Vamos Rafa
Registered User
Why does it seem like in today's game, the team with the more shots usually loses the game? I think in all of the losses this season, the Kings outshot their opponents.
GDT: 2019-20 GM#12 LA Kings vs Chicago Blackhawks @4:00pm 10/27/19
- Thread starter kingsboy11
- Start date
-
Work is still on-going to rebuild the site styling and features. Please report any issues you may experience so we can look into it. Click Here for Updates
This season will be an interesting study in seeing how useful all of these possession stats really are. The Kings are at or near the top of the league in every possession statistic, from straight Corsi to scoring chances and expected goals. They also are 4-8 on the season. On the other hand, you have a team like the Islanders who are near the bottom in most possession stats, but have just won 7 games in a row.
Raccoon Jesus
We were right there
I know everyone wants to think we've moved on to a different era, but the best possession teams are by and large the best teams.
There will always be outliers and you definitely need shooting talent and goaltending talent but if you look at the percentages it's easy to see why it's just a numbers game. Hell, it's like prospecting for sales, if someone is closing 1% more of the time than you are, but you're making 200 more calls a day, you'll win.
Also generally speaking it's a Sutterism but a team controlling the shots and flow of play are usually far away from their own net. However, we've seen it in recent years and exaggerated now with lesser defensive personnel, it doesn't matter if you limit chances if damn near every chance is a two-on-one haha.
You still have to put it all together but it's very unusual to have a team that can outplay every opponent and lose more than they win. That's a lack of shooting talent, goaltending talent, or bear-down-in-front-of-the-net-on-rebounds talent. If you're going to hit 40 goalie pads a game, get a Grundstrom or Hornqvist or Anders Lee in there to clean up the crap.
There will always be outliers and you definitely need shooting talent and goaltending talent but if you look at the percentages it's easy to see why it's just a numbers game. Hell, it's like prospecting for sales, if someone is closing 1% more of the time than you are, but you're making 200 more calls a day, you'll win.
Also generally speaking it's a Sutterism but a team controlling the shots and flow of play are usually far away from their own net. However, we've seen it in recent years and exaggerated now with lesser defensive personnel, it doesn't matter if you limit chances if damn near every chance is a two-on-one haha.
You still have to put it all together but it's very unusual to have a team that can outplay every opponent and lose more than they win. That's a lack of shooting talent, goaltending talent, or bear-down-in-front-of-the-net-on-rebounds talent. If you're going to hit 40 goalie pads a game, get a Grundstrom or Hornqvist or Anders Lee in there to clean up the crap.
tny760
Registered User
- Mar 12, 2017
- 20,782
- 22,653
that's what i think you need if you expect corsi-dominant teams to win, a good set of mop-up guys up front or guys that are really good at tipping shots from the slot
we saw a few shots against lehner that were good hard shots for deflections, the problem was that the guys crashing the net were looking for those low tip plays and couldn't handle the rebounds at the speed they got them
i'll trust the process that those are coachable moments that these guys will understand eventually
tny760
Registered User
- Mar 12, 2017
- 20,782
- 22,653
ahhh that's why nate schmidt got popped
i guess his trainers just hold their crystals in syringes that they align in his ass cheek meridians
King'sPawn
Enjoy the chaos
- Jul 1, 2003
- 23,190
- 24,101
Well, Sharks won't be getting the pick. They traded it to Ottawa.
lumbergh
It was an idea. I didn't say it was a good idea.
This is not correct. Go to Natural Stat Trick and sort the teams this season by Point %. You'll see that there's little to no correlation between the ability to win games and Corsi For %, Fenwick For %, Shot Attempt %, all indications of possession.
The top six teams by point percentage have a 5vs5 CF% of 49% as a group. The six best 5vs5 CF% teams have a Point % of 0.578 as a group, which over a season would leave you with 95 points, marking you as a bubble team. The leaders of each division are also under 50% as a group.
I'm as much of a numbers guy as anyone, so I'd love to have some advanced stat that tells me what a winning team looks like. Just turns out that possession numbers don't do that. Maybe they did 10 years ago, but they don't today.
In the end, you have to outscore the other team. If you can't put the puck in the net and your goalie can't keep the puck out of your net, you're still gonna lose.
Raccoon Jesus
We were right there
This season is barely 10 games deep.
Go look at last season.
And again I'm not at all suggesting that's the be all and end all, just that there's not some crazy sea change. And it boils down to a really simple, ironically and extremely old school philosophy, shoot more than your opponent every night. That's why I've never understood why that's so controversial.
Yes there are some newer systems that emphasize holding the puck and shot quality instead of just feeding the Sutter funnel but you'll notice they're still high in CF% because they cut down CA rather than just raw pump CF.
On the flipside, that's also why teams with generational talents are pretty flippant about it, they don't care if you want to play pond hockey because they can score seemingly at will relative to others.
You are going to want to look at xGF%, which takes into account not only shots, but shot quality as well. Take a look at last season and you will see that most of the top 16 were playoff teams, with only a couple of outliers either way(Capitals/Wild).
lumbergh
It was an idea. I didn't say it was a good idea.
My argument was about the idea that high possession teams are better teams, which this year has turned out not to be true. But just for giggles, for 2019 so far, xGF% also has very little correlation with Point %. Even less than CF%, SF% or FF%.
If you're gonna go to xGF%, you might as well just look at GF%, which actually is strongly correlated with Point %. That's obvious, though. If your team outscores its opponent at 5vs5, it's a good team.
Well, you kind of answered your own question. If you are looking for a stat with a direct correlation to points percentage, you won't find one better than GF%.
xGF% is more of a predictor, the idea is that as the season goes along teams with high xGF% will end up near the top of the league in points percentage by the end of the season. It is looking for teams who have some unsustainable streaks early on that should normalize. None of this is perfect and there are outliers(I have my own questions about them), but this is more for just looking into who has potentially been performing better than their record shows.
lumbergh
It was an idea. I didn't say it was a good idea.
178 games have already been played by teams in the NHL this season. The season is 1/7 over already. The "small sample size" argument doesn't hold much water at this point.
Simple and old school does not equal true. The data this season say shoot more does not produce more wins. Vamos Rafa is onto something.
Last edited:
CF% isn't as reliable as it once was with the evolution of playstyle, but by the end of the season you'll still see most of the strong CF% teams experience success. Last season 13 of the top 16 CF% teams made the cup playoffs. You'll have outlier years every once in a while, but more often than not teams with strong possession numbers make the playoffs.
As far as stats go, there isn't a much better predictor of success than possession. Not perfect but it's probably the best we got.
178 is a small sample size for a stat like CF% which is reliant on lots of data. The reason isn't so much volume, but when a team has only played 10 games things like injuries and strength of schedule can have huge effects on the data. Ideally you need at least a full season to make anything of team Corsi numbers. You can glean a lot more from individual Corsi in a shorter period though, maybe 25-30 games. It doesn't take as long for things to balance out when a player is seeing many shifts per game, but matchups have to be taken into account.
lumbergh
It was an idea. I didn't say it was a good idea.
Just for fun I went through Natural Stat Trick and determined the number of times that the winning team has outshot the losing team this season. In 178 games this season, the winning team has outshot the losing team 79 times. The losing team outshot the winning team 75 times.
Basically, this season, outshooting your opponent doesn't lead to more wins.
It would be interesting to see this with shot attempts rather than just shots. A lot of the most dangerous opportunities are through traffic where the puck gets deflected or bounces away from the net. They don't get registered as a shot.
Corsi was an incredible predictor for about 5 years there when big, dominating possession was the thing all teams were aiming towards, and it still holds it's own pretty well. Not because it's this tell-all stat, but because there isn't much else out there that has ever shown to be a great predictor. I fully expect a different stat to be better suited now that things are moving towards speed/skill. It wouldn't surprise me if quality of shot attempts surpasses quantity soon, things are definitely trending that way.
Choralone
Registered User
Yeah, Corsi is the best of a limited, flawed set of statistics we have had to date. The NHL is behind other leagues in recording these more granular statistics, but it's moving in the right direction. It's such a fast, flowing game I think it's probably the hardest of the big 4 to measure. Football is difficult, too, I think, because there are so many players moving at the same time and so much physical chaos, but the constant stoppages help.
lumbergh
It was an idea. I didn't say it was a good idea.
Easy peasy. 79 games the winner (not including shootout) had more shot attempts, 78 games the winner (not including shootout) had fewer shot attempts. Again, no real correlation between shot attempts and winning the game.
Choralone
Registered User
For everyone's clarity:
Corsi takes into account:
Fenwick takes into account:
Corsi takes into account:
- Shots on goal
- Missed shots on goal
- Blocked shot attempts towards the goal
Fenwick takes into account:
- Shots on goal
- Missed shots on goal
Raccoon Jesus
We were right there
178 games is a much smaller sample size than 2460.
Would love to see how this shakes out throughout the season but every year people decry it and every year the result is the same.
Every team that thinks they found 'the secret' crashes back to earth hard.
Even if systems change to "hold the puck until you get a quality shot and nothing else," those teams will still have a higher CF% by virtue of dragging down CA due to true possession time. Maybe things are changing, but it will be very slowly, certainly not quickly, as recent years have demonstrated.
With respect, I think you're missing that half of the equation. It's not just "shoot everything," it's "prevent shot attempts in any manner" as well.
Thanks. I think this illustrates how data cannot be taken on it's own, but has to be taken in context. For years, CF% has had a positive correlation with wins. There's a fairly high variance, but the relationship is definitely there. We are seeing a shift away from this though.
I've attached a graph I compiled of CF% vs Points Percentage since the 2010 season. While there are outliers, you can see there's a very solid correlation between CF% and increased points/wins.
So CF% has traditionally been a good indicator of points. What's most interesting are the teams on the right side of the graph. From right most to left they are: LA, BOS, VGK, CAR, SJ, CHI, StL. Over that time period all but one of those teams have won cups or been in Finals. It's pretty impressive really for a stat to have a correlation with cup winners like that, and it's why so many teams were pushing the possession type of play for so long. That tall outlier at the top-middle is Washington. Their try to outscore you style doesn't lend itself to strong possession but they did win a cup. It also explains why they were a disappointment in the playoffs for a lot of those years.
Lately though it's been changing, so in 10 years CF% might be a terrible stat. Would be neat to revisit this after a run of the newer style trends.
lumbergh
It was an idea. I didn't say it was a good idea.
I don’t know if this has occurred to anyone, but a good team that produces good chances would in principle score more goals with fewer shots. Say a highly skilled team might score on the first shot or two in the zone while the Kings might take 4 shots with little chance of scoring. Sure one looks good from a possession point of view, but the other one gets you a goal.
I’m becoming more and more convinced that your Trevor Lewises and Alex Iafallos don’t actually win the Kings games. So much effort, so few results.
I’m becoming more and more convinced that your Trevor Lewises and Alex Iafallos don’t actually win the Kings games. So much effort, so few results.
Raccoon Jesus
We were right there
Which is one of the reasons they started tracking high danger scoring opportunities rather than just raw volume as well, but the same problem applies with the Kings, despite the large jump in events.
But believe me yes that's occured to everyone. Still, the math adds up--provided you have NHL level scoring, a few difference points in shooting percentage would still find a team pouring more shots in (while limiting the opposition's) the winner more often than not. The "provided you have NHL level shooting" thing has been the hitch for us, though.
Latest posts
-
Canucks & NHL News, Rumours, and & Fantasy GM | The Morning After- Latest: Reverend Mayhem
-
-
-
