The great Kadri vs Kerfoot debate thread.

[666]

Kadri's team shot at 13.6% when he was on the ice, or 5% higher than his next best full season. The only time he ever had that kind of luck was in the shortened season where he put up close to PPG over 40-odd games.

Is this something you expect from him going forward, or is it more reasonable that it's another 40-50 game streak of unsustainable luck like we've already seen him come crashing back to earth from once before? At his career average on-ice shooting%, he'd be sitting at roughly 27 goals for.

That's not to say expected production > actual production in terms of the value he brought this season, but if you expect his line to continue shooting at 13+% you must think that Kadri is unquestionably the most effective offensive player in the league. Or... it's luck over a small sample size.

Actually, what I'm saying is watch the games BUT if you're gonna use stats don't obsess over shot attempt based stats. Zeke literally attempted to explain Kadri's entire career basing it on a single number or two. That's just absurd.

Click to read more...

 

[Martin Skoula]

Where did you find this? Link?

I was thinking it would be interesting to see who the others are in the top 20. If they're randoms then the stat doesn't carry much weight IMO but if they're mostly elite players, then that's a different story of course.

It's pretty clear that the Avalanche's real pecking order is actually:

Nichushkin
Jost
Graves
Kadri

Followed by MacKinnon at 7th and Rantanen at 16th on the team.

If you want to see what GF% is measuring, try sorting by PDO (luck). Weird how the top-3 is still Nichushkin Graves and Kadri, must be a coincidence.

Actually, what I'm saying is watch the games BUT if you're gonna use stats don't obsess over shot attempt based stats. Zeke literally attempted to explain Kadri's entire career basing it on a single number or two. That's just absurd.

I watched probably 80+% of every game Kadri has played prior to this season, what part of that was inaccurate? Kadri started his career with a lucky year where he was in better shape going into the season than everyone else coming off the lockout, got underused by Wilson, peaked as a match-up 30/30 center, and fell off hard in his last year here. Consider that he's having a lockout-year amount of shooting luck right now as well as getting huge PP minutes and he's still only just barely matching what he did here in his peak with less PP time, worse linemates, harder minutes, and worse luck.

If the numbers clearly don't line up with what's actually happening, look into what you're missing. In this case, they line up pretty well with what I saw. You're the one who pulled up a heavily luck-dependent stat to try to show he's actually having a great year.

 

[666]

It's pretty clear that the Avalanche's real pecking order is actually:

I watched probably 80+% of every game Kadri has played prior to this season, what part of that was inaccurate? Kadri started his career with a lucky year where he was in better shape going into the season than everyone else coming off the lockout, got underused by Wilson, peaked as a match-up 30/30 center, and fell off hard in his last year here. Consider that he's having a lockout-year amount of shooting luck right now as well as getting huge PP minutes and he's still only just barely matching what he did here in his peak with less PP time, worse linemates, harder minutes, and worse luck.

If the numbers clearly don't line up with what's actually happening, look into what you're missing. In this case, they line up pretty well with what I saw. You're the one who pulled up a heavily luck-dependent stat to try to show he's actually having a great year.

I'm trying to help you understand that you are using stats the wrong way by pointing out stats that you are ignoring. Hockey cannot be measured the way that you are trying to measure it. With that being said:

You and Zeke keep mentioning the PP. I posted 5v5 stats just as he does to keep things equal. Kadri gets even better if we include the PP but I didn't include it and neither do you because it makes him look even better.

In 18-19 playing for the Leafs, Kadri had a 5v5 CF% of 53.2%, his best season ever, yet somehow that's supposed to be a horrible season but only after it's converted into magical expected goals. You are taking shot attempts (for all 12 players on the ice) and trying to convert them into expected goals (while ignore actual goals) and then assign that to an individual. Do you not see how absurd that is? There's a reason that shot based stats do not correlate well with success. It's because NOTHING correlates well with success because hockey can't be measured that way.

If you think that Kerfoot is better than Kadri then you are a fool. Watch the games.

 

[Martin Skoula]

I'm trying to help you understand that you are using stats the wrong way by pointing out stats that you are ignoring. Hockey cannot be measured the way that you are trying to measure it. With that being said:

You and Zeke keep mentioning the PP. I posted 5v5 stats just as he does to keep things equal. Kadri gets even better if we include the PP but I didn't include it and neither do you because it makes him look even better.

In 18-19 playing for the Leafs, Kadri had a 5v5 CF% of 53.2%, his best season ever, yet somehow that's supposed to be a horrible season but only after it's converted into magical expected goals. You are taking shot attempts (for all 12 players on the ice) and trying to convert them into expected goals (while ignore actual goals) and then assign that to an individual. Do you not see how absurd that is? There's a reason that shot based stats do not correlate well with success. It's because NOTHING correlates well with success because hockey can't be measured that way.

If you think that Kerfoot is better than Kadri then you are a fool. Watch the games.

Where did I use expected goals? I just pointed out that his production this season is incredibly unsustainable going forward and is the result of puck luck. It's almost the exact same sample size as the lockout shortened year, and he has almost the same oiSH% as he had in the lockout shortened year. The same way it would have been foolish to expect Kadri to be a PPG player going forward based off the lockout year, it's incredibly foolish to expect Kadri to be a 60 point forward going forward unless he gets even more PP time somehow. He just isn't a strong ES producer when his line is shooting at normal rates.

No group of 5 players in the league shoots at 13% longterm (maybe if McDavid and Draisaitl got an equally talented linemate and played out of their minds all year), least of all Kadri and the Av's middle 6 wingers. Yes, his actual goals for/against are good this year, but you're a fool if you expect that to be the norm going forward, just like you're a fool if you think his last year with us was in any way offensively productive at ES. Kerfoot is not better than Kadri as a whole, but Kerfoot is capable of having similar offensive seasons at ES as Kadri had in his last year here. Frankly, most middle-6 centers are capable of having those kinds of offensive years consistently. There is a drop off from Kadri to Kerfoot as overall players but it's not nearly big enough to hurt us or be upset about.

 

[Ziggdiezan]

People's fascination with xGF% is pretty interesting. It uses data from like a decade or more ago in it's complex regression analysis so it doesnt account for rule changes like goalie pads etc.

The other thing that is a problem is it is very self fulfilling. For example in the shot quality calculation they assign a coefficient based on who the shooter is. This makes sense in that good shooters are expected to score more. However what happens is good scorers do shoot more and score more on their own without inflating it by multiplying it by a coefficient. This causes a self reinforcing behavior for good players and a self detrimental behavior to bad shooters, basically amplifies the good/bad. This is illustrated by if a good shooter takes a muffin shot it will be rated as a higher xGF than another player in the same position who is a worse shooter traditionally regardless if it was a much better shot or a 2 on 1 etc.

This brings me to one of the bigger issues with xGF, it doesnt incorporate the relative positioning of teammates or opponents. This is due to that data not having been tracked previously so it makes sense. The biggest issue that arises is that a cross seam pass from the hash marks would be rated similarly if the player just skated in under pressure and took a shot in that same position. It has a lot of issues (as do any stat), however it definitely is important to look at all stats as they provide context assuming you understand their limitations. xGF can tell you if a player is getting really lucky, really unlucky, if they are providing a lot of shots from a dangerous locations etc.

 

[Gary Nylund]

Natural Stat Trick as Corsica seems to have let it's domain expire.

OK I just had a look. Kadri being 17th in the NHL, sounds great until I see that Matthews is in 69th place, Malkin is 90th and McDavid comes in at 246. Add in the fact that half the names ahead of Kadri I don't even recognize and I have choice but to conclude that this particular stat is as useless as they come. I'd ask why you think this stat is useful for evaluating players but I can't even imagine what you could possibly say so I'll just leave it at that.

 

[666]

Where did I use expected goals? I just pointed out that his production this season is incredibly unsustainable going forward and is the result of puck luck. It's almost the exact same sample size as the lockout shortened year, and he has almost the same oiSH% as he had in the lockout shortened year. The same way it would have been foolish to expect Kadri to be a PPG player going forward based off the lockout year, it's incredibly foolish to expect Kadri to be a 60 point forward going forward unless he gets even more PP time somehow. He just isn't a strong ES producer when his line is shooting at normal rates.

No group of 5 players in the league shoots at 13% longterm (maybe if McDavid and Draisaitl got an equally talented linemate and played out of their minds all year), least of all Kadri and the Av's middle 6 wingers. Yes, his actual goals for/against are good this year, but you're a fool if you expect that to be the norm going forward, just like you're a fool if you think his last year with us was in any way offensively productive at ES. Kerfoot is not better than Kadri as a whole, but Kerfoot is capable of having similar offensive seasons at ES as Kadri had in his last year here. Frankly, most middle-6 centers are capable of having those kinds of offensive years consistently. There is a drop off from Kadri to Kerfoot as overall players but it's not nearly big enough to hurt us or be upset about.

Well I'm glad that you aren't a Corsi fanatic like Zeke. As for puck luck and oiSH% I also appreciate that you mention that it's a group of players being measured not an individual. I also agree that Kadri's oiSH% is a bit high but even if you knock that down to 8 or 9 Kadri is still going to have a 5v5 GF% over 55% this season.

While Kerfoot might be able to have "a similar offensive season" as Kadri did last year, we are back to where this started where there is much more to Kadri's game than offense and alot of it is hard to impossible to measure with "advanced stats".

My beef is with Zeke misusing xGF% as some sort of all encompassing stat which it is not.

 

[zeke]

Zeke, this post of yours is hilarious. Your obsession with shot based stats (impacted by all players on the ice) that are converted into expected goals is bizarre.

But much more importantly why do you ignore actual goal scoring “advanced stats”? This game is not won based on attempted shots taken. It’s based on goals scored.

This season 5v5 Kadri was on the ice for 40 goals for and 22 goals against for a GF% of 64.5. Guess what? That’s higher than MacKinnon, 18th in the entire league and Kadri’s best season ever.

Dude, it's about goals, not attempted shots.

It is for the past, not going forward.

And again, kadri on pace for 35 espts to kerfoot's 32.

 

[Martin Skoula]

Well I'm glad that you aren't a Corsi fanatic like Zeke. As for puck luck and oiSH% I also appreciate that you mention that it's a group of players being measured not an individual. I also agree that Kadri's oiSH% is a bit high but even if you knock that down to 8 or 9 Kadri is still going to have a 5v5 GF% over 55% this season.

While Kerfoot might be able to have "a similar offensive season" as Kadri did last year, we are back to where this started where there is much more to Kadri's game than offense and alot of it is hard to impossible to measure with "advanced stats".

My beef is with Zeke misusing xGF% as some sort of all encompassing stat which it is not.

Kadri at his career oiSH% would be at 27 GF to 29 GA. His season looks better than it really is because of his PP scoring which he wouldn't get here with Nylander taking over the last 1st unit spot finally.

There's more to Kadri's game than offense, but it's not consistent. Kadri is good defensively in 1v1 shadow situations like his games against McDavid. He can agitate physically and get players off their games when he's on. If he's going up against Boston's top unit, he doesn't have the defensive IQ to take away space properly, if he tries to shadow one of them the other two will burn him hard. There are contexts in which Kerfoot has a better "all around" game (not to say he's particularly great defensively, but he's smarter and faster at playing disruptive area defense).

If it comes down to choosing between playing Kadri in 3rd line minutes with a 2 million dollar winger vs Kerfoot with a 3-ish million dollar winger like Kapanen, we're probably better off with the latter considering we won't be making the most of Kadri's surplus value which comes on the PP.

 

[ShaneFalco]

Move on from Kadri, the team had to. He's gone

This team can challenge anyone in a playoff series. We just need, like any other team, a hot goalie. And this will be a huge test for Freddie. He may not be staying anyway, but if he's sub-par in the playoffs again, that would cement it

 

[666]

OK I just had a look. Kadri being 17th in the NHL, sounds great until I see that Matthews is in 69th place, Malkin is 90th and McDavid comes in at 246. Add in the fact that half the names ahead of Kadri I don't even recognize and I have choice but to conclude that this particular stat is as useless as they come. I'd ask why you think this stat is useful for evaluating players but I can't even imagine what you could possibly say so I'll just leave it at that.

Exactly, all of this stuff is silly. xGF% is even dumber. Using attempted shots by everyone on the ice it attempts to "predict" expected goals for individuals.

For Kadri 5v5 this season it "predicts" 22.42/26.26 where Kadri ACTUALLY was on the ice for 40/22. Yet Zeke uses the predicted value to value players ignoring what actually happened.

These "advanced stats" particularly anything that uses Corsi, count events for everyone on the ice EQUALLY regardless how involved the player was just like plus/ minus. It would maybe be ok if we were carefully measuring lines or teams or configurations of players but these stats then assign the values to the individuals. That's why poor players on great lines always have great advanced stats.

To top it off these stats are horrible at predicting anything because hockey can't be measured and predicted this way yet guys like Zeke eat it up and misuse it every day.

Why do I care? Because some people around here think that Zeke knows what he's doing and it infects our fan base. Further he spouts a lot of this nonsense on the main boards and it embarrases our fan base. Eventually we are actually having arguments of who is better between Kerfoot and Kadri and it's huge waste of time.

 

[X66]

Obviously Kadri is better than Kerfoot, but last I checked, Kadri wasn't traded for just Kerfoot.

He was traded for Kerfoot AND Barrie, who after Keefe took over had more value than Kadri alone IMHO.

The trade might look different down the road because the cap will likely stay flat and the chance of bringing Barrie back is not great, but who knows how Kerfoot and Kadri perform next year. So lets to about it then.

 

[ACC1224]

Obviously Kadri is better than Kerfoot, but last I checked, Kadri wasn't traded for just Kerfoot.

He was traded for Kerfoot AND Barrie, who after Keefe took over had more value than Kadri alone IMHO.

The trade might look different down the road because the cap will likely stay flat and the chance of bringing Barrie back is not great, but who knows how Kerfoot and Kadri perform next year. So lets to about it then.

Obvious to most, not all.

 

[Merrrlin]

I do miss Kadri, but I don't have the long term adoration for him that some have here. I don't love the trade, but I also don't think it will be the reason we fail if we don't win the cup.

First and foremost, we need Andy to at least be average.
Secondly, we need our PK to be elite or close to it.
Thirdly, we need some offence from our non-key guys who will be receiving better matchups and less attention.
Finally, we need our PP to be a strength, not just middling.

To me, those are the things that determine this teams success, descending from most important to least.

 

[zeke]

Exactly, all of this stuff is silly. xGF% is even dumber.
.

You can't respond with "exactly" to a post which just showed that the stat you just used was useless, fyi.

 

[Bomber0104]

Exactly, all of this stuff is silly. xGF% is even dumber. Using attempted shots by everyone on the ice it attempts to "predict" expected goals for individuals.

For Kadri 5v5 this season it "predicts" 22.42/26.26 where Kadri ACTUALLY was on the ice for 40/22. Yet Zeke uses the predicted value to value players ignoring what actually happened.

These "advanced stats" particularly anything that uses Corsi, count events for everyone on the ice EQUALLY regardless how involved the player was just like plus/ minus. It would maybe be ok if we were carefully measuring lines or teams or configurations of players but these stats then assign the values to the individuals. That's why poor players on great lines always have great advanced stats.

To top it off these stats are horrible at predicting anything because hockey can't be measured and predicted this way yet guys like Zeke eat it up and misuse it every day.

Why do I care? Because some people around here think that Zeke knows what he's doing and it infects our fan base. Further he spouts a lot of this nonsense on the main boards and it embarrases our fan base. Eventually we are actually having arguments of who is better between Kerfoot and Kadri and it's huge waste of time.

Great post.

 

[Merrrlin]

Exactly, all of this stuff is silly. xGF% is even dumber. Using attempted shots by everyone on the ice it attempts to "predict" expected goals for individuals.

For Kadri 5v5 this season it "predicts" 22.42/26.26 where Kadri ACTUALLY was on the ice for 40/22. Yet Zeke uses the predicted value to value players ignoring what actually happened.

These "advanced stats" particularly anything that uses Corsi, count events for everyone on the ice EQUALLY regardless how involved the player was just like plus/ minus. It would maybe be ok if we were carefully measuring lines or teams or configurations of players but these stats then assign the values to the individuals. That's why poor players on great lines always have great advanced stats.

To top it off these stats are horrible at predicting anything because hockey can't be measured and predicted this way yet guys like Zeke eat it up and misuse it every day.

Why do I care? Because some people around here think that Zeke knows what he's doing and it infects our fan base. Further he spouts a lot of this nonsense on the main boards and it embarrases our fan base. Eventually we are actually having arguments of who is better between Kerfoot and Kadri and it's huge waste of time.

If you are wasting energy worrying about what the other "fans" on the main boards think about you, I feel bad for you.

 

[666]

You can't respond with "exactly" to a post which just showed that the stat you just used was useless, fyi.

You're using a hammer to insert screws. I'm using your hammer to extract them so that you'll understand how to use a screws.

 

[zeke]

You're using a hammer to insert screws. I'm using your hammer to extract them so that you'll understand how to use a screws.

He showed that the number you just tried to claim was what actually matters, produces obviously nonsense results.

You responding with "Exactly!" to that is pretty amusing.

 

[Wafflewhipper]

*"Well not really" = you agreeing that Kadri's suspensions were worse, "both are bad" = equivocation. Which is ok, you like Kadri more than Kerfoot, but it was sort of a yes or no question
*Yes, Kadri is better and has been better than Kerfoot. I don't believe anyone has said otherwise. The question is, as a 3C and at cost moving forward, which player best suited the needs of the team. Kadri may have played on our PP this year, but it would have been the 2nd PP. No coach would load a PP unit with it's top three centers on a consistent basis...who would follow up? You're exposure to seeing your 4C against their 1C would be frightening. Thus Kadri's point contributions would be diminished.
*This may prove to be true. Both Kerfoot and Barrie were less effective than I'd hoped for. But, should we start up again in a month or two, will not their contributions be measured on how we perform in the playoffs (and in Kerfoot's case, the following years of his contract)? Please see below.



If you're suggesting that a 1 year sample size is insufficient in one case, does it not follow that it would be insufficient when considering Kerfoot's value?



*Agreed, I say with excitement.
*Agreed, I say with less certainty.
*There is no path to the Cup if your goalie is not in good form.
*Curiously, I am somewhat confident that this will not occur. After all the confusion and second guessing that has been the Leafs season, I've a sneaky suspicion they'll have a bit of a 'we're going to show them' attitude, that they have a 'something to prove to the world snark in their strides'.
*You and me both, let the games begin

I really hope Kerfoot gets better and you are right i am giving Kadri a bit more a pass than Kerfoot. Years of Kadri’s game kind of set a benchmark that i have never seen in Kerfoot’s game to this point. I don’t expect to either. Just two different players. Kerfoot is younger and growing his game with us now. Maybe a bit more pass from me for a larger sample is fairer. You hit the nail on the head there i believe. I really like Kadri and still wouldn’t have traded him for this deal with Colorado. I think Dubas blew the trade. We even have to give them a third round pick also. Damn seems hard to get my head around right now. Picks and prospects from Colorado would be better for my liking when i look at the entire picture. That said, Barrie was really building his game up with Keefe behind the bench. He recovered from the horrible fiasco with Babcock.
Kerfoot being a trade chip wouldn’t bother me either

 

[Gary Nylund]

Exactly, all of this stuff is silly. xGF% is even dumber. Using attempted shots by everyone on the ice it attempts to "predict" expected goals for individuals.

For Kadri 5v5 this season it "predicts" 22.42/26.26 where Kadri ACTUALLY was on the ice for 40/22. Yet Zeke uses the predicted value to value players ignoring what actually happened.

These "advanced stats" particularly anything that uses Corsi, count events for everyone on the ice EQUALLY regardless how involved the player was just like plus/ minus. It would maybe be ok if we were carefully measuring lines or teams or configurations of players but these stats then assign the values to the individuals. That's why poor players on great lines always have great advanced stats.

To top it off these stats are horrible at predicting anything because hockey can't be measured and predicted this way yet guys like Zeke eat it up and misuse it every day.

Why do I care? Because some people around here think that Zeke knows what he's doing and it infects our fan base. Further he spouts a lot of this nonsense on the main boards and it embarrases our fan base. Eventually we are actually having arguments of who is better between Kerfoot and Kadri and it's huge waste of time.

OK I just had a quick look and while I don't know if "xGF% is even dumber". I'll settle for just plain old dumb and IMO it's also useless for player evaluation. So your point is taken, let's put it that way. I'm not sure as it was a while ago but there was one stat I believe Zeke was touting as the best as it "considers everything" and I think it might have been xGF%. I will try to remember that and if I happen to notice him talking as if that stat is valuable for player evaluation, I will be sure to call BS.

I personally like stats, I think they can be useful but agree that Zeke relies on them too much. So many of his posts consist of numbers and nothing else (as if the numbers he posts tell the whole story and they never do).

Not worried about what people on the main boards think though. At least it doesn't matter to me. Lot's of dump posters there, no shortage of them here either.

 

[Gary Nylund]

He showed that the number you just tried to claim was what actually matters, produces obviously nonsense results.

You responding with "Exactly!" to that is pretty amusing.

To be fair (to 666), I will point out that you were posting xgf% numbers for Kadri and IMHO, xgf% are equally useless and if you're using them to evaluate Kadri then you are on the wrong track.

 

[zeke]

To be fair (to 666), I will point out that you were posting xgf% numbers for Kadri and IMHO, xgf% are equally useless and if you're using them to evaluate Kadri then you are on the wrong track.

With all due respect, this area is very well researched, and I don't really have to defend the usage. I continue to use the best stats we have at our disposal.

But we don't have to get stuck in another stats debate.

Nobody here has yet been able to explain the most relevant comparison - 3rd line C on the leafs vs 3rd line C on the leafs:

Even Strength Paces

Kadri 18-19: 13gls/35pts, -2 (team #10sv%)
Kerfoot 19-20: 11gls/32pts, -2 (team #21sv%)

And nobody critiquing Kerfoot has mentioned the one clear skill advantage Kerfoot has over Kadri, which also happens to make it unsurprising that he puts up much better possession numbers.

 

[Gary Nylund]

With all due respect, this area is very well researched, and I don't really have to defend the usage. I continue to use the best stats we have at our disposal.

You don't have to defend anything, this is true of course.

BTW, McDavid ranks 386th in the NHL in xGF%. But hey, if you feel that this is one of "best stats we have at our disposal" for player evaluation, you're certainly entitled to that opinion. And i'm entitled to have the opinion that if you think this stat is a useful tool for player evaluation (never mind among "the best stats we have") then it's difficult to take seriously anything you have to say about statistical analysis.

Edit - out of curiosity, I looked up Sid the Kid, he ranks 217th in xGF%. I mean you do what you like but as far as player evaluations go, sGF% seems to be as useless as it gets. If that's one of "the best tools" we have then good luck to you.

 

[Mess]

Here is the perfect thread to discuss the traded and all the other fine stuff.