James Tanner: Coyotes As Dangerous as Your Grandma

Tanner be like: Arizona has played 20 games yet to get that first win but the team is great statistically, things will turn around things will be fine lol

James I know you write for this team so I guess you always try to spin the positive but maybe this team just isn't as good as you think it is. Didn't you go on record saying they are a playoff team this year?

Galchenyuk wasn't there last year and didn't do well in Montreal with the coach not really liking the kid. We have no idea if they made a good trade or a bad one yet so saying they miss his scoring is premature. Chychrun has had two serious leg injuries back to back. He may never turn out to be the player that many hoped he would be. Can't say anything about the Dvorak problem. They should have traded OEL last summer, the poor guy is being wasted on this team.

James I know you write for this team so I guess you always try to spin the positive but maybe this team just isn't as good as you think it is. Didn't you go on record saying they are a playoff team this year?

Galchenyuk wasn't there last year and didn't do well in Montreal with the coach not really liking the kid. We have no idea if they made a good trade or a bad one yet so saying they miss his scoring is premature. Chychrun has had two serious leg injuries back to back. He may never turn out to be the player that many hoped he would be. Can't say anything about the Dvorak problem. They should have traded OEL last summer, the poor guy is being wasted on this team.
- sparky

It's entirely possible that I was wrong, but I did and still do think they're a playoff team. It's not premature to say you miss the scoring of a 30 goal scoring 24 year old third overall pick.

Take three of the top six players off any team and see how well they do.

well jimmy maybe thats the problem, if those 3 are 3 of their best players maybe just maybe they aint that good of a team.

Stop with the it impossible to overcome crap. Its really boring.

You overate chychrun so bad it is amusing.

At what point do they fire Tocchet? There's only one thing consistent about this team and we all know what it is. They addressed it by rebuilding and firing a bad coach, but hired an equally poor one. The guy has proven nothing at the NHL level. Wrapped up Wayner in his gambling ring, lost with Tampa, then lost with Arizona. Time to say goodbye and let someone else have a crack.

No mention of how Strome took a penalty with a minute left with his team down by 1? If I was the coach I don't think I'd be playing him either!

You guys love to criticize Tanner but facts are:

AZ is dead last in PDO right now. That will absolutely change.

They are 12th in HDCF%

They are 3rd in SCF%

They are 4th in CF%

These numbers suggest this is a team that is in the better half of the league.

Here's a corsi logic question to wrap your head around. If I took a team and filled the forward roster with 12 zach hyman clones, would it not have excellent corsi, but terrible finish?

No one would look at that team and be like, they are out-corsi-ing everyone! They will come around!

Remember that corsi is used to predict goals, not the other way around. Goals are what matters.

Here's a corsi logic question to wrap your head around. If I took a team and filled the forward roster with 12 zach hyman clones, would it not have excellent corsi, but terrible finish?

No one would look at that team and be like, they are out-corsi-ing everyone! They will come around!

Remember that corsi is used to predict goals, not the other way around. Goals are what matters.
- The_Big_Heen

Actually shot location is what is used to predict goals.

This is a huge oversimplification.

There are too many factors needed to predict goals with the measure of certainty that people use corsi/shot location for. Shot location alone is not the be-all/end-all

The problem with “the stats” saying they are a better team is that “the stats” have no impact on results. The yotes are more of a placostamous than a playoff team.

This is a huge oversimplification.

There are too many factors needed to predict goals with the measure of certainty that people use corsi/shot location for. Shot location alone is not the be-all/end-all
- The_Big_Heen

No it isn't but it is the best tool currently being used.

The problem with “the stats” saying they are a better team is that “the stats” have no impact on results. The yotes are more of a placostamous than a playoff team.
- dcoms77

Except these stats are data points of on ice events that indicate a quality team.

This is a huge oversimplification.

There are too many factors needed to predict goals with the measure of certainty that people use corsi/shot location for. Shot location alone is not the be-all/end-all
- The_Big_Heen

Though it may be counter-intuitive, the simplest way is the best way. If you want to make short-term predictions, you need to consider a ton of variables. But over a large sample size, these things all balance out.

For example, if you take every single shot in the history of the NHL and find out what the shooting percentage is (lets say its 8%) then you know that for every 100 shots, you're bound to score eight goals, and that shot quality has already been factored in.

Of course, shot quality can vary wildly game to game, so the shorter the sample size, the more you need to factor it in. But if a player takes 300 shots and only scores 9 goals (3%) you know he got a bit unlucky because history tells us its impossible to get that many shots without some of them being dangerous. You just wouldn't be in the NHL and take 300 useless shots. 40 sure, it's possible. 100. outside chance. But you couldn't come close to leading the league in shots and be such a bad shooter that you deserved to score only 3% of the time. Not in a professional league.



The main reason people argue about stats anyways is because people using stats always assume a level of probability and the critics treat it like an either/or proposition.

Let's say the Coyotes continue to put up excellent numbers.

The Odds are, they make the playoffs. If you're in the top half of the league in all the stats that lead to goals, a reasonable probability of making the playoffs might be 60%.

But four out of ten times, they will still miss in that situation.

Conversely, a team that defies the analytics and makes the playoff might have only had a ten percent chance at making it. But they still made it.

The games obviously have to be played, and what makes them worth watching is that anything can happen. Using analytics to make better guesses at who will do well is fun, and it can be profitable if you're betting or running a team.

There will always be teams that deserve to make the playoffs who don't, and vice versa. But this doesn't disprove the ability of stats to help you make better decisions.


If you're the Coyotes right now, your main question is Are we this bad, or is this just a fluke?

Lots of teams overreact for the worse during a bad run. Smart teams stay the course and don't fall for a recency bias.

You guys love to criticize Tanner but facts are:

AZ is dead last in PDO right now. That will absolutely change.

They are 12th in HDCF%

They are 3rd in SCF%

They are 4th in CF%

These numbers suggest this is a team that is in the better half of the league.
- Feds91Stammer

To me one of the factors that this kind of analysis seems to ignore is the 82 game season. Sure, over many games, I believe all other things being equal, the stats probably prove correct, they will win more and have better results and things will change. But will they be able to do it in time to win enough games to make the playoffs? You only have 82 games, if the stats allow for enough error to go on a 10 game losing streak or have a big chunk of the year go the other way, then how useful is the information? That's not even taking into account the fact that they might not be able to maintain these numbers for whatever reason.

Im no disagreeing, I just havent seen any evidence that take the 82 game season into account and would really like to. I would like to see the sample size analysis that proves these numbers, Im guessing it's more than 82 games.

The Yotes are playing well, shots are great, its just weird to not score 5v5 and "luck" is not a good enough answer. It's the same as "puck luck" which an idea that has been around since before anyone even thought of advanced stats.

To me one of the factors that this kind of analysis seems to ignore is the 82 game season. Sure, over many games, I believe all other things being equal, the stats probably prove correct, they will win more and have better results and things will change. But will they be able to do it in time to win enough games to make the playoffs? You only have 82 games, if the stats allow for enough error to go on a 10 game losing streak or have a big chunk of the year go the other way, then how useful is the information? That's not even taking into account the fact that they might not be able to maintain these numbers for whatever reason.

Im no disagreeing, I just havent seen any evidence that take the 82 game season into account and would really like to. I would like to see the sample size analysis that proves these numbers, Im guessing it's more than 82 games.

The Yotes are playing well, shots are great, its just weird to not score 5v5 and "luck" is not a good enough answer. It's the same as "puck luck" which an idea that has been around since before anyone even thought of advanced stats.
- 13sundin13

AZ has shot 0% at 5v5 this season.

If they shoot even a modest 7% they have 9.87 more goals this season.

In that case we are looking at a 3-2 maybe 4-1 team.

If they continue playing how they are now, the pucks will go in and they will win a bunch of games.

" Once Galchenyuk, Chychrun and Dvorak are back, scoring will be much easier. " - I didn't know that they got 3 50 goals scorers.

AZ has shot 0% at 5v5 this season.

If they shoot even a modest 7% they have 9.87 more goals this season.

In that case we are looking at a 3-2 maybe 4-1 team.

If they continue playing how they are now, the pucks will go in and they will win a bunch of games.
- Feds91Stammer

Again, that doesn't at all talk about sample sizes and why analytic numbers are a large enough determining factor in an 82 games season.

All you really said here was if they scored more, they would have won. I'm sure everyone here would agree with that without any supporting numbers. What I think people are looking for in the analytics is why they didn't, will they in the future, and will it be enough to win x amount of games to make the playoffs. Which brings us back to sample sizes.

@Feds

Those team numbers are somewhat skewed by some measure of score effects (in the 3-0 games) and by playing Carlyle's Ducks twice (garbage-tier possession team without Getzlaf and Kesler).

Not gonna disagree that they're better than their results, but the underlying numbers are skewed a tad too much.

Though it may be counter-intuitive, the simplest way is the best way. If you want to make short-term predictions, you need to consider a ton of variables. But over a large sample size, these things all balance out.

For example, if you take every single shot in the history of the NHL and find out what the shooting percentage is (lets say its 8%) then you know that for every 100 shots, you're bound to score eight goals, and that shot quality has already been factored in.

Of course, shot quality can vary wildly game to game, so the shorter the sample size, the more you need to factor it in. But if a player takes 300 shots and only scores 9 goals (3%) you know he got a bit unlucky because history tells us its impossible to get that many shots without some of them being dangerous. You just wouldn't be in the NHL and take 300 useless shots. 40 sure, it's possible. 100. outside chance. But you couldn't come close to leading the league in shots and be such a bad shooter that you deserved to score only 3% of the time. Not in a professional league.



The main reason people argue about stats anyways is because people using stats always assume a level of probability and the critics treat it like an either/or proposition.

Let's say the Coyotes continue to put up excellent numbers.

The Odds are, they make the playoffs. If you're in the top half of the league in all the stats that lead to goals, a reasonable probability of making the playoffs might be 60%.

But four out of ten times, they will still miss in that situation.

Conversely, a team that defies the analytics and makes the playoff might have only had a ten percent chance at making it. But they still made it.

The games obviously have to be played, and what makes them worth watching is that anything can happen. Using analytics to make better guesses at who will do well is fun, and it can be profitable if you're betting or running a team.

There will always be teams that deserve to make the playoffs who don't, and vice versa. But this doesn't disprove the ability of stats to help you make better decisions.


If you're the Coyotes right now, your main question is Are we this bad, or is this just a fluke?

Lots of teams overreact for the worse during a bad run. Smart teams stay the course and don't fall for a recency bias.
- James_Tanner

With the huge confidence and certainty that people have in advanced stats, I'm very surprised that it's only a 60% chance. With 16 out of 31 teams in the league making the playoffs, just by having a team you have over a 50% chance by default. Giving a team an extra 8-10% using "advanced stats" as proof seems marginal, especially considering how many factors change over the course of a season.

@Feds

Those team numbers are somewhat skewed by some measure of score effects (in the 3-0 games) and by playing Carlyle's Ducks twice (garbage-tier possession team without Getzlaf and Kesler).

Not gonna disagree that they're better than their results, but the underlying numbers are skewed a tad too much.
- MaximumBone

Oh for sure it is still way too small of a sample. Personally I prefer a 10 game rolling average. I don't think anyone though would say if they continued at the numbers they have now that they won't win a bunch of games.

After reading your 500 paragraph blog before the season started, I thought for sure the Yote's would be a top 3 team in the Pacific from day one. You brainwashed me into thinking they might actually be alright but after watching 3 of their first 5 games. They look as bad as they did the first 2 1/2 months of last season. Yeah they are missing a key piece but so are a lot of the other teams.

Again, that doesn't at all talk about sample sizes and why analytic numbers are a large enough determining factor in an 82 games season.

All you really said here was if they scored more, they would have won. I'm sure everyone here would agree with that without any supporting numbers. What I think people are looking for in the analytics is why they didn't, will they in the future, and will it be enough to win x amount of games to make the playoffs. Which brings us back to sample sizes.
- 13sundin13

The underlying numbers suggest that over a larger sample size they will win more than they lose.

Obviously as pointed out by MaxBone the underlying numbers may be a little skewed based on the small sample size.

The underlying numbers suggest that over a larger sample size they will win more than they lose.

Obviously as pointed out by MaxBone the underlying numbers may be a little skewed based on the small sample size.
- Feds91Stammer

Goddam Randy Carlyle