THE BOOK--Playing The Percentages In Baseball

Like a lot of things, it only confers an advantage if no one else does it, i.e. 123 points won’t be a sure playoff spot once anyone else starts doing this.

Not to mention the fact that teams and players try to win the Stanley Cup for some combination of pride (satisfaction in achieving a difficult task) and monetary gain (via increased attendance for the team, or increased salary for the player), and this sort of thing would confer neither of them…

Incidentally, would this have to be a conspiracy, or could a team do it unilaterally?  I don’t know if the rules forbid stacking players like cordwood in front of their own goal

Of course, there’s an easy fix - give out 3 points for a non-shootout win.  That way the same points are given out for every game.

The point, of course, is not that some team is really going to play an entire season or even an entire game that way, but that the rules give teams incentive to play conservative, boring hockey, especially late in games.

If I remember my NHL history correctly:
The point for an OTL was instituted in the time when there was a short OT but no shootout, so games could still end in a tie. The thought was that teams were playing too conservatively in OT to “preserve their point.” The new rule encouraged teams to play more aggressively in OT, but at the expense of giving them even more incentive to play conservatively toward the end of regulation. All in all, that was probably a wash.

Once they added the shootout following OT, the positive effect of the OTL point went away, while the negative effect remained. It makes no sense to me that it’s still given.

I like the suggestion in #2 a lot, but simply making a win a win and a loss a loss would also work.

The CCHA (the first NCAA conference to adopt the shootout (last year)) this year went to the system proposed by Richard: 3 pts for a win, 2 for a shootout win, and 1 for a shootout loss. 

I have always been puzzled by the NHL system, which makes shootout games worth more in the standings (total) than non-shootout games.  All games are equal seems like a pretty fundamental starting point for league standings.

If going all-out to play like this means 123 points compared to the 70-ish points if you played your typical crappy game.  Then maybe going somewhat toward the all-out will get you somewhat toward the 123.

If Hawerchuk is up for it, maybe he can look at teams in the bottom part of the (regulation) standings, and see what happens when they face each other.  Is there more than the usual number of points per game?  And what happens when two heavyweights face each other?  And two lopsided teams?

Patriot / 4:

Soccer has the same problem. In most leagues, including World Cup play, a game with a winner and a lose has three points (all to the winner). A game that ends in a tie has two points (one apiece, obviously).

The way soccer does it is nice because it gives you a lot of incentive to go for a win if a game is tied late (except in certain cases where you only need a draw to advance).

Tango @5 - interesting question.  I’ll look into that.

Soccer used to do 2 for a win 1 for a draw but because soccer series are usually home and home (over all 20 teams in a league for 38 games a year) teams would play to draw on the road and win at home.  This led to boring soccer, hence a win being worth more than a draw.  In hockey, if you want to encourage teams playing for a win, make a win worth 4, a shootout win -2 and a draw 1.

Wow, thats a total failure in the hockey scoring system.  You should not make teams have an incentive to draw.  It should award at most the same number of points for 2 draws, as for a win and loss.  You shouldnt make there be an incentive to take the game to OT, essentially giving the equivalent of a 75% win rate to a team that goes to OT every game and wins half.

Many systems reward a win as more than 2 draws, like soccer, to avoid the draw incentive, and this works well.

If hockey is going to give 1 point to the loser of an overtime game, if needs to take it from the winner, with a 3/0 system normally and 2/1 in overtime.

I’d like to see teams agree to play games in the way Tango describes, just to expose the scoring system flaw.

The CCHA, mentioned by 4/Patriot above, provides excellent evidence for the negative incentives in this scoring system.

In 2007/08, the CCHA used this system:
Regulation & OT games--2 for the winner, 0 for the loser.
No shootouts after tie games.

In 2008/09, the CCHA instituted shootouts:
Regulation & OT games--2 for the winner, 0 for the loser.
Shootouts--2 for the winner, 1 for the loser.

In 07/08, 39 percent of overtime games resulted in a winner.  In 08/09, 19 percent of overtime games had a winner.

It’s really interesting to look at save percentages--

07/08 regulation:  .903
07/08 overtime:  .913

08/09 regulation:  .911
08/09 overtime:  .950

Not only were teams taking fewer shots in overtime once the scoring rules changed (OT shots per minute were down 18 percent), but the shots were clearly less dangerous shots--from the save percentage, it looks like teams were mostly taking “pot shots” from the blue line, and playing more conservative defense.

Now that they are using their third different scoring system in 3 years, I expect we should see another change in the numbers, back toward the 07/08 levels.

Excellent info, alt_n.  I’ve linked your comment and a couple of other articles - I was going to just comment here but it got too darn long.

Alt/11: fantastic info.  You guys are great…