Does Season Series Translate to Predictable Playoff Winners?
When discussing playoff predictions, it is common to hear people bring up the season series as an argument for predicting the winner (the season series is another word for head-to-head records within a season). For example, if team A beats team B three out of four times in a season, team A had the better head-to-head matchup. There are times when teams split the season series with equal wins, but one team won all games in regulation, whereas the opposing team needed extra time for at least one of their wins. Because of this, it is important to analyze points percentage in conjunction with win percentage to determine if the season series record is an important variable. For this analysis, only round one matchups since the salary cap was introduced in the 2005-2006 season will be analyzed. Although it may be interesting in the future to analyze the season series impact in later rounds, the first round gives the best indication, since there are less confounding variables such as injuries from a previous series or length of previous series that can impact the play of a future round.
Since the 2006 post-season, there have been a total of 144 first round playoff series. Of these, 124 had the scenario where one team had a better season series points percentage than their opponent, with the remaining 20 playoff series having an even head-to-head record in terms of points percentage. Of the 124 occurrences of a team having a better head-to-head record than their opponent, that team won 81 times, or 65.3% of the time. Similarly, strictly looking at win percentage, there were 104 series where one team had more wins than their opponent, of which 68 times the team with more wins won the first round series for a win rate of 65.4%, slightly better than looking at points percentage. Although this would support the theory that head-to-head matchups are important, it may also be impacted by the fact that higher seeded teams win more games and lose less games in regulation, and therefore have a better head-to-head record against most teams, as opposed to an indicator of a dominant matchup to that specific team. Because of this, it is also important to take into account the degree to which season series can determine a first round outcome. It is not as telling if one team has a head-to-head point percentage that is better by one loser point, versus a team that won all head-to-head games in regulation and left their opponents with zero points.
When breaking the season series points percentage into sections, we can better grasp if this variable is a good predictor. Beginning with the most extreme scenario is the 100% versus 0%. This means that one team won every game in regulation, which has occurred nine times in the salary cap era. Of these nine teams, two managed to beat their opponent. The 2017-2018 Winnipeg Jets went 0-5-0 against the Minnesota Wild, yet beat them 4-1 in the first round. The 2018-2019 Columbus Blue Jackets not only went 0-3-0 against the Tampa Bay Lightning, but were outscored 17-3 in these three games. Despite this, the Blue Jackets swept the Lightning, who tied the single-season wins record at the time. Of the remaining seven, the 2019-2020 Bruins versus Hurricanes had a one game sample size due to the shortened season, effectively resulting in six of eight teams winning in this scenario, or 75%. Surprisingly, none of the six season series sweeps resulted in any playoff sweeps.
The next section is looking at teams who didn’t win a single head-to-head game, but still achieved loser points, or in other words, had a points percentage greater than zero. This scenario occurred fourteen times, with one occurrence being a one game sample size from the shortened 2019-2020 season, effectively making it a thirteen time occurrence. Only four of the losing teams won the first round series, resulting in a win rate of 30.8% for losing teams. These include the 2016-2017 Rangers beating the Canadiens, the 2017-2018 Lightning beating the Devils, the 2018-2019 Avalanche beating the Flames, and the 2018-2019 Hurricanes beating the Capitals. The lowest points percentage of these teams comes from the 2007-2008 Boston Bruins, with a record of 0-7-1 against the Montreal Canadiens. Although down 3-1 in their first round series, they managed to push it to seven games before ultimately losing. Unlike the 0% head-to-head teams, there were sweeps that occurred in this scenario. The 2013 Sharks swept the season series against the Canucks, and then swept them in the playoffs, as did the 2014-2015 Ducks with the Jets.
The next category of teams are ones that only won less than half of their head-to-head games, or in other words, teams that had a winning percentage less than 50%. This situation occurred 82 times, of which 30 times the team with less than half of the season series wins won the series. This gives teams that have a winning percentage between 0% and 50% a first round success rate of 36.6%. The lowest winning percentage among winners are the 2013-2014 Wild, who won 20% of their games against the Avalanche, but eliminated them in seven games. Similarly, in the 2021 shortened season where teams only play amongst their division, the Winnipeg Jets went 2-7-0 against the Oilers for a 22.2% win rate, only to sweep them in the first round.
The final category of teams to analyze are those who split the season series in wins, but had differing points percentage. For example, if team A went 2-1-1 against team B, but team B went 2-2-0 against team A, they would have an equal winning percentage, but team A would have the better points percentage. This scenario occurred 23 times, of which 15 winners had a better points percentage than their opponent. This gives teams that have an equal winning percentage but a better points percentage a 65.2% winning rate. Although the sample size may be small, this would imply that it is important to compare points percentages among teams with equal head-to-head wins.
Common result is that teams with a better head-to-head record have a first round success rate of approximately 65%.
As seen, season series can be an important factor to consider when making predictions. While the correlation between winning the season series and winning the first round is not strong enough to use this variable on its own, it can still be used as additional guidance to other variables when predicting first round series winners. The head-to-head record of each first round series matchup for the 2024 Stanley Cup Playoffs are as follows:
Surprisingly, despite the rarity of the season series sweep, half of all first round matchups contain one. According to the analysis above, the Canucks and Jets should have a 75% chance to win their series, and the Golden Knights and Bruins should have a 69% chance to win their series. However, this is not necessarily true, especially given the smaller sample size of previous season series sweeps. As is the case with most variables, one alone is not enough to determine the odds of a series winner, and instead season series should be used as a variable within a multiple regression analysis, as it is important to analyze its effect in conjunction with other predictor variables. To get a full in depth analysis on predicting first round series winners, visit the services page for probabilities and additional insights.