Bobby Orr's icetime during his prime

[Canadiens1958]

Agree But...............

You have to be careful with this, though. During a game, I expect the last thing on a player's mind is trying to keep track of exactly how long he's been on the ice that game. If you ask him to later estimate it, you have no idea how accurate that estimate will be, if he's relying solely on his memory, especially since it's a figure that he wasn't actually keeping track of at the time.

General agreement here but the issue is clouded by the fact that certain athletes at the higher level tend to have an inner clock.Certain runners have the ability to accurately judge split time, certain boxers have the ability to judge where they are in a three minute round and manage the fight accordingly, while certain FB quarterbacks can manage time mentally while others cannot even with the help of multiple stadium clocks.

On the other handhockey players express themselves about other factors that may or may not be accurate and an astute observer realizes this.Forwards provide SOG estimates yet it is very doubtful if during the game they visually / mentally track their shot to see if it would qualify as a SOG. Like wise comments made by players about other stats and events need to be evaluated within the same context.

Your TOI model. Is it adaptable to extrapolate other data like SOG? Comments relative skills of opponents, etc?

 

[Iain Fyffe]

Your TOI model. Is it adaptable to extrapolate other data like SOG? Comments relative skills of opponents, etc?

I would say no, because the variance in individual SOG stats is far, far greater than in individual TGF and TGA stats. SOG is very dependent on an individual player's style and choices, while TGF and TGA are somewhat dependent on those but also greatly on the four teammates and five opponents on the ice at the same time.

 

[seventieslord]

I would say no, because the variance in individual SOG stats is far, far greater than in individual TGF and TGA stats. SOG is very dependent on an individual player's style and choices, while TGF and TGA are somewhat dependent on those but also greatly on the four teammates and five opponents on the ice at the same time.

In other words, a player's GF/GA say a ton more about their icetime than their SOG totals.

a correlation function done in excel would prove this quite easily.

 

[Iain Fyffe]

In other words, a player's GF/GA say a ton more about their icetime than their SOG totals.

Indeed.

 

[Canadiens1958]

Understood

I would say no, because the variance in individual SOG stats is far, far greater than in individual TGF and TGA stats. SOG is very dependent on an individual player's style and choices, while TGF and TGA are somewhat dependent on those but also greatly on the four teammates and five opponents on the ice at the same time.

Understood with an amendment that the number of teammates and opponents is not fixed rather fluid. Teammates may range from 3 to 5 while opponents may range from 4-6. I would take this analogy to apply to most tangible from memory assertions.

What about various so-called intangible factors? Keeping it simple let's focus on one - fatigue.

 

[Canadiens1958]

Not the Issue

In other words, a player's GF/GA say a ton more about their icetime than their SOG totals.

a correlation function done in excel would prove this quite easily.

But that was not the question or the issue, The question was could SOG totals be extrapolated from GF/GA in a conjunction with a TOI model.

Iain provided the answer.

 

[Iain Fyffe]

Understood with an amendment that the number of teammates and opponents is not fixed rather fluid. Teammates may range from 3 to 5 while opponents may range from 4-6.

Obviously; the specific numbers don't matter so long as they are greater than one.

I would take this analogy to apply to most tangible from memory assertions.

What about various so-called intangible factors? Keeping it simple let's focus on one - fatigue.

No idea what you're asking about here.

 

[Canadiens1958]

Fatigue.

No idea what you're asking about here.

Fatigue is often attributed as a reason for diminished performance but there are no tangible numbers to support drops in performance due to fatigue.

May your TOI results model and results be fine tuned to study fatigue?

Examples of fine tuning - looking at the variances in average time per shift.. Looking at the results in games with a day off on each side vs games on two consecutive days? Results re PK time. PK time has long been held to be the most fatiguing aspect of hockey since 3 or 4 players are trying to do the work of 5. Can your efforts be adapted to study such questions?

 

[Iain Fyffe]

Examples of fine tuning - looking at the variances in average time per shift.. Looking at the results in games with a day off on each side vs games on two consecutive days? Results re PK time. PK time has long been held to be the most fatiguing aspect of hockey since 3 or 4 players are trying to do the work of 5. Can your efforts be adapted to study such questions?

The TOI model is designed to do one thing and one thing only. I don't see any reason why the questions you pose above about fatigue could not be studied, but the TOI model would not be the tool to do it.

I did a very simple preliminary team-based study years ago based on the numbers of days rest a team had. You can find it here. I've never done anything on individual players.

 

[Canadiens1958]

Funny

The TOI model is designed to do one thing and one thing only. I don't see any reason why the questions you pose above about fatigue could not be studied, but the TOI model would not be the tool to do it.

I did a very simple preliminary team-based study years ago based on the numbers of days rest a team had. You can find it here. I've never done anything on individual players.

Please do not take the following the wrong way.

Your TOI study may have been designed for a singular purpose but I sense that it has more potential then you are willing to explore.

The link to the fatigue study you did is appreciated. While it is interesting it is an island.

Something I learned years before you were born. There are reasons and causes that are so compelling that everyone wishes them to be the reason. WWII is such a situation. Used to believe and play with numbers but eventually saw the WWII reason as simply a non-factor.

 

[Iain Fyffe]

Your TOI study may have been designed for a singular purpose but I sense that it has more potential then you are willing to explore.

I'd certainly be willing to explore any possible uses of such a system. Such possibilities are not readily apparent, since it is such a singular purpose. I don't think it can do what you're suggesting, though, since it's calibrated only to deal with a season's worth of aggregate numbers.

Something I learned years before you were born. There are reasons and causes that are so compelling that everyone wishes them to be the reason. WWII is such a situation. Used to believe and play with numbers but eventually saw the WWII reason as simply a non-factor.

That the reason for the massive player losses is specifically WWII is irrelevant to me; it's descriptive and not prescriptive. That there were such massive player losses is what matters. It could have been the bubonic plague for all I care.

Can you provide the numbers which demonstrate that the wartime player losses are a non-factor? I provided some numbers earlier that suggest they aren't.

Edit: Sorry, not in this thread obviously. The other one, where WWII is actually a topic of the conversation. Can we keep the various threads on topic?