Post-Game Talk: Kraken 4 - Jets 3

[Gm0ney]

First line scoring more goals?

YOU'RE almost there

"First line scoring more goals" in general, or more goals than the opposition?

"First line more often scoring more goals than the opposition" would be ideal, agreed?

 

[SLAYER]

First line scoring more goals?

YOU'RE almost there

I mean, this is along the same lines as the "Pens win more without Crosby in the lineup" argument. Is it true? Yes, the Pens have a higher win % with Crosby out of the lineup (or at least they did at the time this original debate surfaced). Are they a better team with Crosby out of the lineup? No, absolutely not, that's silly.

However, when we look at the "should we play Ehlers with Chef" debate, the results are clear - the pairing outplays their competition by far. If we really want to analyze the situation, we need clear data for how the top two lines perform with both paired and separated. I understand the concept of spreading the talent within our lineups and having Ehlers drive his own line. Is having our top line perform poorly just to try to have Ehlers drive another line worth it? I don't know, I'd have to actually put the effort into performing the data analysis.

Also, I appreciate the effort that @Dale53130 has put into his analysis; however, no one is going to take the time to read 50 screenshots of game logs from Hockey DB. Maybe if the data was in a workable format that we could actually summarize and analyze it properly... but it's pretty useless as presented and no one is going to actually read through it. You can spam screenies of line-stats all your want, but they're pretty unconvincing.

 

[Buffdog]

"First line scoring more goals" in general, or more goals than the opposition?

"First line more often scoring more goals than the opposition" would be ideal, agreed?

"Whole team scoring more goals than the opposition" is the correct answer

I'll give you partial credit for showing your work, but I have to deduct a couple of marks for being condescending and still wrong the same time

I mean, this is along the same lines as the "Pens win more without Crosby in the lineup" argument. Is it true? Yes, the Pens have a higher win % with Crosby out of the lineup (or at least they did at the time this original debate surfaced). Are they a better team with Crosby out of the lineup? No, absolutely not, that's silly.

However, when we look at the "should we play Ehlers with Chef" debate, the results are clear - the pairing outplays their competition by far. If we really want to analyze the situation, we need clear data for how the top two lines perform with both paired and separated. I understand the concept of spreading the talent within our lineups and having Ehlers drive his own line. Is having our top line perform poorly just to try to have Ehlers drive another line worth it? I don't know, I'd have to actually put the effort into performing the data analysis.

Also, I appreciate the effort that @Dale53130 has put into his analysis; however, no one is going to take the time to read 50 screenshots of game logs from Hockey DB. Maybe if the data was in a workable format that we could actually summarize and analyze it properly... but it's pretty useless as presented and no one is going to actually read through it. You can spam screenies of line-stats all your want, but they're pretty unconvincing.

Two things are possibly correct at the same time:
1. Ehlers and schief outperform their opponents
2. Jets lose anyway

That's what the data that @Dale5313 dug up that youre too lazy to read through seems to suggest

 

[SLAYER]

Two things are possibly correct at the same time:
1. Ehlers and schief outperforms their opponents
2. Jets lose anyway

That's what the data that @Dale5313 dug up that youre too lazy to read through seems to suggest

lmao how could you tell? It's unworkable/unreadable. Buncha screenies of HockeyDB is trash for any sort of data analysis. No one is going to take the time to read through that wall of images imo. This is what "modern" or "fancy" stats can be really good at - summarizing data and presenting it in a workable/digestable format.

This is also a poor premise for the argument, a la "Pens without Crosby" or "Jets without Enstrom" debate. Correlation does not show causation.

Have the Jets won more when Ehlers plays less minutes? Maybe. Is it because Ehlers performs much better with less minutes, or is it because Ehlers' usage has traditionally been within a certain range of minutes/game? Would the Jets win more with Ehlers playing more minutes? Maybe - we can't actually know that but all of the stats I've seen indicate we should play him more because he performs very well compared to his competition. Would the Jets win less with Ehlers playing more minutes? Maybe, but no performance stats I've seen would suggest this so I wouldn't bet on it.

 

[Gm0ney]

"Whole team scoring more goals than the opposition" is the correct answer

I'll give you partial credit for showing your work, but I have to deduct a couple of marks for being condescending and still wrong the same time


Two things are possibly correct at the same time:
1. Ehlers and schief outperform their opponents
2. Jets lose anyway

That's what the data that @Dale5313 dug up that youre too lazy to read through seems to suggest

Your top line is getting the most icetime. It makes no sense to trade Line 1 at 50% GF instead of 60% so your 2nd line can score at 60% instead of 50%...

 

[Buffdog]

lmao how could you tell, it's unworkable/unreadable? Buncha screenies of HockeyDB is trash for any sort of data analysis. No one is going to take the time to read through that wall of images imo. This is what "modern" or "fancy" stats can be really good at - summarizing data and presenting it in a workable/digestable format.

This is also a poor premise for the argument, a la "Pens without Crosby" or "Jets without Enstrom" debate. Correlation does not show causation.

Have the Jets won more when Ehlers plays less minutes? Maybe. Is it because Ehlers performs much better with less minutes, or is it because Ehlers' usage has traditionally been within a certain range of minutes/game. Would the Jets with more with Ehlers playing more minutes? Maybe - we can't actually know that but all of the stats I've seen indicate we should play him more because he performs very well compared to his competition. Would the Jets win less with Ehlers playing more minutes? Maybe, but no performance stats I've seen would suggest this so I wouldn't bet on it.

Correlation doesn't necessarily show causation, but it certainly can. And innocence, that's a good starting point for investigation and analysis- something nobody has done with the data that @Dale53130 has presented

People (now including yourself) have generated possible hypothesises to try to explain why the Jets win more often when ehlers and schief don't play together and when ehlers plays limited minutes... but from where I'm sitting, it all seems like very deep cognitive dissonance/confirmation bias

Dig up the numbers, figure out why the data says what it does

 

[Buffdog]

Your top line is getting the most icetime. It makes no sense to trade Line 1 at 50% GF instead of 60% so your 2nd line can score at 60% instead of 50%...

There are 2 more lines and special teams as well

I'm still waiting for someone to explain the data dale posted

Maybe I'm just used to reading medical literature and I value hard endpoints over soft endpoints. Is a medication useful if it shows that it can lower cholesterol by 25% but causes side effects that increase all cause mortality by 400%? You're arguing for that medication right now

 

[SLAYER]

Correlation doesn't necessarily show causation, but it certainly can. And innocence, that's a good starting point for investigation and analysis- something nobody has done with the data that @Dale53130 has presented

People (now including yourself) have generated possible hypothesises to try to explain why the Jets win more often when ehlers and schief don't play together and when ehlers plays limited minutes... but from where I'm sitting, it all seems like very deep cognitive dissonance/confirmation bias

Dig up the numbers, figure out why the data says what it does

The reason no one has referenced the data that Dale presented is because he has presented it in a poor, unworkable/unusable format. The data that Dale has presented also contains no nuance or context. I'll just throw up some Hockey DB screenies and claim that my data is better or proves my point, no one will know because it's too cumbersome to actually be useful.

There is no cognitive dissonance, it's just that you're making a poor argument without further consideration for the conditions which produced the situation you're attempting to analyze. It's shallow and doesn't prove the point you're thinking it does.

I'm still waiting for someone to explain the data dale posted

Yeah, me too. I'd like him to explain how he expects anyone to find it useful.

 

[Buffdog]

The reason no one has referenced the data that Dale presented is because he has presented it in a poor, unworkable/unusable format. The data that Dale has presented also contains no nuance or context. I'll just throw up some Hockey DB screenies and claim that my data is better or proves my point, no one will know because it's too cumbersome to actually be useful.

There is no cognitive dissonance, it's just that you're making a poor argument without further consideration for the conditions which produced the situation you're attempting to analyze. It's shallow and doesn't prove the point you're thinking it does.



Yeah, me too. I'd like him to explain how he expects anyone to find it useful.

OK, but in the meantime you realize that it's impossible to take anyone seriously who says "I refuse to look at the data but I don't believe it"

If you're curious, dig it up yourself. If you aren't, then obviously your opinion isn't very valuable to people that actually looked through it in the format it was presented

 

[Gm0ney]

There are 2 more lines and special teams as well

I'm still waiting for someone to explain the data dale posted

Maybe I'm just used to reading medical literature and I value hard endpoints over soft endpoints. Is a medication useful if it shows that it can lower cholesterol by 25% but causes side effects that increase all cause mortality by 400%? You're arguing for that medication right now

I think this was already explained somewhere. Scheifele and Ehlers played more together in 2015-16 and 2016-17 (when the Jets were pretty bad) than they have over the last 4 seasons combined. So that might be skewing some of your W-L data.

 

[SLAYER]

OK, but in the meantime you realize that it's impossible to take anyone seriously who says "I refuse to look at the data but I don't believe it"

If you're curious, dig it up yourself. If you aren't, then obviously your opinion isn't very valuable to people that actually looked through it in the format it was presented

I'm not saying that at all. I'm saying I refuse to read through the wall of screenshots some rando posted because it's extremely difficult to analyze at best.

My opinion is not uninformed, it was based on logic and statistics. When I have a bit more time I'll throw something together... that's not 50 Hockey DB screenshots.

 

[Buffdog]

I think this was already explained somewhere. Scheifele and Ehlers played more together in 2015-16 and 2016-17 (when the Jets were pretty bad) than they have over the last 4 seasons combined. So that might be skewing some of your W-L data.

@Dale53130 did you have this data broken down season by season? I could have sworn you did

See, that's a good point @Gm0ney

Now we're having a discussion. Next step would be to look into that. You could be right