Post-Game Talk: Uncomfortably comfortable win

[Ol' Jase]

.

 

[McCombo]

the oilers lost their starting goalie 2 games into the season. the result is 6-1 record since then. they are building a cushion now to protect vs lulls that happen to EVERY team. the st louis blues went from last place like actual last place at xmas and won the cup. your advanced stat agrument is just going circular and making excuses for other teams while not giving the oilers the same benefit? the oilers beat the shit out of the flames and they have only lost a couple times since. The oilers are a top 5 team despite their underlying stats, that dont mean much anyway when your PP is near 50%, the PK is top 5-6ish and you are giving up limited grade A scoring chances every game.

That exact same St. Louis team had crap underlying numbers to start the season, but from Xmas on they led the league xG. Very deserved SC.

Yea the Oilers lost a starting goalie, but it hasn't been a problem because Koskinen has been awesome. No injury is a problem if there is internal replacement that plays as well or better.

PP will not stay at near 50% for long time. But like I said in some post, it will most likely still be best in the league by a decent margin and is something to lean on.

It is good to build cushion, but no need to worry about playoffs. Weak pacific, great start and best forward group in the league. Via Dom's model Oilers will make playoffs 97%.

"Limited grade A scoring chances", if I look at Natural Stat Trick or Moneypuck it says a different story. Via NST Oilers have created 80 high danger scoring chances and given 87 5on5. xG% is a hair below 50.

Moneypuck.com xG% 46.34 for Oilers.

Also JFresh seems to agree. You can find Koskinen from "Tough workload, great performance" segment.

I am going to trust three models more than your eye-test.

 

[McSuper]

I think Vegas has made a big mistake. They're quickly losing all their depth and they've given away all their young players too.

Not to mention the are almost at the cap for next year with 2/3 of a team . Chia level bad at negotiating contracts . Worst team for committing to a player LT and trading him the next off season

 

[McCombo]

Wow, you absolutely don’t understand the concept of a variable when it comes to statistics, do you?



Ah yes, there it is...

Sure are variable in statistics too, but even more on results. I will put more weigth on xG, scoring chances, high danger scoring chances than actual goals from ten game sample size. Some teams have inflated shooting or save percentages, some teams can't score from anywhere, hit posts/crossbars. Loose/win all the one goal games for a while.

Extreme example would be when Ben Scrivens had ~50 save shutout (was it against LA? ) and the Oilers won eventhough they were badly outclassed. Oilers got two points and W, but when predicting future outcomes that game would not have positive effect on Oilers.

There are lesser examples every round when a team with lower xG/scoring chances win the game (un-deservedly). Some teams will win multiple games in a row eventhough they are out-chanced. They will rise in standings, but unless they actually improve they will eventually come back to ground.

There it is and it is so bang on. Some people sure have done some math calculations or they are familiar from poker or betting that how crazy good/bad runs you will get. And there are probabilities in a hockey so it isn't outside of those variables.

I was arguing earlier with a poster who said that best teams almost always (~90% of time) beat worst/below average teams . That just isn't true. The odds for third worst team in the league to win third best team are about the same as if you can guess which suit is the first card in the deck. And that is for away game.

 

[Ol' Jase]

Sure are variable in statistics too, but even more on results. I will put more weigth on xG, scoring chances, high danger scoring chances than actual goals from ten game sample size. Some teams have inflated shooting or save percentages, some teams can't score from anywhere, hit posts/crossbars. Loose/win all the one goal games for a while.

Extreme example would be when Ben Scrivens had ~50 save shutout (was it against LA? ) and the Oilers won eventhough they were badly outclassed. Oilers got two points and W, but when predicting future outcomes that game would not have positive effect on Oilers.

There are lesser examples every round when a team with lower xG/scoring chances win the game (un-deservedly). Some teams will win multiple games in a row eventhough they are out-chanced. They will rise in standings, but unless they actually improve they will eventually come back to ground.

There it is and it is so bang on. Some people sure have done some math calculations or they are familiar from poker or betting that how crazy good/bad runs you will get. And there are probabilities in a hockey so it isn't outside of those variables.

I was arguing earlier with a poster who said that best teams almost always (~90% of time) beat worst/below average teams . That just isn't true. The odds for third worst team in the league to win third best team are about the same as if you can guess which suit is the first card in the deck. And that is for away game.

Hey, if you’re willing to put any weight on any stat from a nine game sample size, knock yourself out, but statistically speaking it has no validity in terms of predictive effectiveness.

That’s not hockey, that’s statistics.

I’m not even going to touch how incredibly ridiculous it is to equate the variability of hockey events with that of poker or coin-flipping. That’s just embarrassing.

 

[McSuper]

Righteousness is a real drug. Guys who would rather see the team fail just to be right.
Then when it’s obvious they are wrong the main dish of Crow is easily available.
Ignore feature is great, absolutely amazes me why the same ten posters aren’t ignored by everyone.

I have always said ,when I disagree with a fellow Oilers fan that " I would rather him be right if the team is winning " Anyone who would rather the team lose to be right isn't a fan . I take all the crow I can get if the team is winning .

 

[Ol' Jase]

That exact same St. Louis team had crap underlying numbers to start the season, but from Xmas on they led the league xG. Very deserved SC.

Yea the Oilers lost a starting goalie, but it hasn't been a problem because Koskinen has been awesome. No injury is a problem if there is internal replacement that plays as well or better.

PP will not stay at near 50% for long time. But like I said in some post, it will most likely still be best in the league by a decent margin and is something to lean on.

It is good to build cushion, but no need to worry about playoffs. Weak pacific, great start and best forward group in the league. Via Dom's model Oilers will make playoffs 97%.

"Limited grade A scoring chances", if I look at Natural Stat Trick or Moneypuck it says a different story. Via NST Oilers have created 80 high danger scoring chances and given 87 5on5. xG% is a hair below 50.

Moneypuck.com xG% 46.34 for Oilers.

Also JFresh seems to agree. You can find Koskinen from "Tough workload, great performance" segment.

I am going to trust three models more than your eye-test.

Statistically speaking, J-Fresh’s “goalie performance comparison” has the usefulness of toilet paper stuck to a shoe.

Just mind-blowing people just accept this as not only valid, but hardcore predictive.

 

[McCombo]

Statistically speaking, J-Fresh’s “goalie performance comparison” has the usefulness of toilet paper stuck to a shoe.

Just mind-blowing people just accept this as not only valid, but hardcore predictive.

So I should trust someones opinion that has a very high chance for bias, rather than a model that uses actual shot locations (no bias) and how often there have been scored goals in NHL in the past from those areas?

And the model at least approves my eye-test. Koskinen has played great and made multiple big saves on most nights, Oilers haven't been that great at limiting chances against.

 

[McCombo]

Hey, if you’re willing to put any weight on any stat from a nine game sample size, knock yourself out, but statistically speaking it has no validity in terms of predictive effectiveness.

That’s not hockey, that’s statistics.

I’m not even going to touch how incredibly ridiculous it is to equate the variability of hockey events with that of poker or coin-flipping. That’s just embarrassing.

Well I need to be aware which teams are most likely over/under-performing, to be able to adjust and make correct decisions. If a team with weak roster, about the same core as last year (finished lets say 26th in the league) has won 6 out of 10 games this year. Has a 44% xG, 103 PDO. Most likely that will still suck this year, no need think that they are any better eventhough they have won 6/10. That is just randomness and they will come back to ground at some point. And most likely average fans are thinking that they are better than they actually are because of results. After that the very same team loses 7 out next 10, same wins will say "Ahh they still suck, I thought they were better". But every single fact was pointing that, other than results.

Though it isn't stupid. If two really close teams play against each other in a neutral environment the probabilities are 50-50 (or a slightly worse team has a home game, home ice advantage makes it a really close match-up). Exactly the same as in a coin toss, the outcome is decided very differently but odds are same. Two very close teams can play five games against each other in same season and the team X wins all of those matches. I see a lot of those arguments used here (not just in Oilers section) that in that scenario the team X must be clearly better because of those games. While I agree that sometimes team X's playing style may fit better against team Y, but mostly that is just randomness.

You don't need to look far for few examples. Last year Oilers dominated Winnipeg in regular season, but Winnipeg won the playoff series. Oilers were favourites going into the series, but not by a huge margin. 19-20 season Detroit Red Wings played four games against Montreal Canadiens and won all of them. Red Wings finished the season with just 17 wins and four of those came against Montreal. Detroit were a much weaker team, but still they ended up sweeping the season series.

Just think how many 4-8 game sample sizes one NHL regular season includes. 32 teams, 82 games for everyone, so if I am not mistaken 1312 games. Then flip a coin 1312 times and see how many crazy runs of same side you get.

 

[Rengorlex]

And the model at least approves my eye-test. Koskinen has played great and made multiple big saves on most nights, Oilers haven't been that great at limiting chances against.

His results have been great. I'd argue there's a ton of variation in goaltender performance based on luck, for that goalie too, just based on the nature of the position. Doesn't control where the shooters shoot the puck.

To my eye, Koskinen hasn't looked great in the net. But I'm not a goaltending expert by any means.

 

[McCombo]

His results have been great. I'd argue there's a ton of variation in goaltender performance based on luck, for that goalie too, just based on the nature of the position. Doesn't control where the shooters shoot the puck.

To my eye, Koskinen hasn't looked great in the net. But I'm not a goaltending expert by any means.

His movement looks clumsy/weird, but for 6'7'' it is a lot harder to be as smooth as Saros for example.

Chara' s skating hasn't been a treat to watch ever. If you get the job done better than most in your position I am fine with that no matter how you look.

Mike Smith isn't the textbook example either.

Edit. Not a goalie expert either