi Olympics. The time investment required by her would make most people cringe at the thought of doing it. She has scored multiple Brier, Scotties and World championships, played at a high level herself (winning a provincial title not too long ago), and went through an extensive interview process. Once she was accepted, she was then told to go a score the most recent world junior’s in Sochi (another close to 2 weeks vacation/unpaid time) and finally once the Olympics role around, she’ll have to take more time off. All this time commitment has eliminated most of the “fly by nighters” from the stats world at a high level.

    Oh, and for the comment about the bonus point, I agree, if you can give bonuses, you should be able to take away marks. Having done a little scoring myself, the bonus point is intended to only be used once, maybe twice in an entire game. It’s meant to be a game changing/saving shot and have high difficulty associated with it.

    VA:F [1.9.20_1166]
    Rating: 5.0/5 (1 vote cast)

Leave a Comment