Stats that show Toews is elite defensively

[190Octane]

I posted these numbers in another thread...

Ozsp = Offensive zone start percentage
Ozfp = offensive zone finish percentage
GA/20 = goals against per 20 minutes of 5 on 5

Toews: 63.9 ozsp, 52.6 ozfp, .818 ga/20
Benn: 57.1 ozsp, 52 ozfp, .796 ga/20
Kopitar: 53.7 ozsp, 55.2 ozfp, .428 ga/20
Getzlaf: 48.5 ozsp, 51.5 ozfp, .715 ga/20
Tavares: 62.6 ozsp, 55 ozfp, .905 ga/20
Giroux: 55.6 ozsp, 51.8 ozfp, .779 ga/20
Thornton: 47.5 ozsp, 49.7 ozfp, .720 ga/20
Stamkos: 51 ozsp, 50.9 ozfp, .815 ga/20
Malkin: 59.6 ozsp, 53.5 ozfp, .810 ga/20
Crosby: 51.4 ozsp, 54.1 ozfp, .801 ga/20

I don't understand how someone who is elite defensively ends up having more ga/20 than almost everyone on this list despite starting in the offensive zone 10-15% of the time more than most of the other players on this list.

Is there something that I'm missing or is "Toews is great defensively" purely anecdotal with no hard evidence to back it up?

Click to read more...

 

[DrVanntastic]

Because GA/20 doesn't take into account the other players on the ice?

 

[190Octane]

Because GA/20 doesn't take into account the other players on the ice?

Sharp, Hossa, and Keith are the top 3 players he plays with at 5 on 5, a forward who some say is one of the best defensive wingers in the game and a Norris trophy candidate. I doubt they are defensive liabilities.

Are there any stats that say he's good on defense?

 

[Jacques Trap*]

The eye test trumps stats 98.999% of the time

 

[Bear of Bad News]

The eye test trumps stats 98.999% of the time

That sure as hell isn't true.

I recommend Daniel Kahneman's "Thinking Fast And Slow" as a primer to narrative biases and fallacies.

Stats aren't perfect, but neither is "the eye test".

 

[190Octane]

I haven't seen any stat that says he's elite defensively and everyone that says he's a great player has nothing but anecdotal evidence to support their point.

The only reason he's nominated for the Selke is a reputation thing in my opinion. I'm open to different view points but the only thing people seem to be able to say he passes the eye test. The problem with the eye test is that it's extremely biased and leads you to see what you want to see.

 

[Jacques Trap*]

That sure as hell isn't true.

I recommend Daniel Kahneman's "Thinking Fast And Slow" as a primer to narrative biases and fallacies.

Stats aren't perfect, but neither is "the eye test".

Maybe. But let me ask you. If you forced to rely on only one to evaluate talent, looking solely at numbers but never watching the game or watching but never scrutinizing numbers, which would you choose?

 

[Bear of Bad News]

Maybe. But let me ask you. If you forced to rely on only one to evaluate talent, looking solely at numbers but never watching the game or watching but never scrutinizing numbers, which would you choose?

That's a silly question, since both are available.

Anyone who suggests using one solely over the other probably has a vested interest in the outcome.

 

[Jacques Trap*]

That's a silly question, since both are available.

Anyone who suggests using one solely over the other probably has a vested interest in the outcome.

So it seems to me (and I may be wrong) that you are suggesting that they are both equally accurate in evaluating talent. Do you think that pro scouts rely on one over the other? Or use both equally?

 

[Bear of Bad News]

So it seems to me (and I may be wrong) that you are suggesting that they are both equally accurate in evaluating talent. Do you think that pro scouts rely on one over the other? Or use both equally?

They're generally orthogonal to one another, so it's hard to say whether one is better than another (it's also not clear whether you're talking about scouting vs. statistics as they currently stand, or whether you're talking about the achieveable future level of both items).

I'd say that pro scouts rely primarily on scouting (it's part of the name "scout", after all). That's partly because of the disparity in my parenthetical above.

 

[veganhunter]

I posted these numbers in another thread...

Ozsp = Offensive zone start percentage
Ozfp = offensive zone finish percentage
GA/20 = goals against per 20 minutes of 5 on 5

Toews: 63.9 ozsp, 52.6 ozfp, .818 ga/20
Benn: 57.1 ozsp, 52 ozfp, .796 ga/20
Kopitar: 53.7 ozsp, 55.2 ozfp, .428 ga/20
Getzlaf: 48.5 ozsp, 51.5 ozfp, .715 ga/20
Tavares: 62.6 ozsp, 55 ozfp, .905 ga/20
Giroux: 55.6 ozsp, 51.8 ozfp, .779 ga/20
Thornton: 47.5 ozsp, 49.7 ozfp, .720 ga/20
Stamkos: 51 ozsp, 50.9 ozfp, .815 ga/20
Malkin: 59.6 ozsp, 53.5 ozfp, .810 ga/20
Crosby: 51.4 ozsp, 54.1 ozfp, .801 ga/20

I don't understand how someone who is elite defensively ends up having more ga/20 than almost everyone on this list despite starting in the offensive zone 10-15% of the time more than most of the other players on this list.

Is there something that I'm missing or is "Toews is great defensively" purely anecdotal with no hard evidence to back it up?

Basically every forward on CHI (save for 4 players who receive few minutes) has an OZSP in the 60's...all that says to me is that CHI has a fantastic possession game (not Toews receiving cushy minutes).

Toews, Hossa, Kane (not counting Morin - sample size) have basically the same OZFP (highest on the team for forwards). Not entirely sure of the explanation but there is likely a reason for this pattern.

In terms of the GA thing the only players with a lower ON-ICE-SV (SV% is pretty much only reflective of the goaltender ie. no significant correlation between team and a goaltender's SV%) than Toews are Bickell and Smith combined with the fact he has one of the lowest GA/60 of regular players on the CHI says to me he may have just been a bit unlucky (in fact I might even say it says a lot about Cory Crawford).

I think Corsi much better reflects his defensive/possesion impact (likely the reason Kopitar, Bergeron, and Toews all rank in the top 10 and that's without even adjusting for players with tiny sample sizes eg. 1GP)

Could be wrong though as my understanding of all these metrics is at a fairly novice level.

 

[Jacques Trap*]

They're generally orthogonal to one another, so it's hard to say whether one is better than another (it's also not clear whether you're talking about scouting vs. statistics as they currently stand, or whether you're talking about the achieveable future level of both items).

I'd say that pro scouts rely primarily on scouting (it's part of the name "scout", after all). That's partly because of the disparity in my parenthetical above.

Simplifying things is always a better option than complication.

So quite simply I'm asking when evaluating the talent and skill level of a hockey player, which is more useful and accurate. Watching them in game action, or looking at numbers?

 

[Bear of Bad News]

Simplifying things is always a better option than complication.

So quite simply I'm asking when evaluating the talent and skill level of a hockey player, which is more useful and accurate. Watching them in game action, or looking at numbers?

And I told you - BOTH. None of this made up "or" garbage. You also didn't answer my clarifying question, which would allow me to answer you.

You're free to continue with your line of argument. I won't be responding further, because we've now done a complete circle.

I'll ask that you stop derailing the thread.

 

[Jacques Trap*]

And I told you - BOTH. None of this made up "or" garbage.

You're free to continue with your line of argument. I won't be responding further, because we've now done a complete circle.

Well you can choose to not comment further if you'd like and I'm glad you're letting me me know I'm free to continue. Thanks.

Here's the thing. I've been playing fantasy hockey for a long time so things like corsi and fenwick are not new to me and are very useful in some cases. Especially in something purely stats based like fantasy hockey.

However, in real life hockey there are things that numbers can't equate and very important such as a players intangibles, work ethic, game reading ability (hockey I.Q) and level of professionalism. For those things the "eye test" is relied upon to gauge the true value of a player in real life game situations.

 

[Bear of Bad News]

Here's the thing. I've been playing fantasy hockey for a long time so things like corsi and fenwick are not new to me and are very useful in some cases. Especially in something purely stats based like fantasy hockey.

Who claimed that Corsi and Fenwick are the be-all and end-all of the "statistics" faction?

Yes, the eye test is better than some of the statistics. Is that what you were hoping for?

Can we get back to the question posed by the OP?

 

[190Octane]

Basically every forward on CHI (save for 4 players who receive few minutes) has an OZSP in the 60's...all that says to me is that CHI has a fantastic possession game (not Toews receiving cushy minutes).

Toews, Hossa, Kane (not counting Morin - sample size) have basically the same OZFP (highest on the team for forwards). Not entirely sure of the explanation but there is likely a reason for this pattern.

In terms of the GA thing the only players with a lower ON-ICE-SV (SV% is pretty much only reflective of the goaltender ie. no significant correlation between team and a goaltender's SV%) than Toews are Bickell and Smith combined with the fact he has one of the lowest GA/60 of regular players on the CHI says to me he may have just been a bit unlucky (in fact I might even say it says a lot about Cory Crawford).

I think Corsi much better reflects his defensive/possesion impact (likely the reason Kopitar, Bergeron, and Toews all rank in the top 10 and that's without even adjusting for players with tiny sample sizes eg. 1GP)

Could be wrong though as my understanding of all these metrics is at a fairly novice level.

Are shots attempted a fair measurement when he starts in the offensive zone 63.9% of the time vs. Getzlaf who starts in the offensive zone 48.5% of the time?

Another thing I'd like to point out is that he is starting in the offensive zone 63.9% of the time yet still allowing more goals per 20 minutes than all of those other centers who aren't starting there as often. You would think that an elite defensive center wouldn't be on the ice for so many goals against if he had the advantage of starting 170+ feet away from his own goal 63.9% of the time.

 

[EbonyRaptor]

I posted these numbers in another thread...


Is there something that I'm missing or is "Toews is great defensively" purely anecdotal with no hard evidence to back it up?

Let me start by stating that I don't think Toews was at his best defensively this season, despite the Selke nomination which is as much a "reputatation award" as any in Hockey - and that's based solely on the eye test. I think he was better last season and the couple seasons before that than he was this season.

But, having said that, I take issue with the evidence you presented because it doesn't cover all metrics that would be considered for defensive assessment such as takeaway/giveaway and faceoffs, plus the attack style the Hawks play necessarily gives up more odd man chances against. The fact is that the Hawks didn't play nearly as good defensively as a team this season finishing in the middle of the pack in goals against after winning the Jennings Trophy last season. Toews deserves some of the blame for that but it was a team wide problem this year. He still finished a +26 this year, second only to Hossa's +28.

So while I agree his defensive reputation was based more on past defensive prowess than this season, the metrics you used does not paint the complete picture.

 

[Hammer Time]

I posted these numbers in another thread...

Ozsp = Offensive zone start percentage
Ozfp = offensive zone finish percentage
GA/20 = goals against per 20 minutes of 5 on 5

Toews: 63.9 ozsp, 52.6 ozfp, .818 ga/20
Benn: 57.1 ozsp, 52 ozfp, .796 ga/20
Kopitar: 53.7 ozsp, 55.2 ozfp, .428 ga/20
Getzlaf: 48.5 ozsp, 51.5 ozfp, .715 ga/20
Tavares: 62.6 ozsp, 55 ozfp, .905 ga/20
Giroux: 55.6 ozsp, 51.8 ozfp, .779 ga/20
Thornton: 47.5 ozsp, 49.7 ozfp, .720 ga/20
Stamkos: 51 ozsp, 50.9 ozfp, .815 ga/20
Malkin: 59.6 ozsp, 53.5 ozfp, .810 ga/20
Crosby: 51.4 ozsp, 54.1 ozfp, .801 ga/20

I don't understand how someone who is elite defensively ends up having more ga/20 than almost everyone on this list despite starting in the offensive zone 10-15% of the time more than most of the other players on this list.

Is there something that I'm missing or is "Toews is great defensively" purely anecdotal with no hard evidence to back it up?

First question: Where did you get these numbers from? I'm looking at Behindthenet and I'm seeing Thornton at 0.81 ga/20, Giroux at 0.89 ga/20, Benn at 0.88 ga/20 (all 5v5 numbers).

Second, Toews has a higher Corsi Rel QoC than most of the players on that list showing that he faces tougher opponents.

15% looks like a lot but it's really not. The majority of shifts don't even start with a faceoff, and of those that do, many start with a faceoff in the neutral zone. If Toews is taking 15% more faceoffs in the OZ compared with the DZ, I'm guessing that's about 2-3% more of his shifts played in the offensive zone compared with the average player.

Toews actually had a very good season defensively last year which got him a Selke. This year he's been a little worse but he can still get votes on past reputation.

2013-14: 2.44 GA/60, 26.5 ShA/60
2012-13 playoffs: 1.40 GA/60, 21.7 ShA/60
2012-13: 1.60 GA/60, 24.3 ShA/60
2011-12: 2.84 GA/60, 26.9 ShA/60
2010-11: 2.27 GA/60, 26.7 ShA/60
2009-10: 2.41 GA/60, 22.7 ShA/60

 

[190Octane]

http://stats.hockeyanalysis.com/index.php

Quality of competition is important, which players on that list had better and worse if you don't mind me asking.

Also, it looks like last season he had his best season by far defensively by the numbers you posted. Unfortunately I'm on my phone right now and can't see how the others stack up but it's something I will look at later.

 

[pluppe]

Maybe. But let me ask you. If you forced to rely on only one to evaluate talent, looking solely at numbers but never watching the game or watching but never scrutinizing numbers, which would you choose?

Numbers, easily. Or you might end up with a Rick Nash type of player over a Gretzky type of player. There are a lot of numbers you know.

 

[aleshemsky83]

Numbers, easily. Or you might end up with a Rick Nash type of player over a Gretzky type of player. There are a lot of numbers you know.

Who would pick nash over gretzky having watched both play

 

[Marotte Marauder]

Numbers, easily. Or you might end up with a Rick Nash type of player over a Gretzky type of player. There are a lot of numbers you know.

Who would pick nash over gretzky having watched both play

A blind person?

 

[KaseMeOutside]

zone starts matter, getting offensive zone starts in the 60's makes it easier to prevent goals

 

[Plural]

Numbers, easily. Or you might end up with a Rick Nash type of player over a Gretzky type of player. There are a lot of numbers you know.

Huh? Looking at the numbers, who would anyone take Gretzky-like over Nash-like?

 

[Micklebot]

Sharp, Hossa, and Keith are the top 3 players he plays with at 5 on 5, a forward who some say is one of the best defensive wingers in the game and a Norris trophy candidate. I doubt they are defensive liabilities.

Are there any stats that say he's good on defense?

How about this;

Here is a table of his most frequent line mates over the last 4 years, and the net difference in their Corsi against per 60 when not on the ice with him

Player | CA diff w/o Toews
KEITH, DUNCAN | -1.25
HOSSA, MARIAN | -1.6
SEABROOK, BRENT | -1.36
SHARP, PATRICK | -1.4
KANE, PATRICK | -1.37
HJALMARSSON, NIKLAS | -1.47
LEDDY, NICK | -0.48
SAAD, BRANDON | -2.36
STALBERG, VIKTOR | -2.4
ODUYA, JOHNNY | -4.02
CAMPBELL, BRIAN | -1.18
BROUWER, TROY | -2.6
RAANTA, ANTTI | -0.15
BICKELL, BRYAN | 2.38
BRUNETTE, ANDREW | -1.48
ROZSIVAL, MICHAL | -3.07
KOPECKY, TOMAS | 0.76
FROLIK, MICHAEL | -3.23
BROOKBANK, SHELDON | -2.79
MONTADOR, STEVE | -4.07
VERSTEEG, KRIS | -3.17
BOYNTON, NICK | -2.75
CARCILLO, DAN | -1.55
O_DONNELL, SEAN | -2.21
SHAW, ANDREW | -2.13
SKILLE, JACK | -4.19
CAMPOLI, CHRIS | -1.82
CULLIMORE, JASSEN | 2.9
HENDRY, JORDAN | -3.32
SCOTT, JOHN | 1.97
SMITH, BEN | -1.97
OLSEN, DYLAN | -1.64

All those negatives are players who have over the last 4 years allowed less shot attempts per 60 when on the ice with Toews than when apart. That's quite the feat imo.