Anything Can Happen in One Game

As Canadian world junior fans learned yet again yesterday, anything can happen in a one game playoff. By now the more reactionary fans are well into their usual routine of blaming the goalie and/or the coaching staff, but all that really needs to be said is that the format of international tournaments works against the best teams and creates a high degree of randomness. The recent history of world junior and Olympic tournaments speaks to that, with some of the huge upsets that have taken place, together with the oft-repeated story of teams that looked like world-beaters through four or five preliminary games (as Canada did in this year's tournament), before unexpectedly falling to an inferior foe because of one bad outing in an elimination game (and as far as poor games go, it's certainly possible to do much, much worse than outshooting the opposition 56-24).

The international tournament format caused me to think of an interesting hypothetical: What if the NHL postseason was a series of one-and-done showdowns? Obviously it's not possible to replay those postseasons based on that counterfactual, but it is possible to just look up the results for the first game of each playoff series. It's not a given that things necessarily would have turned out the same way if both teams knew it was do-or-die, but it's probably a fairly reasonable approximation.

It turns out that having a single game elimination format would almost completely alter NHL history. Out of the last 22 Stanley Cup champions, only one of them never found themselves trailing 1-0 in a series at any point in their postseason run. The 2008 Detroit Red Wings are the only Cup winner to win all of their series openers in a single playoff season since Edmonton traded Wayne Gretzky. Every other Cup champion since then lost an opening game, meaning that if they were playing a one game series they would have been eliminated and would never have earned the chance to drink from Lord Stanley's mug.

It would have been a similar story, although not quite as extreme, if all series had been best-of-three affairs. In that scenario, the clear majority of winning teams would still probably not have made it all the way through to win as they did. Fifteen out of the 22 teams lost 2 out of the first 3 games in at least one series on their way to a Cup. Even if the format was changed to a best of 5, that would still have a major impact on the final results, as somewhat amazingly half of the eventual champions trailed 3-2 after five games at some point during their Cup run.

In total, based on these assumptions, Cup winning teams since 1988-89 would have won 61.4% of one game playoffs, 76.1% of three game playoffs, and 87.5% of five game playoffs.

That supports the obvious fact that the larger the sample size, the more likely it is for talent to win out over luck. At the world juniors, it's obvious that Canada routinely has the most talent. Over the past five tournaments, Canada has a record of 26-5 with a goal differential of 185-65. Some portion of that is from pounding on the minnow nations, but even against the traditional top five hockey nations (USA, Sweden, Finland, Russia, Czech Republic) Canada still went 15-5. That's a .750 winning percentage, and the goal differential suggests that the team's record was fully earned (the Canadian juniors scored 99 and allowed 56 against the same opponents for a Pythagorean expected winning percentage of an even slightly better .758).

That's an incredibly dominant record, but it still leaves the simple reality that if Canada is a 75% favourite in back-to-back playoff games against two solid opponents, that still leaves them with only about a 56% chance to win any given tournament. Given the talent of some of the American, Russian and Swedish squads in recent years, that's almost certainly overstating the odds of even a truly dominant team getting through two single elimination contests unscathed. The odds drop even further if the team did not secure a quarterfinal bye by finishing first in their pool.

This is not a simple attempt to justify a loss. Winning a single elimination tournament is also less meaningful, for the exact same reasons. Sometimes a weaker team wins, sometimes a good team plays poorly but gets the breaks anyway. The longer the series, the less variance and the more confidence that the better team ends up triumphant. All Canadian hockey fans would like to claim that the 2010 Olympics win proved that Canada is the world's best hockey nation, for example, but that tournament alone is not enough to prove that assertion. A better argument would be to look at Canada's overall record in winning three of the last six best-on-best tournaments, but even that analysis would show that it is relatively close between the top nations. At the end of the day, many Canadian fans need to have more reasonable expectations about how much success to demand from the teenagers representing their country internationally.

The international format is what it is, it's certainly exciting and it's not going away any time soon, but it's still important to not try to draw too much significance from such a short tournament. If you are tempted to do so, just remember that if the same rules applied to the NHL playoffs, a different team would probably end up winning well over 90% of the time.