Skip to main content

This column was first published in the November and December 2019 issues of The Curling News. I’ve re-upped it for two reasons. First, the World Curling Federation went through significant political maneuvering in the summer of 2019 and decided, at the last moment, not to implement the expected change from 10 ends to eight ends. As such, this debate is set to re-ignite the curling world about a year from now.

The second reason is that this piece serves as a primer for the next episode of the Rock Logic podcast, which drops here at on February 6. Listen in as myself and two accomplished sports data experts dig in.

At the 1924 Olympics, games were 18 ends in length. Based on some of the scores (38-7, 48-6 and 18-10), no concessions likely took place. The 1932 Olympics dropped contests to 16 ends, and then curling was dropped from the Games altogether until demonstration status in 1988.

Following the first Brier in 1927, it was Saskatchewan’s Oswald Barkwell who suggested the games be reduced from 14 ends to 12. The change was made the following year.

Forty-nine years later, at the Brier in Montreal, games were shortened to 10 ends for the Canadian Championship. Prior to 1974, teams were not allowed to concede before a complete game, meaning every Brier team would play a full 120 ends (once Newfoundland entered the competition in 1951). If you’re under the age of fifty, you will have to watch old video of the corn broom era to begin to imagine the physical toll this might take on a front end.

Now the World Curling Federation is evaluating a move to eight ends for all four-player competitions. The ultimate decision will be made after the 2022 Olympic season.

We declared eight ends to be a done deal. The decision was delayed until 2022.

We declared eight ends to be a done deal. The decision was delayed until 2022.

When the Brier moved to 10 ends in 1977 it was largely because of time. At the 1976 Brier, two extra-end games ran beyond five hours. Who knows, perhaps if time clocks had been introduced in the seventies, 12 ends would still be the format today?

Television was a major influence to introduce these clocks and keep games moving. Today, broadcasters again are central to the discussion but it’s not necessarily on the side you might expect. Streaming and pre-recorded shows allow viewers to skip commercial breaks. Live events like sports and the Oscars are able to capture more eyeballs for sponsors in the moment than scripted television. Originally there was a desire for curling to be shorter in order to fit a broadcasting window of less than three hours, but now that may not be the case. Extended airtime from a high stakes 10-end game could conceivably garner more advertising than an eight-end contest.

So why is the WCF considering the move to eight-end games?

There appears to be many reasons and most top players have put their support behind the move. The physical toll has been mentioned by some, though the corn sweeper of yesteryear who battled blisters and lathered on RUB A535 may chuckle at that notion. Keeping playing conditions optimal is also a consideration; longer games can further break down and soften ice in the sliding path, causing the dreaded “fudging” effect. Many have pointed to the Mixed Doubles becoming a breakout hit at the 2018 Olympics and its swift play as a contributing factor in its appeal.

I’m not in the camp of fundamentally changing a sport to benefit those who only watch it every four years. Granted, this would not be a dramatic change, given curling is predominantly played as eight-ends today. Games on the World Curling Tour, at the Grand Slam series and the home game versions at your local club are all the shorter form, and have been for some time … roughly 70% of world ranking games during the 2018-19 season were eight-end contests.

Some argue, however, that if Grand Slam tennis (albeit men’s and not women’s) can be five sets in length rather than three, then surely curling should promote a different, more gruelling test for its highest achievements.

The question remains: Should the sport move to eight-end games for the Olympics and World Championships? For any sport, two of the most influential considerations for any change must be fairness of competition and entertainment value. Sport is often attempting to find a balance between determining the most deserving champion while also aligning with the considerations of fans and sponsors who want to watch and appreciate the spectacle. These two factors are often in competition. The very nature of one final game to determine a champion produces great drama, but it can also be argued that this does not always deliver the most deserving winner.

When Canada’s Macdonald Brier and Lassie Championships were round-robin format only, the team with the best round-robin record was simply declared the winner. Barry Fry of Manitoba won the 1979 Brier with a 10-1 record, while Rick Folk of Saskatchewan and Larry Pineau of Northern Ontario were both two games back at 8-3; there was no playoff format and nothing for CBC to televise on the weekend. To many players at the time, it was considered the most equitable method to crown a champion.

Pat Ryan in Netflix’s “Losers”

Pat Ryan in Netflix’s “Losers”

In 1985, Alberta’s Pat Ryan went 11-0 in the round-robin and lost a dramatic championship final game to the 7-4 Al Hackner squad from Northern Ontario. Perhaps it was “unfair” that Team Alberta’s great performance during the week was overshadowed by a playoff format which forced them to prove their worth one more time against a team they had already beaten. But the result became a Sunday televised event, an historic upset and one of the most memorable moments of curling lore–recently trumpeted in a Netflix sport series–rather than a trophy ceremony on a Friday evening.

No pressure

No one is pressuring Curling Canada to eliminate the playoff system they welcomed in 1980, and it’s usually the case that fan and broadcast interest, or even event scheduling, will usually win favour over fairness of competition.

If a curling game was played as a single end, it would be unfair to the team without hammer, leaving them perhaps a 20% chance at victory. A two, three or four-end game would also be heavily skewed to the team starting with hammer.

If curling was played over an infinite number of ends, the better team, barring injury, would always win and the longer a team is given an opportunity to establish its authority over its opponent, the higher likelihood a victory is the result. A heavy underdog will have a better chance in a game of one single end, especially if they are starting with hammer … and this leads into our first question: Is an eight-end game a fair contest when compared to a 10-end game?

We will only look at results for the 2019-20 season, using the five-rock free guard zone rules exclusively. In all world ranking games from that season, men’s teams starting with hammer (last-rock advantage) won 62% of the time in 10-end games and 60% in eight-end contests. Women’s teams were 58% and 57% respectively. This is perhaps counter to thinking that a team with hammer has an advantage in the shorter game.

These numbers may be affected by better teams starting with hammer more often. If we look at only games between top 30 teams, the numbers are similar: men’s teams enjoy a 64% winning percentage in 10-end games and 63% in eight-end games; for the top 30 women it’s 58% (10-end) and 56% (eight-end). A shorter eight-end game appears to increase the chance of victory for the team starting without hammer, even if it is only a slight difference.

There are other factors at play, but it appears sufficient to conclude there is no significant advantage to the team starting with hammer in an eight-end game versus a 10-end game. One theory in support of 10-end games is it gives teams a better chance to overcome an early deficit, should they get down early. The current numbers do not support this, and teams that fall two or three points behind after one end have a near-equal winning percentage in both eight-end and 10-end games.

It appears that we’ve determined that an eight-end game is a fair contest, now what about its entertainment value? How can we possibly add mathematical precision to either prove or disprove the typical fan claim that eight-end matches lead to fewer close games … which would, of course, supposedly wreck the entertainment value of matches at the Olympic Winter Games and world championships?

Eight ends will probably enlarge the advertisements • Michael Burns-Curling Canada

Eight ends will probably enlarge the advertisements • Michael Burns-Curling Canada

In sport it’s generally assumed close scores and undetermined outcomes keeps viewers from switching channels. In the late ends of a curling game, an engaging contest occurs when the score is tied, a team is down two points with hammer or one team is ahead by a single point. These situations provide a likelihood of one team winning no more than 90 percent of the time and in many cases it’s less than 80 percent.

Let’s examine the theory that eight-end games lead to fewer close contests. When reviewing the data, the opposite is actually the case. For men, based on our definition, 39% of eight-end games made it to the final end as an “entertaining” match. Men’s 10-end games only scored a 33% finish as a close contest. Women’s results were much closer, with entertaining eight-end games edging out entertaining 10-end matches by 40% to 39%.

Close game likely

Looking backwards at each end before the final frame, in every case, a game is more likely to be closer in the eight-end format. The numbers are more significant in the men’s results with each end further from the last widening the gap between eight and 10 ends. For example, with four ends remaining in an eight-end game, it’s likely to be close 56% of the time, while a 10-end game is only close 45% at the same point (beginning with the sixth end). Women’s results are closer with eight-end games ahead 55% to 50% at the same point in a match.

The longer, 10-end version of curling provides more opportunity for the better team to get ahead and increase their lead, hopefully (for them) getting an early handshake. Traditionally and intuitively, curling fans think the longer game must provide more opportunity for a trailing team to mount a comeback, but the results indicate it doesn’t happen enough to make a measurable difference. What this data helps us understand is that an eight-end game is more likely to be entertaining to the viewer. Again, this assumes “entertaining” is defined as a close game in the later stages. Digging even further, tied and one down with hammer scenarios are closer and arguably more entertaining. Examining these results produces similar findings.

As a fan and someone who covers the sport, I’ve often bunked with the undecided camp. Initially my thoughts were torn. On one hand, the older 10-end version of the sport was marred with no practice time and uncertain ice conditions. Teams commonly hit everything for the first two ends and “played them up-and-down.” Assuming two blank ends to start the match, this essentially creates an eight-end game under today’s conditions. However, having watched a healthy number of high-performance eight-end games, I couldn’t help feeling that a 10-end game felt more … substantial?

Looking at the data disputes my gut reaction, but it was hard to shake the idea the longer version seemed more important. Perhaps this gets skewed by the event itself (a Grand Slam game in October versus a Brier draw in March) and my own feelings about the history of certain events. If those Brier and Olympic contests were also eight-end games, perhaps my apprehension would disappear.

Having parked my instincts and investigated the data, I’m leaning heavily towards the eight-end side these days, and the numbers currently tell me there is no glaring reason not to proceed. But there’s another item to consider. A single eight-end game may be fair and entertaining, but does a full event of eight-end games determine a worthy champion?

Intensity cranked

Two more ends per game, spread over an entire week, is a potential 25 percent increase in playing time. This is not an insignificant number considering a long, gruelling tournament with the intensity meter cranked up and everything at stake. The physical and mental impact cannot be underestimated. As just one example, the late stages of the men’s final at the 2017 Canadian Olympic Trials was intense. The anxiety of both teams was felt by the live audience and was tangible within the arena. Two inexplicable mistakes on second stones by each team in the ninth end looked like nerves, perhaps the result of being emotionally pushed to the brink of mental exhaustion after a long week.

Sport, particularly those considered Olympic-calibre, is meant to fully test an athlete or team. It seems logical that an event of 10-end games provides a stiffer obstacle than one filled with eight-end games.

Kevin Koe wins the 2019 Brier • Michael Burns-Curling Canada

Kevin Koe wins the 2019 Brier • Michael Burns-Curling Canada

It’s interesting to note that a player such as Kevin Koe has thrived in large arenas under a brighter spotlight when the prize is wearing a Maple Leaf. Koe has won four titles and lost one final in five trips to the Brier. He has won one Olympic Trials in three attempts. In that same span over the past decade, Koe has won only four Grand Slam crowns out of a potential 53 (!) for a conversion rate of just 7.5 percent. There are several factors to these results, but one might be a greater ability to persevere under challenging conditions when the pressure is the highest, and the games are longer.

When considering the magnitude of an achievement in sport and deciding if it was earned, it is the opinion of competitors, media, administration and fans that will bring final judgement. I expect most athletes would agree an Olympic gold medal earned through eight-end team contests would garner the same respect as those won via the 10-end format.

Now ... what does everyone else think?