Wild Card Race (Western Conference)

[Blaster86]

Originally Posted by terminator

Please guys stop, my heart can't take the paradigm shifts.

The Flames' players have a chance to possibly maybe make, if everything goes well, an event that occurs after the conclusion of the regular season.

[dino7c]

Originally Posted by Resolute 14

He's also not oversimplifying by basically arguing "that model is wrong, period, because I don't like it", which is what you are doing.

its wrong because its wrong in this case, overall? I honestly don't care picking most of the playoff teams is not hard...I could have picked most last summer.

saying that LA has a 70% chance to make it when they are 2 points back and don't have the tie breaker doesn't make any sense

not one person is willing to put their money where their mouth is with those odds and that says it all

[Street Pharmacist]

Originally Posted by dino7c

he said he personally thinks the chances are 50/50

I said the model is accurate based on what history it measures. It's accuracy has been tested to be more predictive than simple win/loss records through out at least most of the season. So it's objectively accurate within its parameters.

Sportsclubstats uses strict home/loss records in a "Monte Carlo" method by running millions of simulations which at this point may be just effective, more, or less, we don't know.

Regardless of what happens from here on out, the case of one team's fortunes aren't going to prove anything. I do believe there are distinct exceptions to the power of possession merics in making the playoffs (being in the top 16 of 30 teams). I believe shot/passing lane blocking and accurate shooting can get a bad possession team to the middle of the pack, but not elite. We know goaltending can do that, but the skill of good goaltending appears to be streaky and only a few goalies can do it reliably enough to overcome bad possession. I don't have the data to back that up except an n of 1, so I'm not going to pretend to be smarter than the guys making the predictive chart.

I hope that clarifies my stance

[dino7c]

Originally Posted by polak

Great. Doesn't change the point. If you only win one more two more games then you lose in 10 games, other teams will probably keep up.

6-2-2 over 10 games is a very good record

115 point pace not many teams keep up to that

[Street Pharmacist]

Originally Posted by bizaro86

You should take bets where the odds are in your favour, even if they are uncertain. It's like corsi's signature, I would take that bet (45% on a coin flip) even if you only offered it to me once. There are a number of theories on how much to bet in circumstances like that, the most famous is the kelly criterion.

Only if you can be certain of the odds, and are willing to bet multiple times. Neither of which I am.

[Weitz]

Lets say LA wins tonight and Calgary loses in regulation.

You now have both teams ties at 84 points, with LA playing one less game. That is what would really frighten me. The game in hand really helps this late in the season.

[ricardodw]

Ironic that after years of righteous whining about being good enough to make the playoffs in the East under the new format the Flames would likely be out of the playoffs without their softer schedule against the weak sisters of the Pacific.

The Flames are under .500 against the teams from the Central 6-7-1. Vancouver is 7-7 against the Central.

Flames were helped (along with all the teams in the Pacific) having Arz AND Edmonton for 9 games.

so far 4 wins against Edmonton 3 wins from Arizona. (total 7-0-0) with a game remaining against each.

If they were in the Central these games would be replaced with games with Dallas and Colorado who are the weak teams of the central. The Flames are 1-2 with 3 games remaining.

[Igster]

Originally Posted by Weitz

Lets say LA wins tonight and Calgary loses in regulation.

You now have both teams ties at 84 points, with LA playing one less game. That is what would really frighten me. The game in hand really helps this late in the season.

Only matters if they win the game in hand. Otherwise, it means nothing.

[Weitz]

Originally Posted by Igster

Only matters if they win the game in hand. Otherwise, it means nothing.

Which is fair. But it still is an extra chance.

At the end of the day you have to play hard and not mail in any games or lose games you shouldn't. Take care of business.

[troutman]

Regulation losses are really bad for LA now. Not much room for error.

http://www.sportsclubstats.com/NHL/Western.html

[Enoch Root]

Originally Posted by bizaro86

You should take bets where the odds are in your favour, even if they are uncertain. It's like corsi's signature, I would take that bet (45% on a coin flip) even if you only offered it to me once. There are a number of theories on how much to bet in circumstances like that, the most famous is the kelly criterion.

Okay then. The odds are 176:100.

I'll give you 165:100

You take LA and the odds in your favour. I take the Flames. Done?

[CorsiHockeyLeague]

^Did you miss the part where the chart does not take into account other things that you'd want to correct for if you were setting odds on a bet?

Originally Posted by bizaro86

You should take bets where the odds are in your favour, even if they are uncertain. It's like corsi's signature, I would take that bet (45% on a coin flip) even if you only offered it to me once. There are a number of theories on how much to bet in circumstances like that, the most famous is the kelly criterion.

Yeah, and I'd probably take the "will the Kings make it" bet if you offered it to me at even odds in person and the stakes were within my risk tolerance. But I wouldn't be shocked to lose or anything.

The only bet I've made recently was on the SCF last year and it was a free roll (Lundqvist for Conn Smythe hedged by Kings to win).

Originally Posted by Buster

"Large error bars"....translation: "the data irrelevant".

This is just completely and utterly... wrong.

[Enoch Root]

Originally Posted by CorsiHockeyLeague

^Did you miss the part where the chart does not take into account other things that you'd want to correct for if you were setting odds on a bet?

I assume this is directed at me?

So your stance, now that it's time to stand behind the odds from the chart (that you have defended for weeks) is that it doesn't account for other things that you would want to correct for?

[CorsiHockeyLeague]

I've always noted that it doesn't account for certain things that you'd want to correct for. You've read, and responded to, my posts saying precisely that. Are you amnesiac or just dishonest?

[Enoch Root]

Originally Posted by CorsiHockeyLeague

I've always noted that it doesn't account for certain things that you'd want to correct for. You've read, and responded to, my posts saying precisely that. Are you amnesiac or just dishonest?

No model accounts for all things (well, no model except reality). It is what it is and it has tabulated odds based on its defined criteria.

Either you respect the model, and by extension, the odds that it presents, or you do not.

I expected more from you though, than a pathetic, personal attack.

[polak]

Games at Hand suck for comparison purposes but they're meaningless until you win.

[Textcritic]

Originally Posted by CorsiHockeyLeague

I've always noted that it doesn't account for certain things that you'd want to correct for. You've read, and responded to, my posts saying precisely that. Are you amnesiac or just dishonest?

So, what is the chart communicating then, if it is full of noise, and if there are things that need to be corrected for? Is it practically useful for something?

[Street Pharmacist]

Originally Posted by Textcritic

So, what is the chart communicating then, if it is full of noise, and if there are things that need to be corrected for? Is it practically useful for something?

Predicting the future? No method as of yet has been shown to be better

[Textcritic]

Originally Posted by Street Pharmacist

Predicting the future? No method as of yet has been shown to be better

No method better than this specific chart? Really??

[CorsiHockeyLeague]

Originally Posted by Enoch Root

I expected more from you though, than a pathetic, personal attack.

You're right; it wasn't a serious question. I don't actually think you forgot that I've consistently pointed out that things like goaltending, special teams play, players with higher or lower than average finishing ability and injuries (not to mention random variance) can lead to different results than possession stats would lead one to expect. You've demonstrated you're capable of reading and understanding my posts. So the fact that you would act as though I'd never said anything of the sort yields the obvious conclusion that you're doing so for rhetorical effect, rather than honestly responding to my actual views.

Originally Posted by Textcritic

So, what is the chart communicating then, if it is full of noise, and if there are things that need to be corrected for? Is it practically useful for something?

Better predictor of results than any other available method that McCurdy tried using and certainly better than intuition over the long term. Over the short term, the things that aren't taken account of there - including just random puck bounces - play a larger role in determining results. Hence, if you're playing fifty games between the Blackhawks and the Sabres, the methodology used in the chart would lead you to pick the Hawks, and being wrong in that prediction is far less likely than it would be if you were to pick the Hawks to win one game over the Sabres. That's an oversimplification but it demonstrates why, as we get closer to the finish line, the "best guess" provided is more prone to error.