Login

***Wally*** · 08-16-2022, 11:10 AM

08-16-2022, 10:45 AMMatteo Wrote:
08-16-2022, 10:04 AMWally Wrote: Was hoping to see someone delve into the adv goalie stats finally… sad :*(
GSAx in the index when

Oh I was just meaning seeing if there are correlations in tpe versus advanced save % in FHM

*Keven* · 08-16-2022, 12:06 PM

Obviously it would be nice if there was a stronger correlation between TPE and save %. It's unfortunate that 155 TPE freshly created goalies can survive and win games at the SHL level. However, it's important to remember that the data you're looking at isn't perfect. It's worth keeping in mind that the schedule for backups is advantageous. They're usually well rested and are only played against the lowest teams in the standings.

In terms of shooting percentage, the worst 5 teams (PHI, EDM, MIN, TEX, NEW) ranged from 9.6% to 10.1%. This means over the course of the entire season, the goalies they faced were rewarded with save percentages of 0.899 to 0.904. Now maybe this is due to the opposing goalies always standing on their heads during these games, but I think it's much more likely that this can be attributed to the play of the skaters. Compare this to the top 5 teams for shooting percentage (CHI, TOR, TBB, LAP, MTL) who ranged from 11.2 to 12.5%. Goalies playing against these teams would be getting save percentages of 0.888 to 0.875. I'd argue that the lower TPE goalies in the league had much greater chances of facing the five teams at the bottom rather than the five teams at the top. This already gives them a boost of at least +0.010. Now imagine if the schedule for starters and backups were switched, with the high TPE goalies getting to play with well-rested stamina against the teams who shoot with less accuracy. I would imagine that the correlation between goalie TPE and save % would increase by quite a bit in that scenario. By how much? It's tough to know. I just think it's important to remember that the data we're looking at isn't created in a perfect test environment.

Great work as always though. You're awesome, Matty Heart

PremierBromanov · 08-16-2022, 12:16 PM

It is interesting that low TPE goalies have such variance and high TPE goalies have much less. is this due to sample size or is it inherent in the sim?

PremierBromanov · 08-16-2022, 12:31 PM

the other thing I'd mention is that i'm not sure what a comparable skater graph would look like. We know that higher TPE typically results in better performance, but this is sort of a self fulfilling prophecy as higher TPE forwards are usually rewarded with higher ice time, which in my experience is the biggest factor in individual player success. We've seen rookies paired with elite players in an effort to protect them, and we've seen those players perform very well. This is an obvious truth: players dont play alone and their lines/team matter more than their individual skillsets. We couldnt even begin to measure individual success of skaters without something akin to hockeyviz, which is an attempt at isolated impact of a given player by removing common factors (other players, coaches, schedule, etc).

Of course, we know that low TPE rosters typically perform very poorly. Any rebuilding team knows this. But, we also know that it's not a linear, one to one relationship. HAM should be the obvious evidence here, or BAP for that matter. That is to say, you can't BUY a cup.

But, where is the disconnect between skaters requiring a good team to do well and goalies requiring a good team to do well? What is the actual difference? I dont think anyone would say that a similar graph of skater TPE would see so flat a slope, but who can say what the actual slope would be? Once we know that difference, we can have a more accurate discussion.

As it stands, the fact that goalies require good teams to play well -- in and of itself -- is not exclusive to goalies. It's the grade of the slope that matters, and without a comparison we have no idea what we are shooting for. Every skater in this league is beholden to their team composition. If we want to fine-tune how good it feels to play a goalie, we have to know exactly how good it feels to not be a goalie.

***Anthique*** · 08-16-2022, 01:45 PM

08-15-2022, 10:57 PMMatteo Wrote: All of the goalies above this mark are fairly grouped together, with the highest save percentage being a .908 and the lowest not terribly far behind at a .887. Looking below this mark tells a vastly different story, the highest save percentage of a goalie with over one game played being a .930 (and a .952 with only a single game played) and the lowest being a .862. This is over three times more variance between these two ranges.

Hey it's me at the top of the chart with the only game played! Smile

For real thanks for pointing out those stats, we were guessing all this, but now with those stats, we know that to a certain tpe number, our stats will not get better.
That's a concern because if I have to be bad my entire career or wait that my team will have success to have some of it on my own, the fun of being a goalie will disappear.
There is also a possibility of having less goalie knowing all that so what are we gonna do if nobody wants to be in net?

Matteo · 08-16-2022, 04:12 PM

08-16-2022, 12:06 PMKeven Wrote: Obviously it would be nice if there was a stronger correlation between TPE and save %. It's unfortunate that 155 TPE freshly created goalies can survive and win games at the SHL level. However, it's important to remember that the data you're looking at isn't perfect. It's worth keeping in mind that the schedule for backups is advantageous. They're usually well rested and are only played against the lowest teams in the standings.

In terms of shooting percentage, the worst 5 teams (PHI, EDM, MIN, TEX, NEW) ranged from 9.6% to 10.1%. This means over the course of the entire season, the goalies they faced were rewarded with save percentages of 0.899 to 0.904. Now maybe this is due to the opposing goalies always standing on their heads during these games, but I think it's much more likely that this can be attributed to the play of the skaters. Compare this to the top 5 teams for shooting percentage (CHI, TOR, TBB, LAP, MTL) who ranged from 11.2 to 12.5%. Goalies playing against these teams would be getting save percentages of 0.888 to 0.875. I'd argue that the lower TPE goalies in the league had much greater chances of facing the five teams at the bottom rather than the five teams at the top. This already gives them a boost of at least +0.010. Now imagine if the schedule for starters and backups were switched, with the high TPE goalies getting to play with well-rested stamina against the teams who shoot with less accuracy. I would imagine that the correlation between goalie TPE and save % would increase by quite a bit in that scenario. By how much? It's tough to know. I just think it's important to remember that the data we're looking at isn't created in a perfect test environment.

Great work as always though. You're awesome, Matty

Well said and a very good point. When I started this process one of the first walls I ran into was how to adjust for the quality of opponents. I had considered taking a weighted average of each team's goalies' TPE, where the starter and backup are represented by minutes played, then taking each team save percentage like I did in the point percentage chart. I decided not to do it like this since it pushed every team to the middle both TPE-wise and save percentage-wise. In the end I couldn't find a solution for what you said, I'm not sure there is one without having a perfect test environment (which thankfully we can't have now that test simming is gone). I went for a quick and dirtyish method of just throwing them all up on the chart and seeing what it looked like, but I didn't want to have a one game callup be represented the same as a 54 game starter, so I went with the bubble chart.

Stamina is another thing I didn't talk about, I just figured that goalies expecting to start the majority of their games would put more TPE into stamina, but having high stamina sacrifices putting TPE in other areas so I should have considered it more.

ALSO Kev, you're awesomer, thanks for the response Heart

Matteo · 08-16-2022, 04:25 PM

08-16-2022, 12:31 PMPremierBromanov Wrote: the other thing I'd mention is that i'm not sure what a comparable skater graph would look like. We know that higher TPE typically results in better performance, but this is sort of a self fulfilling prophecy as higher TPE forwards are usually rewarded with higher ice time, which in my experience is the biggest factor in individual player success. We've seen rookies paired with elite players in an effort to protect them, and we've seen those players perform very well. This is an obvious truth: players dont play alone and their lines/team matter more than their individual skillsets. We couldnt even begin to measure individual success of skaters without something akin to hockeyviz, which is an attempt at isolated impact of a given player by removing common factors (other players, coaches, schedule, etc).

Of course, we know that low TPE rosters typically perform very poorly. Any rebuilding team knows this. But, we also know that it's not a linear, one to one relationship. HAM should be the obvious evidence here, or BAP for that matter. That is to say, you can't BUY a cup.

But, where is the disconnect between skaters requiring a good team to do well and goalies requiring a good team to do well? What is the actual difference? I dont think anyone would say that a similar graph of skater TPE would see so flat a slope, but who can say what the actual slope would be? Once we know that difference, we can have a more accurate discussion.

As it stands, the fact that goalies require good teams to play well -- in and of itself -- is not exclusive to goalies. It's the grade of the slope that matters, and without a comparison we have no idea what we are shooting for. Every skater in this league is beholden to their team composition. If we want to fine-tune how good it feels to play a goalie, we have to know exactly how good it feels to not be a goalie.

It would definitely be interesting to compare this to a similar skater graph, if the stats we can get from FHM 8 allow someone smarter than me to have a way to make that it that would be really cool to look at. First thing that came to mind was points vs TPE but obviously then defensive players aren't represented well, then maybe I thought relative cf% vs TPE but that's also far from perfect.

Skaters definitely do benefit from better teams around them as well, I would guess not as much as goalies do, but I was also wrong about both of my predictions before and I've also only been a goalie so far so I'm sure I have biases there too.

PremierBromanov · 08-16-2022, 05:41 PM

08-16-2022, 04:25 PMMatteo Wrote:
08-16-2022, 12:31 PMPremierBromanov Wrote: the other thing I'd mention is that i'm not sure what a comparable skater graph would look like. We know that higher TPE typically results in better performance, but this is sort of a self fulfilling prophecy as higher TPE forwards are usually rewarded with higher ice time, which in my experience is the biggest factor in individual player success. We've seen rookies paired with elite players in an effort to protect them, and we've seen those players perform very well. This is an obvious truth: players dont play alone and their lines/team matter more than their individual skillsets. We couldnt even begin to measure individual success of skaters without something akin to hockeyviz, which is an attempt at isolated impact of a given player by removing common factors (other players, coaches, schedule, etc).

Of course, we know that low TPE rosters typically perform very poorly. Any rebuilding team knows this. But, we also know that it's not a linear, one to one relationship. HAM should be the obvious evidence here, or BAP for that matter. That is to say, you can't BUY a cup.

But, where is the disconnect between skaters requiring a good team to do well and goalies requiring a good team to do well? What is the actual difference? I dont think anyone would say that a similar graph of skater TPE would see so flat a slope, but who can say what the actual slope would be? Once we know that difference, we can have a more accurate discussion.

As it stands, the fact that goalies require good teams to play well -- in and of itself -- is not exclusive to goalies. It's the grade of the slope that matters, and without a comparison we have no idea what we are shooting for. Every skater in this league is beholden to their team composition. If we want to fine-tune how good it feels to play a goalie, we have to know exactly how good it feels to not be a goalie.
It would definitely be interesting to compare this to a similar skater graph, if the stats we can get from FHM 8 allow someone smarter than me to have a way to make that it that would be really cool to look at. First thing that came to mind was points vs TPE but obviously then defensive players aren't represented well, then maybe I thought relative cf% vs TPE but that's also far from perfect.

Skaters definitely do benefit from better teams around them as well, I would guess not as much as goalies do, but I was also wrong about both of my predictions before and I've also only been a goalie so far so I'm sure I have biases there too.

Oh we're smart enough, baby. The question is: are we lazy not enough?

***HoltzFan10*** · 08-17-2022, 02:03 PM

08-15-2022, 11:02 PMZoone16 Wrote: Lol. Glad I'm retired now.

same LOL

***Reno*** · 08-17-2022, 03:21 PM

i sux :(

Matteo · 08-17-2022, 03:25 PM

08-17-2022, 03:21 PMRenoJacksonHS Wrote: i sux :(

Says the goalie who's fourth place in save percentage among starters :simon:

PremierBromanov · 08-17-2022, 03:26 PM

08-17-2022, 03:25 PMMatteo Wrote:
08-17-2022, 03:21 PMRenoJacksonHS Wrote: i sux :(
Says the goalie who's fourth place in save percentage among starters :simon:

maybe he was talking about something else

Matteo · 08-17-2022, 03:32 PM

08-17-2022, 03:26 PMPremierBromanov Wrote:
08-17-2022, 03:25 PMMatteo Wrote: Says the goalie who's fourth place in save percentage among starters :simon:

maybe he was talking about something else

[Image: ur,pin_small_front,wide_portrait,750x1000.jpg]

*JURT* · 08-17-2022, 03:37 PM

"A look at goalies this season"

not a single pic of Matty to look at

visible disappointment

FRIENDS DON'T LIE

***Reno*** · 08-17-2022, 06:31 PM

08-17-2022, 03:26 PMPremierBromanov Wrote:
08-17-2022, 03:25 PMMatteo Wrote: Says the goalie who's fourth place in save percentage among starters :simon:

maybe he was talking about something else

ah sorry i hit enter too soon.

i sux :( at minecwaft because i don't have enough dedicated RAM

Navigation

Extra Menu

About us