correct. the average fans' eye tests are uneducated, undisciplined, and biased. the stats have none of those weaknesses.
You have this adorable belief that you think everyone must answer your demands, even though you never answer anyone else's questions.
Nobody has ever said stats tell the whole story. You running around demanding that I answer one of your endless straw men is amusing.
I know why I look at the stats I do, and I'm very careful to add in as much context as possible to address every concern.
Again, i more than happy to consider any suggestion you may have to improve on them - suggest away.
You have no right to demand answers from me when you have left so many of my questions unanswered. It's funny you keep asking this - do you seriously believe that your numbers cover everything? Just to (hopefully) shut you up, I will answer. There are many things missing in your numbers, too many to list. And there are stats which could add more context which aren't even complied as far as I know so the list of ways to improve would be a long one.
I will list just one thing, but that should be plenty. Ice time, that's it.
The beautiful thing about ice-time is that it's a composite of everything else. Teams have a bunch of people working on gathering/compiling/interpreting stats, they have trainers and assistant coaches who report their findings to the head coach who in addition to all that, sees the players on a daily basis in games and in practice. These coaches are highly paid professionals who are considered to be the best in their field and they have access to all kinds of information that we the fans don't have.
And they want to win so the better the player is, the more ice time they get, simple as that. If we had to judge players by a single stat (not saying we do), I can't think of a better one the ice time and I think it does a damn good job too. You have ignore this completely in your analysis and it's a pretty glaring omission.
There are a few minor issues but that is true for any stats. For one thing, Dmen get more ice time than forwards so tough to compare them but that's an issue for many stats so not a big deal. And of course if a player is playing hurt his ice time might drop so there's that as well. The main one is that different coaches/teams have slightly different philosophies as far as how much ice time top players get so you might have to adjust the numbers a bit in some cases (again, this is true for many stats) but
what is always true as a general rule is that the better the player, the more ice time he gets. I only had a peek at two teams and I'm so sure this is true for every team that I didn't bother checking further.
The Leafs 4 top forwards by ice time are Matthews, Marner, Tavares and Nylander. No surprise there.
For EDM it's Draisaitl, McDavid, RNH and then a huge gap until the next forward. No surprise there.
Ice time is especially illuminating in this case because Kadri/Kerfoot flipped teams without any substantial changes in ice time so we don't have to worry about the potential issue of different teams doing things differently.
In 17-18, Kadri played 16.46 per game, Kerfoot played 13.27.
In 18-19, Kadri played 16.11 per game, Kerfoot played 14.53.
In 19-20, Kadri played 17.26 per game, Kerfoot played 14.46.
Those numbers speak for themselves and the conclusion is crystal clear, Kadri is the better player by a good margin. He plays more on special teams and at even strength (look it up if you don't believe me) - it doesn't matter what the situation is, Kadri is the better player, period. And this isn't the opinion of the "average fan", this is the opinion of 3 different NHL coaches during this time span coaching 2 different NHL teams who all say the same thing - Kadri>>> Kerfoot. Spin it any way you like but the facts speak for themselves and leave no room for doubt.
Yes, the more you follow and use the numbers, the better you get at understanding them and deciphering what is more important and descriptive.
In the case of that statline there - you can see how great a season Kadri had on every level. it is missing one key stat which i always use when available and that is TOIqoc - and in this case I remember it clearly, as him and Matthews were neck and neck that year with solid (but not elite) 1st line qoc. So we add that in and we see:
1. Usage: low minutes compared to most 1st liners, but comparable qoc to most of them.
2. Production: Getting above 2p60 is quality 1st line produciton, though not near the elite guys. 8.7oish% is totally normal.
3. Overall Impact: obviously 52.7xgf% is well above average (i.e. 50%), as is +2.8 relative to team (i.e. +0.0). These are very, very good numbers for a center facing legit 1st line level matchups.
So based on those numbers we have a guy that was being used in a borderline 1st line C role, and performaing like a quality 1st line C in that role. Just an excellent season for Naz. Unfortunately, he's been on a steep decline since, despite his oish-inflated offense this year.
That's a lot of words but you didn't answer the question. How can you not answer this fundamental question and instead, say that
You have this adorable belief that you think everyone must answer your demands, even though you never answer anyone else's questions.
It would seem this applies to you perfectly but I'll give you a chance to redeem yourself. Here is the question again:
when you look at these stats you love, xgf%, rel, p/60, oish etc., how do you use them to come up with a final grade so that you can compare players? How much weight does each stat carry - are they all equally important or are some more important than others? What's your formula, your algorythym? Or do you just have a quick peek at the numbers and use your eye test to say Kerfoot is better? I'm going to guess you have nothing resembling a formula and you use your eye test, tell me I'm wrong.
all xgf% is is the best measure we have that compares the quality and amount of offense generated by the player's team when he is on the ice vs. the quality and amount of offense generated by the other team. It's value is obvious, and it would be silly to dismiss it. Of course, it is just as silly to use it in it's raw form and not adjust it for quality of team, quality of competition, zone deployment, and score effects.
Dom at the athletic has a pretty good model. I trust it for the most part. He seems to be the only one that manages to factor in qoc effectively, though it's not public how he does.
But even then, that single number is never enough for me. I'm after more than just overall value - I like my stats to describe the type of player we're looking at. That's why I'm happy to give a package that describes all of usage, production, and overall impact.
Even with baseball, WAR is fine and all, but not nearly enough for me. With baseball WAR, the first thing I do is separate offense from defense, because the defensive component of WAR is a much less reliable stat than the offensive compenent. And even further, I don't quite trust the defensive war comparisons between, say, a 1B and a CF, so I prefer to compare defense based on the best defensive metrrics we have comparing players at the same position to each other. And even offensively, I like to include the usage (i.e. PAs and batting order slot), and the type of hitter we're looking at - i.e. average, walks, strikeouts, power. Comparing players simply on a WAR basis for me simply doesn't tell us enough.
Just remember, the stats I use - which you call a superficial analysis - include a detailed description of that player's role and usage, his production in every game situation and whether it's sustainable, and includes an overall impact competent that demonstrates how the team does when he's on the ice, while factoring out team and opponent quality.
It does all of that. If you think that is superficial compared to the majority of the eye test posts in this thread, you are just lying to yourself.
Before saying that my stats are superficial, ask yourself how many of the eye test posts refer to all of those elements of play in their comparisons.
I never said your stats are superficial, just your analysis.
My favorite part about the Kerfoot v Kadri comparison is that I KNOW THE EYE TEST WILL ALWAYS FAVOR KADRI.
Kadri is a great eye candy. He's got the 3 things the eye test sees the easiest - goals, dangles, big hits. Those are the easiest thing for any eye test to see, so he will always stand out to an average viewer. Kerfoot is the exact opposite of that. He's a passer not a shooter, he doesn't dangle too much, and he doesn't hit. He is precisely the kind of guy that an eye test will have a hard time assessing.
Meanwhile, Kadri's weaknesses and Kerfoot's strengths are the hardest thing for the typical eye test to appreciate. i.e. where a player is away from the puck, how his speed allows him to stretch defenses, get back and maintain defensive positioning, and recover from being out of position. All these are things which Kadri is very weak at (legitimately a liability in these areas at this point), while these are all things that Kerfoot is very good at.
This is a classic comparison between a guy that the eye test is biased towards, and a guy that it's biased against.
And yes, Kadri will also always be a better PP player with his skillset, which helps his topline numbers look even better. In this case, though, Kadri was the worst guy on our top PP unit and our PP is actually better with him gone, so I don't even care about that.
But, again, the funny thing is here we don't need any fancy stats at all to see something which you still haven't yet to explain - their comparative performances in the exact same role - as 3rd line C for these Leafs:
Even Strength, as #3c for these Leafs:
Kadri 18-19: 13:35, 12gl/34pt pace, -2
Kerfoot 19-20: 13:09, 11gl/32pt pace, -2
How does your eye test explain this? Because my fancy stats explain it quite easily.
These are nothing more than simple numbers which provide a small piece of a larger puzzle. On their own, they don't mean much at all so what's to explain?
BTW +- has been widely discredited as a tool for player evaluation. For someone who claims to know a lot about stats, I'm surprised you don't know this. You also claimed that you "always use the same numbers", what happened to that?
One final note re. your hypocrisy. In post #39, you dismissed my "methodology" with this comment:
You don't actually doubt that your one-click leaderboard search might not actually disprove all stats, of course.
But just before that in post #19 you were just fine with the exact same "methodology" when my conclusions suited yor purposes when you had this to say:
He showed that the number you just tried to claim was what actually matters, produces obviously nonsense results.
You claim that you're consistent and that you always use the same stats is clearly false, you're more than happy to use whatever numbers support your case but when they don't suit your purpose, you claim they're valueless. The truth is that you have not shown any consistency whatsoever. In just in this one thread, you have also used raw PTS and the laughable +- when it suits your purposes so this 3rd thing is just the icing on the cake. Strike 3 - you're a hypocrite.
The most important thing in statistical analysis IMHO is intellectual honesty, something which you don't seem to have. I don't mind you posting stats, I'm always happy to read what you post. And your write-up on Kadri wasn't bad at all, I'm more than happy to give credit when credit is due. You may not remember this but you and I have been on the same side of various discussions so there is no reason for you to act as if I'm somehow out to get you, or that I want you to stop posting stats as you've said before. Post what you like my good man, it's a forum for discussion after all. But when you start making ridiculous claims like this Kerfoot>Kadri nonsense, and making claims about consistency and objectivity which are clearly false I reserve the right to call you out on it.
Have a nice night!
