The Quest stands upon the edge of a knife. Stray but a little, and it will fail, to the ruin of all. Yet hope remains while the Company is true. Go Flames Go!
I don't see what difference the age of the stats are - his point is valid.
The critique is he constantly brings up specific stats apropos of nothing, gets called on their relevance and/or validity, and then immediately moves onto new random arguments that people aren't actually having.
I find the posting style to be very comical.
The Following 2 Users Say Thank You to AC For This Useful Post:
The critique is he constantly brings up specific stats apropos of nothing, gets called on their relevance and/or validity, and then immediately moves onto new random arguments that people aren't actually having.
I find the posting style to be very comical.
True, and I am one of his loudest critics.
But IMO, this point was valid. And thus should be treated as such.
This tirade was intended as a defense of analytics, but serves well as the argument against them. Shooting stats and the like will correlate with results, but never add much with respect to predicting them. Because results aren't a function of improved statistics, they are a function of talent and execution.
The difference between winning and losing isn't improving your xGF or whatever stat you want to look at, it's about execution.
As I said, a team of McDavid's with the Kings analytics would destroy a team of McDavid's with the Jets analytics.
While talent does offset some downfalls in players' analytics, it only does so much offensively. Aside from that, you're still being a complete drag on your team defensively, meaning you really have to rely on your goaltender to save your a**.
Not to mention only so many players have the talent worthy of ignoring their analytics to a certain extent. When it comes to slightly-above average players and depth players, you absolutely 110% go for players who are good analytically and can hold their own.
It's also very easy to differentiate the true stars from the elite of the league by analytics, as the true superstars (McDavid, Crosby, Malkin, Ovechkin, Matthews, the entirety of Boston's #1 line, Tkachuk, Gaudreau, etc) tend to be very strong analytically at minimum, if not incredible, while at the same time outproducing their expected scoring stats
The point being, why are you bringing up stats as old as 40 years ago to try and prove something relating to Calgary vs Winnipeg in 2020?
Talk about grasping for straws.
Please look into the overall summed +/- for each team in the 1980s next for reference. I'm sure it will be compelling stuff.
We were discussing whether the Jets and Flames were good teams or mediocre or bad teams based on shot-based stats. The Oilers were great but had mediocre shot based stats.
On the other hand the Oiler dynasty teams had 107 skaters play for them with a combined +/- of +3,066
The Flames over the 1982-1990 time period where they may have been the 2nd best team in the NHL had 92 players skate for them and had a combined +/- of +1,749
There is a far greater correlation between +/- and success than shots on goal and success.
Seems self evident that if you have a bunch of players on your team that are on the ice at even strength when your team score than when the other guys score you will win a lot more games and be a better team.
Having a lot of players players that deliver better results than the guys they are playing against will consistently win games.
To me personally, if you really want to argue that a stat is relevant in predicting the outcome of the Flames-Jets series, you really should be looking at "how often does the regular season leader in this stat win their series". However since this would likely require a lot more work, I don't hold it against Bingo that he didn't do that.
(That said, since it wouldn't be THAT much work, if corsi is indeed a relevant indicator of who wins a playoff series, that information should already be available in the internet somewhere.)
I think if you go back though you won't see me making an argument for Calgary beating Edmonton, or really a prediction of any kind.
And it wasn't corsi at all ... it was xGF% which takes into account a lot more than just shot attempts.
Ricardo kept talking about how well the Jets have played, and I'm just pointing out that their underlying numbers suggest they're riding a hot goaltender and high finish rates.
He brought up LA which had me show they are the only team in the top X teams that are still playing.
The rest was your rabbit hole.
The Following User Says Thank You to Bingo For This Useful Post:
You are nothing if not consistent. You are conceptually challenged and seem proud of it.
The simple fact that great teams dominating the standings and being the obvious best teams of the era were not out-shooting the rest of the teams is a open and direct challenge to the concept that shots based possession stats have any value whatsoever.
If shots based stats were important in identifying good/great hockey teams I would have expected that the Oilers would have been out-shooting their opposition by a huge margin.
This just seems silly though.
Are you suggesting getting out shot as a tactic then?
None of these stats guarantee anything, but it stands to reason that you're better off out attempting, shooting, and scoring chance your opponents if you want to win more games than you lose.
I think if you go back though you won't see me making an argument for Calgary beating Edmonton, or really a prediction of any kind.
And it wasn't corsi at all ... it was xGF% which takes into account a lot more than just shot attempts.
Ricardo kept talking about how well the Jets have played, and I'm just pointing out that their underlying numbers suggest they're riding a hot goaltender and high finish rates.
He brought up LA which had me show they are the only team in the top X teams that are still playing.
The rest was your rabbit hole.
Yeah, xGF%, corsi, whatever. Shot-based stats. Garbage in, sprinkle with faulty logic, garbage out, consistently apply to too small sample sizes. Rinse, wash, repeat. Although I do find "expected goals" to be hilarious in the sense that someone actually named it that.
Sorry. That was a completely different rabbit hole. Carry on. That was just a dig, I'm not going to bother arguing the point.
Yeah, xGF%, corsi, whatever. Shot-based stats. Garbage in, sprinkle with faulty logic, garbage out, consistently apply to too small sample sizes. Rinse, wash, repeat. Although I do find "expected goals" to be hilarious in the sense that someone actually named it that.
Sorry. That was a completely different rabbit hole. Carry on. That was just a dig, I'm not going to bother arguing the point.
Yet you did
xGF% is a step up from corsi as it takes into account quality of shot and situation. It will evolve and get better, get replaced, whatever, but not sure where the logic is faulty.
If one team gives up more high danger chances against than for, that will drive their xgf% down. Where is the logic break in that?
Seems to me if 100 guys jumped off a 10 story building and 87 broke a leg, there isn't much of an argument that it's a good move because 13 got away with just sore shins.
The Following User Says Thank You to Bingo For This Useful Post:
Are you suggesting getting out shot as a tactic then?
None of these stats guarantee anything, but it stands to reason that you're better off out attempting, shooting, and scoring chance your opponents if you want to win more games than you lose.
Not at all.
But trying to identify a good or better team by shots-based stats is not valid.
The Oilers and Islanders were great teams in the day and they were confident in Fuhr and Billy Smith stopping the shots that they need to win.
Teams with confidence in their goalie will play a different possession game.
For the Oilers a hard save by Fuhr was the start of a scoring chance for them.
I am just using Oilers as an example of an unarguable great team.
The Hartley Flames worked the same way without the future HOFers in the lineup. They counted on Hiller and Ramo to make some saves and generate offense from the defense zone and the goalie doing his job.
Bad possession stats were a result of playing a system that was most successful for the team.
No one is going to argue that the Oilers could have been even better if they focused on shutting down shots against.
To me personally, if you really want to argue that a stat is relevant in predicting the outcome of the Flames-Jets series, you really should be looking at "how often does the regular season leader in this stat win their series". However since this would likely require a lot more work, I don't hold it against Bingo that he didn't do that.
(That said, since it wouldn't be THAT much work, if corsi is indeed a relevant indicator of who wins a playoff series, that information should already be available in the internet somewhere.)
I'm not going to go into every team but lets look at the last 5 playoffs and the final 4 (All Strengths, xGF%)
18-19: St.Louis Blues (6th), San Jose (4TH), Boston (8th), Carolina (1st)
17-18: Washington Capitals (29th), Tampa Bay (3), Winnipeg (16), Vegas (5)
15-16: Pittsburgh (1st), Tampa (11th), St.Louis (20th), San Jose (2nd)
14-15: Rangers (14th) , Tampa Bay (10th), Anaheim (13th), Chicago (8th)
13-14: Montreal (21), New York (5), Chicago(3), Los Angeles (2nd)
So it does seem to be better to be ranked top 5/10 in the regular season in xGF% if you are hoping to have playoff success.
Washington is the only major outlier that was able to win a cup, and they really changed their game from regular season to playoffs. Going from 46.4% (29th) in the regular season to 52.3% (5th) in the playoffs.
Last edited by SuperMatt18; 07-27-2020 at 09:57 AM.
