Then why are you even posting in a thread that is specifically about those things?
Melvin's 2021-22 NHL Team Projections
- Thread starter Melvin
- Start date
Then why are you even posting in a thread that is specifically about those things?
oh really? I need to? Why do I need to? I’ll continue working on this as I enjoy working on it and trying to make improvements as was my goal from the outset.
There were 31 other teams that were projected and thankfully next year there won’t be an expansion team included that will be everyone’s focal point despite being 3% of the data with the widest confidence intervals.
I don’t “need to” do anything to please you. You can feel free to ignore me, I certainly have nothing to ever gain from reading anything you ever post.
Tables of Stats
Registered User
I'd imagine this would be a rather large ask, but how realistic would be building a 'Potato Trade Analyser'? It could try to value picks based on the average expected value of that slot and some number of slots below them. The player assets and cap aspects of a trade would be harder to evaluate, but in theory, I think it just might be possible.
Tables of Stats
Registered User
Did you miss the part where Melvin explained their reasoning and said that Seattle was one of the teams their model couldn't predict with a high degree of accuracy? Did you do anything more than simply predict that Seattle would be bad? Did you specifically say they'd have near-record levels of bad goaltending for the first few months of the season? If not you made a worse prediction than Melvin did because at least their prediction can be tested and their model tweaked to provide more accurate results.
geebaan
7th round busted
- Oct 27, 2012
- 10,188
- 9,400
It’s literally a guess. A calculated guess.
posters that are offended by a guess need to grow up.
posters that are offended by a guess need to grow up.
For every team, I posted the detailed numbers. Expected goals for, expected goals against, goaltending, Pp and Pk. It’s these numbers that are interesting to me.
I am waiting for the conclusion of the season to do a deep dive, but Seattle just isn’t that interesting. It’s goaltending. That’s the number that was wrong. Okay, now what about the other 31 teams? Hell, the projection of the flames looks worse than Seattle right now. Or just as bad.
Seattle was the team I most expected to be off, by far, for obvious reasons. This isn’t the Seattle thread. I’m much more interested in seeing how the other teams shook out in terms of the components.
I am waiting for the conclusion of the season to do a deep dive, but Seattle just isn’t that interesting. It’s goaltending. That’s the number that was wrong. Okay, now what about the other 31 teams? Hell, the projection of the flames looks worse than Seattle right now. Or just as bad.
Seattle was the team I most expected to be off, by far, for obvious reasons. This isn’t the Seattle thread. I’m much more interested in seeing how the other teams shook out in terms of the components.
Definitely something that should be possible. To what end though? Like what would it be used for?
Tables of Stats
Registered User
My main idea would be to use it as a novel way to judge who 'won' any given trade. More usefully it could be used to judge what a 'fair' value for a given set of assets might be. It would be tough to be accurate, given that GMs aren't analytical engines and make trades and set values on gut feel, but it could certainly make for a fun way to generate articles near the TDL.
Like it's funny that this is the Canucks board but people are obsessing over Seattle instead of the team that I had the most amount of data for, the Vancouver Canucks:
All we really got wrong here was that the PP has been mediocre instead of very good, and the PK has been even worse than projected at 32nd instead of 26th. Our goal differential is -8 where the model had -12. Our offense is in 27th when the model said 28th. Still 30 games to go though, but I'm quite happy that the model pretty much nailed the team that I follow the closest (and thus had the roster/usage entered most accurately.)
VANCOUVER CANUCKS
ES Offense: 28th
ES Defense: 26th
PP: 5th
PK: 26th
Goaltending: 10th
Overall Goal Differential: -12
Finally. Good grief. I was really starting to sweat if they didn't pop up soon. This team looks like the team from two seasons ago, where ES play was very poor but the PP and the goaltending kept them afloat. People think this offense is so good but I think fail to realize just how much of the scoring comes from the PP; I was tracking it and at one point something like 40% of their goals came on their PP, leading the league by a lot. The Canucks are very good at drawing penalties, and have added another very good forward in Garland who also draws penalties at a high rate. Their PP should be very good, but at ES they have not been great at generating offense for years now, and it is because despite the firepower, it is hard to score when the puck is in your zone.
Unlike two seasons ago, the Canucks do not have Tanev/Edler, but instead have Myers/OEL, which looks like a total train-wreck. Hughes struggled defensively last year (I have a generous bounce-back projected) but no matter what some people say, Poolman and Hamonic are not Chris Tanev. Dickinson has some very good numbers and should be an improvement on Jay Beagle, but this still looks, on paper, like a team that will be spending a lot of time in their own end and relying on the PP to score goals, as well as on Demko/Halak to bail them out of trouble. Another concern is injuries, as Petey has to be seen as something of a risk at this point, and any team that spends so much time blocking shots and defending is bound to pile up the injuries, something that I thought the Canucks had learned by now.
All we really got wrong here was that the PP has been mediocre instead of very good, and the PK has been even worse than projected at 32nd instead of 26th. Our goal differential is -8 where the model had -12. Our offense is in 27th when the model said 28th. Still 30 games to go though, but I'm quite happy that the model pretty much nailed the team that I follow the closest (and thus had the roster/usage entered most accurately.)
Where are calgary? If its just a blanket increase in everything then it could be argued that it really is just a coaching change + un hurt marky.
Probably fighting with the Ducks and Oilers for that last playoff spot in the pacific. They built themselves to be this hard to play, defensive team and then got obliterated by bad goaltending. Like what would happen to the Oilers if McDrai got injured for an extended period of time or had Pettersson levels of underpeformance. Disaster
In a way its good/bad from a van pov that this happened. They get another top 5 pick to go alongside Berniers and will get assets that Francis should have obtained in the expansion draft. Add in being bad next year for a chance at Bedard or Michkov and a clean cap sheet going forward
The Kraken could be the better team over the next 3 years and have a better future outlook. Thats so depressing
VanJack said:
You literally got everything wrong about your predictions except their placement.
You predicted good goaltending and tight checking but no scoring.
Instead they had decent scoring and horrible goaltending.
Do you need to take your own recommendation and have a break from hockey to play xbox because you got everything wrong about their team and are refusing to answer basic questions like where would they be in the standings if their goaltending was as good as you yourself predicted it would be?
It blows my mind that people are stuck on this "expansion team must be bad because expansion teams 40 years ago were bad" even though they have completely different expansion rules and there's a salary cap now, which are two massive differences.
Yeah. Had their goaltending in 28th. Whoops. Obviously Markstrom bounced back in a huge way, and their offense has been better than expected. But also, we're in the middle of the hottest streak of the season right now; we'll see where they settle out.
This is why I keep harping on the components. Like if Seattle had received bad goaltending but somehow had a powerhouse offense I don't think that would make my projections look "good," since even if they ended up in the same place I would have gotten all the pieces of it wrong.
It's kinda like that thing where the kid does all of the steps of the math question incorrectly but somehow still gets the right answer in the end. It's not really all that satisfying to me.
Although it would make this thread more fun.
M2Beezy
Objective and Neutral Hockey Commentator
I find the Seattle focus odd, but i think tehres a few reasons for it:
-Everyone and their brother predicted The Golden Knights would be a bottom feeder their first year and they made the finals and have done pretty dam good since then and people were maybe expecting something like that COULD happen again with Seattle, so many expert and amateur analyses thought it was possible and/or probable
-There was some excitement/interest with Canucks fans as the team had been run so hard into the ground by Benning that some people suggested cheering for the Kraken or that theyd do better than another season of Benning destroying the Canucks
Anyways, i still love reading the first few pages of this thread and a certain group of supporters of the dud Benning regime so confidently attacking some predictions
Peen
Rejoicing in a Benning-free world
- Oct 6, 2013
- 29,742
- 29,664
This is literally everyone. "They're not going to score, but they have a good D and good goalie but you can't expect them to win 50 games 2-1" or whatever.
Like, no one who is trying to dunk about the Kraken now actually got what would be their downfall right.
Like 90% of the time, it's the same three factors.
Goaltending, coaching, rookie variance. Other than that, you can pretty well predict teams likely finishes.
AGREED! The last two games is the only sample size you need to know about that team
Pastor Of Muppetz
Registered User
- Oct 1, 2017
- 25,892
- 16,525
I was just giving an example encapsulating their play of late
190 goals against this season is all you need to know..
Tables of Stats
Registered User
How about the ratio of shots to goals? Or where the goals have been scored from? Those would tend to be indicators of a team defense issue versus a goaltending issue.
Both have the same end result, but the cause and the solution for one is different than for the other.
Aphid Attraction
Registered User
- Jan 17, 2013
- 5,055
- 1,716
You should see how they treat the guys that try guess the end of the world
Pastor Of Muppetz
Registered User
- Oct 1, 2017
- 25,892
- 16,525
But it’s not just goaltending.
Analysis: What’s gone wrong for the Kraken and how can they climb out of last place?
Well, that's what GSAx is supposed to help with, in theory, by doing exactly that. In a perfect world, xGA should be able to tell you how well the defense did (by telling you how many goals you "should" have allowed.) And then the rest is presumably goaltending.
Of course we don't live in a perfect world.
I've calculated my own xGA/GSAx metrics which I've posted in the analytics thread. They are based on shot distance and shot angle. For this season they weren't ready yet so I used Evolving-Hockey's, but I might use my own next year. I haven't decided yet. I don't think mine are any better or anything.
It's also that they can't outscore the goaltending problems?
Tables of Stats
Registered User
I was asking Pastor Of Muppetz because they seem to think that only goals against and goals for matter and the underlying causes are irrelevant.
I expect that you look at more data than GF, GA, and points percentage when assessing a team and making predictions.
Users who are viewing this thread
Latest posts
-
HF Habs: Trade Proposal Thread #94: "We know what we are, but know not what we may be." (131 Viewers)- Latest: sampollock
-
-
Confirmed with Link: Hockey Hall of Fame Runs on Duncan (Keith) (18 Viewers)- Latest: Hull2Roenick2Kane
-
