I'm trying to predict cup winners using regression formulas from regular season statistics
Here is a summary of the results
From 2012-13 to 2017-18 I could have given you six teams each year and one of them would've won the cup. Also, out of those six seasons and those six teams, five times the cup loser would've been included in that list. Pretty damn good, I think.
Keep in mind 12-13 was a shortened lockout season, the cup winner in 13-14 was the 6th best team in the west, and the 14-15 winner was the 4th best team in the west. The other winners for the other seasons were either 1st,2nd,or 3rd in their conference. Usually the best teams with the most points in the regular season are one of the winners, but tell that to the Washington Capitals prior to last season.
The 11-12 season posed some problems as the 6th best team in the East, Devils, lost to the 8th best team in the West, the Kings. Had the Devils won this cup, the system would've continued with 6 teams as listed as one of the winners
The 10-11 season also posed some problems as the Bruins beat the Canucks. The system did not have the Bruins listed in the top 6; but again, the Canucks had they won, would've matched the 6 team system.
The 07-08 and 09-10 seasons included the cup winner from the 6 team list
The 08-09 winner was not in the top 6, but #8
So, hmmm, some thinking. The system was great for the past 6 seasons, the NHL is a bit different now.
Basically as a summary:
I analyzed 11 seasons, with an initial 6 team list to win and/or lose the cup each year before playoffs start
8 times, the winner would've been in the 6 team list
9 times, the loser would've been in the 6 team list
so now the question is, is this performance good enough?
trying to "improve" it further and include the 3 seasons where the cup winner was not in our 6 team list----
I found a category to include the winner from those 3 seasons:
If i include that additional or new category for each of the 11 seasons:
I could have given you a list of 9 teams each seasons and guaranteed you the cup winner was in it every time
Or i could have give you a list of 8 teams and guaranteed the winner was in it 10/11 times
9 teams sounds like a lot out of 16 who make it
I only had cup odds in 2017 after the reg season and before playoffs started, the 6 teams I would've given you
Wash +450
SJ +1500
ANH +750
TB +2000
STL +750
PIT +750
Had you bet each of these teams, (PIT won over Nashville, this is the one year the cup loser wasn't in the list) you would've won 2.5 units. Of course if washington won, you would've lost a half unit.
does anyone have any thoughts?
if someone has pre-playoff cup odds I can provide those results for other seasons