Quick convo-> what if tie, specifically in SCF

[Robin Hood]

FTR, my preffered method of breaking a tie would be total stat differential, with the obvious exception of SHTOI (or if we do use SHTOI, round up to the nearest minute).

That is a fucking awesome idea. Check the season totals and standardize each category, then sum the differences !!!! this is the obvious playoff-series tie breaker.

that's a terrible idea LOL. the flaws are so obvious.

[MSP4LYFE]

its cute how you and nick have little gossiping acronyms like women do

Are you suggesting that only women use acronyms? LOLSIAFR

[Nick]

home ice advantage is a HUGEEEEEEEE factor.

You should check the #, because it really is not, at least not in hockey.

i guess it's close to the relevance of a tie-breaker in our scoring format then

How is that logical? Pleas remember we're not talking about this year anymore, but just the best option for tie-breakers in the playoffs... why is home-ice a determining factor? It seriously has NO statistical or perceived advantage in hockey, especially not in playoffs.

that's a terrible idea LOL. the flaws are so obvious.

List a few for me? I'm struggling to see them

[Robin Hood]

its cute how you and nick have little gossiping acronyms like women do

Are you suggesting that only women use acronyms? LOLSIAFR

No I'm suggesting only women gossip like you and nick. What do you use to clean that vagina of yours?

[MSP4LYFE]

that's a terrible idea LOL. the flaws are so obvious.

Of course you would find the flaws in stat differential, but fail to see any in breaking ties by seeding...

[MSP4LYFE]

No I'm suggesting only women gossip like you and nick. What do you use to clean that vagina of yours?

Shall I copy paste the definition of gossip for you?

To answer your question though, I use a cricket bat.

[Nick]

its cute how you and nick have little gossiping acronyms like women do

Are you suggesting that only women use acronyms? LOLSIAFR

No I'm suggesting only women gossip like you and nick. What do you use to clean that vagina of yours?

Shiv's argument falls apart, he switches to personal insults and overstatements... Try just having a discussion dude.

[MSP4LYFE]

Shiv's argument falls apart, he switches to personal insults and overstatements... Try just having a discussion dude.

Meh, I excel at both

[Robin Hood]

List a few for me? I'm struggling to see them

Every stat has a certain range within which a team performs every week. I.e. it is ridiculous to suggest that a team will total 50 goals in a week just like it is ridiculous to suggest that a team will total 20 FOW in a week (with apologies to teams who this is true for lol).

Given that each stats range is different, you cannot use statistical differential to break a tie. It isn't logical.

You teach stats nick, the standard deviation for each statistic is different. To total them up to come up with a cumulative differential is mathematically silly.

And this excludes the fact that FOW/SHTOI have HUGE ranges compared to every other stat and you cannot exclude them when breaking a tie.

[Robin Hood]

that's a terrible idea LOL. the flaws are so obvious.

Of course you would find the flaws in stat differential, but fail to see any in breaking ties by seeding...

You wouldn't understand the math so I won't even bother. Hopefully Nick can break it down to you one day once I explain it to him.

[MSP4LYFE]

there is a very obvious answer. #1 seeds have home ice for the conference final. thats why.

You finished higher than me in the standings, and I was a superior team through the course of a season. Not all seeds are as cut and dry as Scott vs everyone else. Unbalanced scheduling and divisional strength all play into seeding, unless the league mandates a completely balanced conference and division, as well as a perfectly balanced schedule, it is foolish to determine tie breaks by seeding, especially between cross conference teams.

everything you just said points to the fact that SCHEDULING is the issue. not the Tiebreak. The Tiebreak is a perfect system - > higher seeds have an advantage.

If better teams are getting worse seeds then 1) you would beat me in a matchup anyways and 2) the scheduling needs to be addressed. you guys are trying to address the Effect and not the Cause.

Actually it is extremely flawed, the whole system is reliant on the schedule which can never be perfectly balanced because of the number of teams and weeks in a season. My reference to the perfect schedule was sarcastic, though I suppose that would be hard to pick up the way I phrased it.

[Robin Hood]

anyways ladies i gotta run out. good luck with this.

[kyuss]

why is home-ice a determining factor? It seriously has NO statistical or perceived advantage in hockey, especially not in playoffs.

"Of the NHL's 129 Game 7s to date, the home team holds an overall record of 80-49 (.620) with a 28-19 (.595) edge in the opening round."
http://www.puckreport.com/2009/04/nhl-p ... story.html


P.S. WKEGU

[MSP4LYFE]

that's a terrible idea LOL. the flaws are so obvious.

Of course you would find the flaws in stat differential, but fail to see any in breaking ties by seeding...

You wouldn't understand the math so I won't even bother. Hopefully Nick can break it down to you one day once I explain it to him.

You're right, I probably wouldn't understand the math, but it isn't difficult to explain the flaws of a stat differential based system in laymen terms.

[Nick]

List a few for me? I'm struggling to see them

Every stat has a certain range within which a team performs every week. I.e. it is ridiculous to suggest that a team will total 50 goals in a week just like it is ridiculous to suggest that a team will total 20 FOW in a week (with apologies to teams who this is true for lol).

Given that each stats range is different, you cannot use statistical differential to break a tie. It isn't logical.

You teach stats nick, the standard deviation for each statistic is different. To total them up to come up with a cumulative differential is mathematically silly.

And this excludes the fact that FOW/SHTOI have HUGE ranges compared to every other stat and you cannot exclude them when breaking a tie.

Ummm...please read my post... you've taken stats...standardize each stat means what again?

[Nick]

why is home-ice a determining factor? It seriously has NO statistical or perceived advantage in hockey, especially not in playoffs.

"Of the NHL's 129 Game 7s to date, the home team holds an overall record of 80-49 (.620) with a 28-19 (.595) edge in the opening round."
http://www.puckreport.com/2009/04/nhl-p ... story.html


P.S. WKEGU

Sorry, that's with no control for the superior team... remove the first 2 rounds and exam that stat again

[MSP4LYFE]

List a few for me? I'm struggling to see them

Every stat has a certain range within which a team performs every week. I.e. it is ridiculous to suggest that a team will total 50 goals in a week just like it is ridiculous to suggest that a team will total 20 FOW in a week (with apologies to teams who this is true for lol).

Given that each stats range is different, you cannot use statistical differential to break a tie. It isn't logical.

You teach stats nick, the standard deviation for each statistic is different. To total them up to come up with a cumulative differential is mathematically silly.

And this excludes the fact that FOW/SHTOI have HUGE ranges compared to every other stat and you cannot exclude them when breaking a tie.

Under the proposed new system, every team will have the same number of faceoff taking forwards (or a maximum of one less), so the ranges between those stats are not as silly as you suggest, ditto with SHTOI, every team has 20 players. There is also the option of calculating the single week totals in which a tie occurs.

[Shep]

HSYASFG

[Nick]

The standard score z = x-mew divided by sigma (x-u )/ s

where:
x is a raw score to be standardized;
u (mew) is the mean of the population;
s (sigma) is the standard deviation of the population.

This is a 100% accurate normalization, we can do this because we know the population statistics. Sigma requires some leg work, but it's not really different from what I do in a normal day anyways.

Pretty nice concept actually. This proposal has the tie breaker being the magnitude of the categorical differences, still purely based on the actual week of play!

[Robin Hood]

The standard score z = x-mew divided by sigma (x-u )/ s

where:
x is a raw score to be standardized;
u (mew) is the mean of the population;
s (sigma) is the standard deviation of the population.

This is a 100% accurate normalization, we can do this because we know the population statistics. Sigma requires some leg work, but it's not really different from what I do in a normal day anyways.

Pretty nice concept actually. This proposal has the tie breaker being the magnitude of the categorical differences, still purely based on the actual week of play!

Sorry I was out.

I thought this was pretty obvious. The formula you listed above is pretty basic. Standardizing the whole statistical population for all the categories is oxymoronic. You want to use it for every stat in case of tie break and then determine which team was better overall. It makes sense in principle.

But there is one fatal flaw:

It can be possible for example, for a team to lose a week 6-8-2, yet STILL WIN if all the stats are standardized because the categories that team wins are by a far greater margin. But this is H2H. A win is a win.

FOR EXAMPLE: Assume an H2H league uses 4 categories:

Team A: Goals = 20, Assists = 25, Points = 45, FOW= 5
Team B: Goals = 19, Assists = 25, Points = 44, FOW = 100

According to H2H, Team A wins 2-1-1. If we standardize these scores for the whole league, however, Team B will win.

How can you use a system to determine the outcome of a tie if that same system picks a different winner than our scoring system in certain circumstances?

Using just goals or heck even less GP is better than that.

Are you sure you teach stats?