Does Corsi For %, 5 on 5, Close predict Playoff Success ?

[HamiltonNHL]

Last year's Playoffs:

The Kings won both the Stanley Cup and the Corsi For %, 5 on 5, Close.

Was it a good year for the "Pro-Analytics" Crowd ?

Que Pense Tu ?

Click to read more...

 

[ThirdManIn]

We have this nifty sub-forum for number crunching, and it is frequented by several intelligent people who know how to do things like math and, you know, more math.

Moved

 

[Jason MacIsaac]

The Last 20 games adjusted fenwick seemed to work the year NJ and LA went to the finals.

 

[HamiltonNHL]

We have this nifty sub-forum for number crunching, and it is frequented by several intelligent people who know how to do things like math and, you know, more math. ..... Moved

Cool.
Thanks.

That being said, it's so non-crunchy it might be "better suited" to the Masses in the \NHL\ forum

 

[ThirdManIn]

Cool.
Thanks.

That being said, it's so non-crunchy it might be "better suited" to the Masses in the \NHL\ forum

Maybe, but I feel like the question you are asking will be better answered here. Then again, I don't know how to multiply numbers with more than two digits, so maybe I'm wrong

 

[Ohashi_Jouzu*]

Was it a good year for the "Pro-Analytics" Crowd ?

Depends on what you think about the 4th, 8th, and 10th place teams not even making the playoffs, while the 5th, 7th, and 10th went out in the first round, and the 26th team knocked off the 3rd.

Did anyone need Fenwick to figure out if LA and Chicago were going to be good teams in the playoffs?

 

[GKJ]

Fenwick Close 5v5 is used more freqently, especially using only post-trade deadline results

 

[hatterson]

To answer the general question, yes corsi for% 5v5 can be used to predict playoff success. Fenwick close 5v5 is more accurate.

However, "predict" in this case is used loosely. A 7 game playoff series is such a small sample size that you still have "underdogs" upset the "favorites" fairly regularly.

It's very easy for a goalie to go on a crazy hot streak (or go on a terrible streak) and then most analytics are out the window. Having a few random bounces go your way can change a series.

Last year the Kings were in 3 game 7s, one of them in OT. I'm not saying it's random that they won all 3, but if they played 3 games 7s every single year, you'd expect to see them eliminated a majority of the time.

Higher fenwick close generally means you're a better team, and being a better team generally means you win more, but it doesn't guarantee anything, especially in the playoffs.

 

[HamiltonNHL]

However, "predict" in this case is used loosely. A 7 game playoff series is such a small sample size that you still have "underdogs" upset the "favorites" fairly regularly.

So true !

 

[HamiltonNHL]

Higher fenwick close

Is this the stat you mean ?

FF% = Fenwick For percentage (Close)

http://stats.hockeyanalysis.com/teamstats.php?disp=1&db=201415&sit=5v5close&sort=FFPCT&sortdir=DESC

 

[HamiltonNHL]

Sister thread:
http://hfboards.mandatory.com/showthread.php?p=94277769

 

[hatterson]

Is this the stat you mean ?

FF% = Fenwick For percentage (Close)

http://stats.hockeyanalysis.com/teamstats.php?disp=1&db=201415&sit=5v5close&sort=FFPCT&sortdir=DESC

Yep, that's the one.

 

[Hammer Time]

Fenwick adjusted for score effects (i.e. Fenwick results when up a goal, down a goal, etc. are weighed to account for the effects of a team playing more conservatively when in the lead) is generally considered even better. Eric Tulsky did a lot of work on this. Puckon.net is a site which tracks score effects-adjusted Fenwick (you can go to the "About" tab for formula and rationale). Here is last year's results: http://puckon.net/fenwick.php?s=2013-09-01&e=2014-04-30&f=1&ld=1&l=82

 

[HamiltonNHL]

I've been using Corsi For % (5v5, close) for certain analyses. Should I switch to Fenwick% (5v5, close) ?

 

[Kramerica Industries]

Those two are generally similar; the only thing Fenwick does is eschew blocked shots from the events tally. Personally, I prefer Fenwick, but you aren't doing it wrong, comparatively, by using the Corsi equivalent.

 

[lomiller1]

I've been using Corsi For % (5v5, close) for certain analyses. Should I switch to Fenwick% (5v5, close) ?

Most people are now leaning towards score adjusted metrics instead of close. Score Adjusted Corsi seems to slightly outperform Score adjusted Fenwick so far but this could still continue to evolve.

 

[HamiltonNHL]

This year's Stanley Cup is
#2 vs #4.

 

[Ohashi_Jouzu*]

This year's Stanley Cup is
#2 vs #4.

Who just played #15 and #19, respectively, in the conference finals.

 

[Bps21*]

Who just played #15 and #19, respectively, in the conference finals.

Who won? The team that carried possession or the middle of the pack opponent?

 

[tinyzombies]

I've read that scoring chances differentials, score adjusted (War On Ice) is the best predictor of a series win. And I assume high danger chances even more so. But it didn't work out that way in the Hawks-Ducks series. Ducks held an 81-62 edge in score adjusted high danger chances over the Hawks but didn't get it done.

Scoring chances total, score adjusted was closer: 180.8 - 171.2 in favor of Anaheim.

 

[hatterson]

Who won? The team that carried possession or the middle of the pack opponent?

If your point was dependent upon the outcome of two game 7s, it's a bad point.

 

[Michael Farkas]

Who won? The team that carried possession or the middle of the pack opponent?

That's a tiny bit specious, no?

I mean, on that list there...of the top 12 teams in the league...

5 failed to even qualify for the playoffs
5 failed to win a series (going, what, 9-20 in the process; 1 lost to 15, 5 lost to 19, 6 lost to 4, 7 lost to 2, 8 lost to 13)
Then of course the two in the show...

But six total series wins among the top 12 teams, four total series wins among the bottom 12 teams...

It's neat and all, but is it any more than what regular shots on goal could have "predicted"...?

SOG regular season 2014-15:
1. Chicago
2. Islanders
3. Nashville
4. Pittsburgh
5. San Jose
6. Rangers
7. Dallas
8. Boston
9. Ottawa
10. Los Angeles
11. St. Louis
12. Minnesota

4 failed to qualify
5 failed to win a series
1 of the finalists, 2 of the Conf. finalists

19. Detroit
20. Washington
21. Philadelphia
22. Toronto
23. Arizona
24. Columbus
25. Montreal
26. Edmonton
27. Colorado
28. Calgary
29. New Jersey
30. Buffalo

This time it's six series wins at the top (again) to three...

Same basic deal...no?

 

[Mickey Marner]

^^^
Or goal differential.
NYR-1st
TB-2nd
CHI-4TH
ANA-17th

 

[Doctor No]

^^^
Or goal differential.
NYR-1st
TB-2nd
CHI-4TH
ANA-17th

Schedule-adjusted goal differential has been doing decently well this playoffs. See the end of this thread: http://hfboards.mandatory.com/showthread.php?t=1865419

10-4 (two better than picking higher seeds), with half of the losses due to stupid sexy Anaheim.

 

[tinyzombies]

Round 1:
Scoring Chances For %, Score Adjusted: 1-7
Fenwick For %, Sc Adj: 2-6
Corsi For %, Sc Adj: 2-6
Goals For %, Sc Adj: 3-5
(tho both Van and Cgy had terrible numbers, so you could throw that one out maybe; Pittsburgh and Islanders had injuries so they weren't the same teams)

Round 2:
Scoring Chances For %, Score Adjusted: 2-2
Fenwick For %, Sc Adj: 3-1
Corsi For %, Sc Adj: 3-1
Goals For %, Sc Adj: 4-0

Round 3:
Scoring Chances For %, Score Adjusted: 2-0
Fenwick For %, Sc Adj: 2-0
Corsi For %, Sc Adj: 2-0
Goals For %, Sc Adj: 2-0

Finals:
Tampa holds an edge in all four categories, including a huge edge in SCF%, Sc-Adj.

They should probably adjust for playoff experience at key positions, size, adjusted save percentage and injuries.