Loewen: Applying different standings points systems

Much has been written and said over years to disparage the current NHL point system. Unlovingly called the ‘Loser Point’ system, it’s almost impossible to find someone who likes it. From time to time you’ll find someone who is indifferent toward it or who recognizes that it might be a kind of necessary evil, but the vast majority of hockey fans hate it, and hockey analytics folk hate it even more. Many analysts have written about how bad it is, so I won’t add to that, but I would like to present some alternatives and discuss statistically what the implications of using them would be.

First, I’ll outline the various options. Each of these has it’s benefits, a problem that it addresses, and they each have their drawbacks, some more obvious than others.

Current Loser point system (LP) – 2 points for any win, 1 point for a loss in extra time.
It looks like the NHL wanted to get rid of ties and they wanted a point system that was new, but not drastically new. It’s mathematically inconsistent (ie. some games are worth 2 pts, some are worth 3), and it seems to reward teams for keeping a game tied long enough to get to overtime, but none of it’s problems seem to concern management enough to talk about adopting something new.

Straight Wins and Losses (W/L) – 1 point for a win.
The vast majority of other team sports only count wins, so maybe hockey shouldn’t be any different. The luck factor would increase, but teams would probably play harder for a win near the end of the game, and point calculations (clinching playoff spots, etc.) would be much simpler.

Overtime Punisher (OTP) – 2 points for a regulation win, 1 point for an extra time win, no points for any loss.
This system has the same inconsistency problems as the current model, but this one actively discourages overtime, but most fans like overtime, as long as both teams try their hardest during regulation.

2 Point Tie-Breaker (2PTB) – 2 points for a regulation or overtime win, 1 point for a tie (any game that goes to a shootout, no matter what the result), shootout wins are counted as a tie-breaker.
Already, the NHL essentially awards 1.999 points for a shootout win, since they are not counted in the ROW tie-breaker. People love the entertainment value of a shootout, but there are many who question it’s legitimacy as a way to decide a game. For them, it would be more accurate to hear that two teams played a tie game, and one team won the shootout after the game, and this system makes that distinction.

3 Point System (3PG) – 3 points for a regulation win, 2 points for an extra-time win, 1 point for an extra-time loss.
There are 4 possible outcomes in the current model, the NHL just treats a regulation win and an overtime win the same, but if they understand that an overtime loss and a regulation loss deserve different points, then it should also make sense that an overtime win and a regulation win are not equal.

4 Point Sysem (4PG) – 4 points for a regulation by 3 or more goals, 3 for a regulation win by 1 or 2 goals, 2 points for an extra-time win, 1 point for an extra-time loss, and for a regulation loss by 3 or more goals results in 1 point being taken away.
This is easily the most cumbersome, gimmicky and complex alternative, and there would be revolt if it was applied, but it fixes a lot of problems. It provides nuance and it separates the good from the lucky better than anything else. It would ensure that no matter what the score, teams would try their hardest all the time, because nobody could afford to coast. It’s mathematically consistent and it would set the NHL apart for doing something unique. The question is, how badly do we want to fix these problems.

What would change?

We can easily apply these point systems to the results from every year since the 2004-05 season was needlessly cancelled. These adjusted results don’t tell us how the different systems would change how people would think and strategize, but the mathematical results do tell a story. For the psychological changes, we can only speculate.

Standings

The table below shows the results of the 2013-14 NHL season with the actual results used to produce the various different point totals. Teams that would qualify for the playoffs have their point totals printed in red. It’s plain to see that not much would change. With all of these different systems, the same 16 teams qualify for the playoffs 5 out of 6 times. Other seasons produce similar results. (If the 2011-12 season was decided strictly by wins and losses, the cup winning Los Angeles Kings would have finished 10th in the West, not 8th, and would have missed the playoffs, but this is the exception, not the rule.) So, if we would only change point systems because some teams would needlessly getting short-changed out of a playoff spot, then we wouldn’t do it. The standings don’t change.

Loewen 2

The Record Books

Throughout the whole history of the NHL, winning teams have always been given 2 points. If we were all of a sudden to start awarding them more, we would lose our ability to compare their results easily from one season to another. Now it’s easy to calculate how close a team is to .500 and how many games a team has to win to clinch a playoff spot. If those numbers change, that process becomes more difficult for the everyday fan. This is of course a legitimate concern, but the current system has already done that. Almost 25% of games in the 2013-14 NHL seasons went to overtime, which means that an extra point was awarded almost 25% of the time, and team point totals were on average 25% higher than they would have been in the previous 2 point system. The record books have already changed.

Intensity

This might just be an exercise in psychology, but with different point systems, players and coaches would approach the game with a different mindset. Players might be more cutthroat if there was only a win at stake. Players might be tempted to run up the score on a weaker opponents with the 4 point system. Sportsmanship, injuries and entertainment value are all at play and would need to be taken into consideration if a new system were to be adopted, because the intensity would change. Many fans would argue that this has already happened. Currently, there are countless examples of when a game is tied midway through the third period and the speed of the game goes down, the playing that stars get goes down, and intensity goes down.

The spread

Remember when you would write a test and your only hope was that maybe everyone else did bad too so that the professor would be forced to use the bell curve to adjust your scores? I don’t know if it ever happened to me, but I still held out hope. I probably wouldn’t be comforted knowing that the doctor prescribing my medication had his or her grades artificially assisted by a bell curve. Would I, as a fan, be happy to know that my team had their results inflated by a league imposed bell curve effect?

Loewen 1

Using a 21 column bar graph, the distribution of team results from the 2013-14 NHL season is as follows (for the LP and OTP system is used an adjusted points percentage so that the mean was .500).

Besides it being painfully easy to find Buffalo every time, the results bear out quite a few differences. While I believe that this is mathematically significant, I don’t think most fans would notice that the spread would change, except that it means something very exciting would change.

The Playoff Race

As we approach the trade deadline for the 2014-15 NHL season, there is speculation that as many as 11 teams are essentially out of the playoff race, so they will be looking to sell assets and coast the rest of the way. With that many teams having so little to play for, that’s a lot of match-ups that won’t be worth attending. Fans want to see games involving teams that need points to make the playoffs. So, does the point system affect when teams qualify for the playoffs? To demonstrate this, I chose 5 teams and the dates of when their playoff status was decided last season.

Team WL 2PTB LP OTP 3PG 4PG
Ottawa eliminated April 4 April 4 April 8 April 8 April 8 April 9
Calgary eliminated April 1 March 30 March 20 March 29 March 27 April 1
Boston clinches spot March 17 March 23 March 23 March 22 March 23 March 25
St. Louis clinches spot March 17 March 18 March 16 March 16 March 18 March 20
Dallas status decided April 8 April 13 April 11 April 8 April 10 April 9
GP by all status decided teams 187 171 168 193 190 138
% of all games 7.60% 6.95% 6.83% 7.85% 7.72% 5.61%

With the season ending on April 13th, these date ranges show the change that the differnt point systems would have on the playoff race. Essentially, the longer it takes a team to clinch their playoff spot, the more invested their fans are in the process. The dates never vary by more than a week, but in the last few weeks of the season, a few days here or there can make a big difference. The last two rows are a rough calculation of how many games were played by teams whose playoff status had already been decided, what what that looks like as a percentage of the season. I believe this is the key drawing point of the Loser Point system for the NHL executives. The teams that make the playoffs don’t really change, but the excitement of the playoff chase is dragged out just a little bit longer. The 4 point game system provides the best results, because as long as the team in 9th place can theoretically string a bunch of big wins together to finish the season, it’s pretty tough to lock down a playoff spot.

Conclusion

Determining which is the best point system to use requires us to pick which issue is most important (ie. historical continuity, credibility, uniqueness, mathematical consistency, entertainment value) and pick which problem we are willing to live with (more difficult calculations, newness, mathematical inconsistency, gimmickry, etc.) The NHL has chosen, and while you might disagree with their rationale, there doesn’t seem to be any interest in changing, but at least now you know what the alternatives would look like.

4 thoughts on “Loewen: Applying different standings points systems

  1. Pingback: My Take on the NHL’s Flawed Point System | NumberPuck

  2. I’m in favor of 2PTB. It’s nice to know I’m not the only crackpot who thought of that plan. 😉
    I think the biggest problem to resolving this is Bettman refuses to understand his polls are asking “Do you like shootouts?” and not “Do you like shootouts determining who gets points in the standings?”

  3. Pingback: Mean & Median at Midway: part i – point systems | Check Republics

  4. Hi, what if the standings were without points, but with ties (no shootouts), where wins (regulation & OT) were the primary factor? Ties would serve as a *ahem* tie-breaker for teams level on wins. Would you be able to do your playoff status calculation again using this system (it could be called WLT)?

Leave a Reply

Your email address will not be published. Required fields are marked *

*