|
Post by Owen-Moncton on Dec 4, 2018 19:45:53 GMT -5
Well, you're right about one item, being another post. Beyond that, it's another story my friend. What irks me isnt that Talbot cant stop a beachball, or even that his card next year is going to be as bad or worse save for a trade out of EDM. What irks me is that an unproven asset has the ability to platoon quite comfortably with a legitimate starter. That shouldnt be the case, and if the logic applied to the engine says it should, then the logic should be re-evaluated. This isnt sour grapes, Glenn. If you think the TECH/ATH assignment is correct after ten games, fine. What I think you are refusing to acknowledge, is that limited success is not sustainable over a whole season. It's December 4th and Georgie already has 8 starts, and one of the best GAA in the league at 2.18. That alone suggests the logic might require a re-thinking going forward. And by going forward, I mean next season, not today. None of this benefits me, this year or next. My point is about making the moving parts that make this thing so engaging more realistic. Take it for what it's worth and if you think that despite all of the constructive feedback coming from numerous participants, you got it right and there's no room for improvement, fair enough. Your house, your rules. I respectfully and enthusiastically disagree with you. For the record, Georgie-boy has the 8th best GAA in the league this year behind John Gibson. He's 17th in SV% and 13th in wins. Pretty fucking stellar for a guy with 10 NHL games and a 3.15 GAA over that stretch. He was only slightly better in Hartford with a 2.98 That alone indicates that over a larger body of work, he'd regress to the norm or below. Do we really think those numbers accurately represent what he's doing to the league this year? Seriously? -out. Dane nailed it on the head in the thread below. Again, this is a "what if" simulation and what if AG with an averaged card and a fort that will allow him to prob play 40 games in the blue played in net for one of the best teams in the league. I am guessing he would have a 2.18 GAA and a .908 save percentage. I am not saying the card generation or the sim is perfect but arguments like how he is doing on the league this year for Hamilton is about as speculative as you can get. Bring me cold hard facts to chew on and digest. In the 6 or so pages, I have yet to see them. The facts are, that an average at best AHL goalie doesnt magically become the most successful goalie on your pro team because he had a decent SV% over a limited sample size. if that's what the data suggests, I think the formula needs to be revisited. High end prospects are sheltered in minor leagues because overexposure would ruin them. Tukka Rask would have been a bust in TOR had he not been traded. I challenge anyone to refute that point of contention. Anything less than 15-20 games should prohibit a STAM rating of more than 5. Far more realistic than pretending a guy with a GAA of +3.00 can platoon after 10 games.
|
|
|
Post by Glenn-Philadelphia on Dec 5, 2018 12:37:10 GMT -5
Dane nailed it on the head in the thread below. Again, this is a "what if" simulation and what if AG with an averaged card and a fort that will allow him to prob play 40 games in the blue played in net for one of the best teams in the league. I am guessing he would have a 2.18 GAA and a .908 save percentage. I am not saying the card generation or the sim is perfect but arguments like how he is doing on the league this year for Hamilton is about as speculative as you can get. Bring me cold hard facts to chew on and digest. In the 6 or so pages, I have yet to see them. The discussion I was having was NOT about how good his is/was doing....it is specifically HOW MUCH he COULD play in the GHL based upon his REAL NHL/AHL stats. Now, you all have to see there is a huge difference between actual games played NHL, versus GHL. No? Actually no. Since he played 47 games total last year (10 NHL, 37 AHL) he is on track right now to play 34.7 games for Hamilton so that would actually be a decrease in actual GP to projected GP of -12.3. According to the way stamina is calculated he is being underused.
|
|
|
Post by Glenn-Philadelphia on Dec 5, 2018 12:51:49 GMT -5
Dane nailed it on the head in the thread below. Again, this is a "what if" simulation and what if AG with an averaged card and a fort that will allow him to prob play 40 games in the blue played in net for one of the best teams in the league. I am guessing he would have a 2.18 GAA and a .908 save percentage. I am not saying the card generation or the sim is perfect but arguments like how he is doing on the league this year for Hamilton is about as speculative as you can get. Bring me cold hard facts to chew on and digest. In the 6 or so pages, I have yet to see them. The facts are, that an average at best AHL goalie doesnt magically become the most successful goalie on your pro team because he had a decent SV% over a limited sample size. if that's what the data suggests, I think the formula needs to be revisited. High end prospects are sheltered in minor leagues because overexposure would ruin them. Tukka Rask would have been a bust in TOR had he not been traded. I challenge anyone to refute that point of contention. Anything less than 15-20 games should prohibit a STAM rating of more than 5. Far more realistic than pretending a guy with a GAA of +3.00 can platoon after 10 games. 1) Sample size by it's definition means every sample is going to be test (IE all 82 games for a goalie in a season). None of our goalies have this so all are being measured based on a sample size. 2) You call 10 games limited. I explained above that this is 16% or the average full time goalies sample size. You say it is limited and I say it is statistically relevant. 3) I didn't create statistics I just begrudgingly studied them as part of my college education. I didn't create the formulas or the methods to determine statistical relevance. Someone much smarter than me did that. His name is Sir Ronald Aylmer Fisher en.wikipedia.org/wiki/Ronald_Fisher Your beef should be with him and not me. Unfortunately he died 1 day after I was born so that avenue of recourse has been closed. In case anyone cares to educate themselves on statistical relevance, this British chap seems to be on top of it. I would start here: select-statistics.co.uk/blog/importance-effect-sample-size/
|
|
|
Post by Chris-Suffolk on Dec 5, 2018 13:19:22 GMT -5
GHL is "OUR" pro league model, correct?
NHL is the PRO league. AHL is the MINOR league.
Georgie played 10 PRO games. He is on pace for 37 GHL games.
But, it really doesn't matter. You are combining his PRO and MINOR "suggestive capabilities" into one PRO type model. And, it's fine. If you feel that it is ok for someone who has never played a full season of PRO caliber hockey, and looking into the future, suggesting that a 10 game stint is enough of a sample size to suggest a players future, so be it. I personally, do not think that it is accurate to predict with certainty, that 10 games is enough of a sample size to indicate an 82 game comp.
A youngster entering an opportunity with the knowledge that he is only up for a short time, already has the frame of mind to know that even if he faulters, it will only be for a short time and he has nothing to lose. He will not have to endure a full season of Pro caliber shots, endure the travel, endure the practice times, endure the pressure, endure the physicality and everyday adventures of pro hockey. But, the GHL has determined it for him, EVEN before he EVER does it. That's the irony. I say, we wait for a player to have the experience of a full season, before they can get a full season.
|
|
|
Post by Glenn-Philadelphia on Dec 5, 2018 15:59:56 GMT -5
GHL is "OUR" pro league model, correct? NHL is the PRO league. AHL is the MINOR league. Georgie played 10 PRO games. He is on pace for 37 GHL games. But, it really doesn't matter. You are combining his PRO and MINOR "suggestive capabilities" into one PRO type model. And, it's fine. If you feel that it is ok for someone who has never played a full season of PRO caliber hockey, and looking into the future, suggesting that a 10 game stint is enough of a sample size to suggest a players future, so be it. I personally, do not think that it is accurate to predict with certainty, that 10 games is enough of a sample size to indicate an 82 game comp. A youngster entering an opportunity with the knowledge that he is only up for a short time, already has the frame of mind to know that even if he faulters, it will only be for a short time and he has nothing to lose. He will not have to endure a full season of Pro caliber shots, endure the travel, endure the practice times, endure the pressure, endure the physicality and everyday adventures of pro hockey. But, the GHL has determined it for him, EVEN before he EVER does it. That's the irony. I say, we wait for a player to have the experience of a full season, before they can get a full season. actually on pace for 34.7 as opposed to the 47 he actually played as per my post above.
|
|
|
Post by Owen-Moncton on Dec 5, 2018 16:13:15 GMT -5
GHL is "OUR" pro league model, correct? NHL is the PRO league. AHL is the MINOR league. Georgie played 10 PRO games. He is on pace for 37 GHL games. But, it really doesn't matter. You are combining his PRO and MINOR "suggestive capabilities" into one PRO type model. And, it's fine. If you feel that it is ok for someone who has never played a full season of PRO caliber hockey, and looking into the future, suggesting that a 10 game stint is enough of a sample size to suggest a players future, so be it. I personally, do not think that it is accurate to predict with certainty, that 10 games is enough of a sample size to indicate an 82 game comp. A youngster entering an opportunity with the knowledge that he is only up for a short time, already has the frame of mind to know that even if he faulters, it will only be for a short time and he has nothing to lose. He will not have to endure a full season of Pro caliber shots, endure the travel, endure the practice times, endure the pressure, endure the physicality and everyday adventures of pro hockey. But, the GHL has determined it for him, EVEN before he EVER does it. That's the irony. I say, we wait for a player to have the experience of a full season, before they can get a full season. actually on pace for 34.7 as opposed to the 47 he actually played as per my post above. Show me a backup who plays 47 -- or even 34.7 games a season --- when the starter is healthy. I'll wait here.
|
|
|
Post by Owen-Moncton on Dec 5, 2018 17:26:30 GMT -5
The facts are, that an average at best AHL goalie doesnt magically become the most successful goalie on your pro team because he had a decent SV% over a limited sample size. if that's what the data suggests, I think the formula needs to be revisited. High end prospects are sheltered in minor leagues because overexposure would ruin them. Tukka Rask would have been a bust in TOR had he not been traded. I challenge anyone to refute that point of contention. Anything less than 15-20 games should prohibit a STAM rating of more than 5. Far more realistic than pretending a guy with a GAA of +3.00 can platoon after 10 games. 1) Sample size by it's definition means every sample is going to be test (IE all 82 games for a goalie in a season). None of our goalies have this so all are being measured based on a sample size. 2) You call 10 games limited. I explained above that this is 16% or the average full time goalies sample size. You say it is limited and I say it is statistically relevant. 3) I didn't create statistics I just begrudgingly studied them as part of my college education. I didn't create the formulas or the methods to determine statistical relevance. Someone much smarter than me did that. His name is Sir Ronald Aylmer Fisher en.wikipedia.org/wiki/Ronald_Fisher Your beef should be with him and not me. Unfortunately he died 1 day after I was born so that avenue of recourse has been closed. In case anyone cares to educate themselves on statistical relevance, this British chap seems to be on top of it. I would start here: select-statistics.co.uk/blog/importance-effect-sample-size/Golf clap implied...there should be an emoji for that. I acknowledge your obvious expertise in both statistics and computer sciences, and that of the obvious brilliant minds that you've referenced. I wonder what they might know/understand about hockey, however. Gathering data and compiling it based on a formulation to create tables is one thing. testing the application against practical application is another. And I dont think any sound argument exists for rewarding Georgie with the ability to platoon with a proven starter. If your equation tells you he's good for 37 NHL games after posting average numbers over a very small sample size (sorry bud, 10 NHL games IS a small sample size for any player) then your equation needs to be reconsidered to bring it in line with what would be realistic in pro sports. Mad respect sir, for what you've created here, but this is not something you're ever going to convince me of. In what alternate universe does a team sign Dubnyk to 7.8 M a season only to play him 45 games a year? In what league does a team with a legit hi-priced starter give an unproven rookie with 10 previous starts the chance to play almost half the games? Stats be damned, none of this applies in the real world.
|
|
|
Post by Matt-Colorado on Dec 5, 2018 17:54:27 GMT -5
Mad respect sir, for what you've created here, but this is not something you're ever going to convince me of. In what alternate universe does a team sign Dubnyk to 7.8 M a season only to play him 45 games a year? In what league does a team with a legit hi-priced starter give an unproven rookie with 10 previous starts the chance to play almost half the games? Stats be damned, none of this applies in the real world. Something about Evgeny Kuznetsov being exposed to waivers because he was sent to the minors.
|
|
|
Post by Dane-Hamilton on Dec 5, 2018 17:54:39 GMT -5
actually on pace for 34.7 as opposed to the 47 he actually played as per my post above. Show me a backup who plays 47 -- or even 34.7 games a season --- when the starter is healthy. I'll wait here. Phillip Grubauer played 35 last year behind a healthy Holtby Bernier played 37 behind a healthy Varlamov Khudobin 31 behind a healthy Rask I’m sure there’s more but only ones I can think of off top of my head
|
|
|
Post by Phil-Cornwall on Dec 5, 2018 18:05:55 GMT -5
We ARE in an alternate hockey universe, where things are BASED off of real NHL stats ect BUT things can play out much much MUCH differently.
middling Defenseman Mike Weaver making 'Sam Bradford' money. Boom. Mic drop.
|
|
|
Post by Owen-Moncton on Dec 5, 2018 18:37:14 GMT -5
Show me a backup who plays 47 -- or even 34.7 games a season --- when the starter is healthy. I'll wait here. Phillip Grubauer played 35 last year behind a healthy Holtby Bernier played 37 behind a healthy Varlamov Khudobin 31 behind a healthy Rask I’m sure there’s more but only ones I can think of off top of my head Cool story. Grubauer also had FIVE NHL SEASONS as a backup prior to last year. That's 66 games over a five-year span, or 13.2 games a year. Bernier played 57 games across his first 4 seasons or an average of 14.5 games a season. Khodobin played 94 games across his first 5 years or 18.8 games per season So let's compare that against Georgie-boy and his trajectory of 35 starts in year 2 of his career. Clearly anything more than 15-6 starts for a rookie backup (youre a rookie until you've played 25 games) is pretty much unprecedented. I was incorrect. I should have said show me a rookie goalie who plays 35 games behind a healthy goalie making top dollar. What you've demonstrated by your examples is that a rookie goalie generally doesnt get more than 12-15 starts a season. All I am saying is that STAM should reflect this.
|
|
|
Post by Owen-Moncton on Dec 5, 2018 18:38:42 GMT -5
Mad respect sir, for what you've created here, but this is not something you're ever going to convince me of. In what alternate universe does a team sign Dubnyk to 7.8 M a season only to play him 45 games a year? In what league does a team with a legit hi-priced starter give an unproven rookie with 10 previous starts the chance to play almost half the games? Stats be damned, none of this applies in the real world. Something about Evgeny Kuznetsov being exposed to waivers because he was sent to the minors. Something about recent and ongoing trade negotiations including something about a firm offer coming last night which I am deliberating.
|
|
|
Post by Matt-Colorado on Dec 5, 2018 18:51:39 GMT -5
Something about Evgeny Kuznetsov being exposed to waivers because he was sent to the minors. Something about recent and ongoing trade negotiations including something about a firm offer coming last night which I am deliberating. I don’t care about the context behind why you’re burying an elite player in the minors instead of simply benching him. In your words: “in what alternate universe” Does Kuznetsov get sent down to the minors? If we’re going to cherry pick aspects of the sim that individuals feel are unfair and need to be addressed—Why not employ a waiver system for players on post ELC pro contracts. Oh shit, someone made an offer on Rantanen...off to the minors with him.
|
|
|
Post by Dane-Hamilton on Dec 5, 2018 19:11:15 GMT -5
Phillip Grubauer played 35 last year behind a healthy Holtby Bernier played 37 behind a healthy Varlamov Khudobin 31 behind a healthy Rask I’m sure there’s more but only ones I can think of off top of my head I was incorrect. I should have said show me a rookie goalie who plays 35 games behind a healthy goalie making top dollar. What you've demonstrated by your examples is that a rookie goalie generally doesnt get more than 12-15 starts a season. All I am saying is that STAM should reflect this. Mikko Koskinen stole the starting job from Cam Talbot as a “rookie”
|
|
|
Post by Owen-Moncton on Dec 5, 2018 19:26:44 GMT -5
Right. Koskinnen, the 30-year-old 6'7'' "rookie" NHL-drafted in 2007 at 31st overall with 10680 minutes played over 184 KHL games, with a career GAA of under 2.00 and a SV% of .926. 77 playoff games with a post season career GAA of 1.64 and a SV% of.938.
THAT Koskinnen? we are using the term "rookie" quite loosely here, Dane.
|
|