Is there such a thing as momentum? (Edit: Between games not shifts)

[Canadiens1958]

Consensus

You're focused on attribution, but what you need to be focused on is "does it matter to begin with?".

If it doesn't matter, then attribution is irrelevant.

Assuming you can get anything approaching consensus about what matters.

Fighting seems to matter in some perceptions of a hockey outcome. It has often been attributed to momentum shifts.

See the link below to a thread from 2009. It is demonstrated that the oft reported and commented on fight between Gordie Howe and Lou Fontinato simply did not matter in terms of short term or seasonal results.

http://hfboards.mandatory.com/showpost.php?p=19473130&postcount=1

Yet some people for the sake of narrative, view it in mythical terms, that it somehow *matters* in NHL history.

My point is that the key is understanding attribution, since the*matters* is too superficial at times and proper attribution just moves these instances of superficial *matters* out of the way.

 

[Ohashi_Jouzu*]

But does a "better" win actually have a tangible effect on the next game? That's really what we're looking at.

We all know the feeling of it exists, but the question is if that feeling actually translates to a tangible impact on the game.

On some level it would come down to whether or not the other team makes adjustments, and whether or not those adjustments provide the desired results, I suppose. The other team ALWAYS has a say in which way the needle ultimately sways, and for how long. But for reasons I mentioned earlier, I think game-to-game (or "lingering") momentum is harder to keep/build than in-game momentum. A team gets hours to adjust/adapt between games, but only gets two short intermissions to do it during a ~2 hr. game, and I think preparation/regrouping time makes a difference - mentally AND physically.

 

[hatterson]

Assuming you can get anything approaching consensus about what matters.

Fighting seems to matter in some perceptions of a hockey outcome. It has often been attributed to momentum shifts.

By "does it matter?" we aren't meaning does it matter to the narrative, we're meaning does it matter statistically. That's not something that can be seen at a quick glance at the standings.

On some level it would come down to whether or not the other team makes adjustments, and whether or not those adjustments provide the desired results, I suppose.

But those things aren't momentum. They may make momentum harder to isolate, but whether momentum has a tangible effect on the next game isn't dependent on system adjustments, line changes, injuries, etc.

 

[Canadiens1958]

Either Way

By "does it matter?" we aren't meaning does it matter to the narrative, we're meaning does it matter statistically. That's not something that can be seen at a quick glance at the standings.



But those things aren't momentum. They may make momentum harder to isolate, but whether momentum has a tangible effect on the next game isn't dependent on system adjustments, line changes, injuries, etc.

Either way. Show how it matters statistically beyond the standings and the results.That there was an actual benefit to anything, anywhere from the Fontinato-Howe fight beyond an easy narrative for the media. Generate a statistical anaysis that contradicts what the standings and results show.

Standings and results are after all statistically objective. Bill Parcell's quote "You are what your record says you are." holds.

 

[oljimmy]

So, some people here have insisted that there is such a thing as team-level momentum but denied that it will show up statistically in a team's actual game-to-game performance. They've been insisting that this kind of momentum is, quote, "qualitative".

Just to repeat something that other posters have been trying to say: if that's what momentum is, why would anyone--GM, coach, player or fan--care about it?

"Well, boys, we've lost 4-0 for three games straight and been behind on every measurable statistic, but at least we've had momentum. Good job!"

 

[Canadiens1958]

Selling the Game

So, some people here have insisted that there is such a thing as team-level momentum but denied that it will show up statistically in a team's actual game-to-game performance. They've been insisting that this kind of momentum is, quote, "qualitative".

Just to repeat something that other posters have been trying to say: if that's what momentum is, why would anyone--GM, coach, player or fan--care about it?

"Well, boys, we've lost 4-0 for three games straight and been behind on every measurable statistic, but at least we've had momentum. Good job!"

All parties have an interest in selling the game, holding the audience, buying time to find a solution, etc. So momentum becomes a synonym for hope,patience, a source for motivation, etc. This should be understood.

As an umbrella word that covers various facets of individual, team or individual improvement, it tends to come up short.

 

[hatterson]

Right, but I don't think anyone has ever doubted whether the narrative of momentum exists. You only have to tune into a tiny amount of playoff hockey to hear someone mention momentum, building on past wins, or trying to put losses behind them.

It's much the same as people have asked about 'clutch' ability. No one doubts the narrative of clutch exists. There are players that hockey media and personnel have declared play better in high pressure scenarios. What some people have questioned, is if this narrative shows up because of actual underlying statistical/empirical data, or if it's simply a result of small sample sizes and confirmation bias.

The same holds true for momentum. If my goal is to understand the game of hockey more, to understand why some players do well and others don't, or why some teams win and others don't, then the narrative doesn't really matter to me. What matters is if something most would agree should be called "momentum" actually changes the weighting of the coin you flip when going into a game. With all else being equal, does one side having "momentum" change their win probability?

 

[Canadiens1958]

Progress

Right, but I don't think anyone has ever doubted whether the narrative of momentum exists. You only have to tune into a tiny amount of playoff hockey to hear someone mention momentum, building on past wins, or trying to put losses behind them.

It's much the same as people have asked about 'clutch' ability. No one doubts the narrative of clutch exists. There are players that hockey media and personnel have declared play better in high pressure scenarios. What some people have questioned, is if this narrative shows up because of actual underlying statistical/empirical data, or if it's simply a result of small sample sizes and confirmation bias.

The same holds true for momentum. If my goal is to understand the game of hockey more, to understand why some players do well and others don't, or why some teams win and others don't, then the narrative doesn't really matter to me. What matters is if something most would agree should be called "momentum" actually changes the weighting of the coin you flip when going into a game. With all else being equal, does one side having "momentum" change their win probability?

Finally some progress.

Yes, momentum is mentioned but "fair coins" are never mentioned in the same circumstances.

Clutch. Again, no mention of "fair coins" in the same circumstances.

Key point is that you have to embrace that sport or any competitive endeavour, like life, is inherently unfair and that the event is about overcoming the unfairness to tip the result ever so slightly in the direction of victory or success for one side.

Measuring the effort to overcome "unfairness" is fairly straightforward. No different than measuring efficiency in the workplace which is never based on "flipping coins" as in a probability abstract.

This is what the short, simple Fontinato-Howe fight study did. Showed that the activity was not efficient for anyone directly involved either short or long term.

 

[hatterson]

Finally some progress.

Yes, momentum is mentioned but "fair coins" are never mentioned in the same circumstances.

Clutch. Again, no mention of "fair coins" in the same circumstances.

"fair coins" was an analogy that Doc No used to help convey his point. I perfectly understood what he was saying, obviously others didn't. Of course talking heads don't mention fair coins or win probability. Their job is the narrative. That's not really what OP is looking for.

Key point is that you have to embrace that sport or any competitive endeavour, like life, is inherently unfair and that the event is about overcoming the unfairness to tip the result ever so slightly in the direction of victory or success for one side.

I'm not really sure how anything I've said contradicts that. Games aren't (or are very rarely) 50/50 odds. The question isn't if the scales are tipped towards one team or not, it's if a team has a 60% chance to win a given game ignoring any 'momentum' effect, does the addition of 'momentum' in their favor change that to a 65% chance or does it leave that chance unchanged?

This is what the short, simple Fontinato-Howe fight study did. Showed that the activity was not efficient for anyone directly involved either short or long term.

No, the short study showed that the points earned percentage for both teams prior to the brawl was higher than the points earned percentage after. There was absolutely no causal relationship established. And honestly, the difference in records isn't that much. Detroit was playing at a 65-66 point pace and finished at 58. The Rangers were playing at a 68-69 point pace and finished at 64.

That's hardly a significant difference, especially given that Boston improved their pace by 10 points (from 63 point pace to finishing with 73)

Were both teams performing better before because they had good puck luck? Were there any injuries involved? How were they performing in the 5-10 games prior to the fight? What about the 5-10 games after (instead of just 1)? Did NY and DET just lose a few games to Boston to make the difference?

 

[Doctor No]

Finally some progress.

You'll see "progress" more frequently if you expand its definition beyond "someone else agreeing with your preconceived notions".

 

[Canadiens1958]

Progress II

"fair coins" was an analogy that Doc No used to help convey his point. I perfectly understood what he was saying, obviously others didn't. Of course talking heads don't mention fair coins or win probability. Their job is the narrative. That's not really what OP is looking for.



I'm not really sure how anything I've said contradicts that. Games aren't (or are very rarely) 50/50 odds. The question isn't if the scales are tipped towards one team or not, it's if a team has a 60% chance to win a given game ignoring any 'momentum' effect, does the addition of 'momentum' in their favor change that to a 65% chance or does it leave that chance unchanged?



No, the short study showed that the points earned percentage for both teams prior to the brawl was higher than the points earned percentage after. There was absolutely no causal relationship established. And honestly, the difference in records isn't that much. Detroit was playing at a 65-66 point pace and finished at 58. The Rangers were playing at a 68-69 point pace and finished at 64.

That's hardly a significant difference, especially given that Boston improved their pace by 10 points (from 63 point pace to finishing with 73)

Were both teams performing better before because they had good puck luck? Were there any injuries involved? How were they performing in the 5-10 games prior to the fight? What about the 5-10 games after (instead of just 1)? Did NY and DET just lose a few games to Boston to make the difference?

Everyone understands perfectly.

Win probability - new phrase introduced that is very vague. Perhaps performance history over a period of games vs the league, vs the opposition, but then we are looking at statistical reports not probability.

Granted that games are not 50/50 odds but how do circumstances effect the game should be considered? If one team is playing with three days rest and the other team is playing a second game in two nights, how does fatigue of the second team influence the results? Look at the compiled data and adjust accordingly since not all teams play the same under such circumstances. Show data to a coach that indicates that going to shorter shifts when playing a second game in two nights against a team with three days rest increases performance by and then have the team deliver on such a claim might get some attention.


15-16% drop in performance isn't that much? Nor is 5-6% drop in performance?
Trust this also applies to other analysis like SV% - goalie or team giving up 150-160 or 50-60 extra goals per 1000 shots during a season,or a stretch.

Fact remains they both lost more often which debunked the myth that the specific fight was a benefit. Who they lost to does not matter, they both still lost more frequently and finished out of the playoffs.

 

[hatterson]

Everyone understands perfectly.

Win probability - new phrase introduced that is very vague. Perhaps performance history over a period of games vs the league, vs the opposition, but then we are looking at statistical reports not probability.

Granted that games are not 50/50 odds but how do circumstances effect the game should be considered? If one team is playing with three days rest and the other team is playing a second game in two nights, how does fatigue of the second team influence the results? Look at the compiled data and adjust accordingly since not all teams play the same under such circumstances. Show data to a coach that indicates that going to shorter shifts when playing a second game in two nights against a team with three days rest increases performance by and then have the team deliver on such a claim might get some attention.

I'm not really sure what you're getting at here. Of course fatigue, coaching adjustments and others affect a team, but those aren't momentum and aren't really relevant to this thread.

15-16% drop in performance isn't that much? Nor is 5-6% drop in performance?

Not in the realm of team point paces over 20 games it isn't. Drops and raises of that size happen to multiple teams every single season in the NHL. Also, the percentage drop doesn't mean that much when the difference is caused by only a few events.

If a team played their first 41 games and went 2-39, they'd be on pace for 8 points. If they finish the season 0-41 they'd have had a 100% performance drop in the second half (or 50% in terms of their finishing pace), but is that really a significant difference? The team was a joke in the first half and remained a joke in the second half.

Yes, both teams played worse in the second half, but the difference in the case of Detroit is 3 wins, and in the case of the Rangers is 2 wins. Especially when both teams had a significantly higher number of ties in the beginning of the season, implying they were in a lot of close games, it's not really convincing that it's meaningful in and of itself.

If you're in a ton of 1 goal or tied-late games, going from 6-15-2 to 8-13-2 is hardly a meaningful change.

Trust this also applies to other analysis like SV% - goalie or team giving up 150-160 or 50-60 extra goals per 1000 shots during a season,or a stretch.

Assuming you're talking about 1000 shots, then yes 15% is significant. If you're talking about 7 shots, then no, it's not really significant since that difference is entirely caused by a single event.

Fact remains they both lost more often which debunked the myth that the specific fight was a benefit. Who they lost to does not matter, they both still lost more frequently and finished out of the playoffs.

I personally don't believe the brawl had much of an effect on either team, but a lower pace to finish the season doesn't debunk anything. Maybe both teams would have finished even lower without the brawl and the brawl helped them from completely falling apart.

 

[Canadiens1958]

Progress III

I'm not really sure what you're getting at here. Of course fatigue, coaching adjustments and others affect a team, but those aren't momentum and aren't really relevant to this thread.




Not in the realm of team point paces over 20 games it isn't. Drops and raises of that size happen to multiple teams every single season in the NHL. Also, the percentage drop doesn't mean that much when the difference is caused by only a few events.

If a team played their first 41 games and went 2-39, they'd be on pace for 8 points. If they finish the season 0-41 they'd have had a 100% performance drop in the second half (or 50% in terms of their finishing pace), but is that really a significant difference? The team was a joke in the first half and remained a joke in the second half.

Yes, both teams played worse in the second half, but the difference in the case of Detroit is 3 wins, and in the case of the Rangers is 2 wins. Especially when both teams had a significantly higher number of ties in the beginning of the season, implying they were in a lot of close games, it's not really convincing that it's meaningful in and of itself.

If you're in a ton of 1 goal or tied-late games, going from 6-15-2 to 8-13-2 is hardly a meaningful change.



Assuming you're talking about 1000 shots, then yes 15% is significant. If you're talking about 7 shots, then no, it's not really significant since that difference is entirely caused by a single event.



I personally don't believe the brawl had much of an effect on either team, but a lower pace to finish the season doesn't debunk anything. Maybe both teams would have finished even lower without the brawl and the brawl helped them from completely falling apart.

Fatigue has been recognized as an influence on the game going back to the first year of the NHL with Elmer Ferguson, columnist and original league statistician commenting and tracking scheduling inequities. Three team league due to a fire so one was always rested.

In the fifties, post 1955 the Canadiens and Leafs went to four lines and quicker changes to combat fatique - sustain momentum, in 3 games in 4 days and 4 games in 5 days stretches of the NHL schedule. The 4 games in 5 days scheduling was banned in one of the early CBAs.

Today the NHL is concerned about the Sitting at Home Waiting Effect, discussed in this thread earlier. In due time it will be addressed.

Youth hockey tracked the fatigue aspect in tournaments generations ago, eventually mandating that a team playing a second tournament game the same day had to have at least four hours rest between games.

Just dealing with what actually happened/happens. Fatigue has and will always be an issue in performance.

 

[Doctor No]

Fatigue has been recognized as an influence on the game going back to the first year of the NHL with Elmer Ferguson, columnist and original league statistician commenting and tracking scheduling inequities. Three team league due to a fire so one was always rested.

Here's the problem - this isn't a catchall thread for "every influence that's ever existed on the game of hockey".

This is a thread for MOMENTUM, and whether not its existence can be quantified (given the nature of the sub-forum).

Your continued attempts to derail the actual conversation are not helping. Would you mind tying your topics back to the thread's purpose?

 

[Ohashi_Jouzu*]

Here's the problem - this isn't a catchall thread for "every influence that's ever existed on the game of hockey".

This is a thread for MOMENTUM, and whether not its existence can be quantified (given the nature of the sub-forum).

Your continued attempts to derail the actual conversation are not helping. Would you mind tying your topics back to the thread's purpose?

I think fatigue and attrition factor into ability to carry momentum from one game to the next, for sure. Teams often pull together "bounce-back" games where they lay it all on the line and get a win against a stronger opponent, for example, but we're not often surprised that they can't string 3 of them together in a row while facing elimination in a 7 game series played within a 2 week time frame. Shortening a bench to get better players out there more often, players getting injured "giving 110%" and blocking shots, etc. can limit a team's ability to "shift the odds" in the longer term (but focus is always on one series at a time in the playoffs), and the rest times that Canadiens mentions impact recuperation needed to sustain/repeat such performances.

If momentum eventually gets related in any way to execution and/or efficiency/effectiveness of play (or confidence, even), then it'll be kind of hard to dismiss the role of collective fatigue and health when it comes to ability to maintain/reproduce a certain level of play, or not.

 

[Doctor No]

Thanks for tying this more directly back to the topic, O_J - and I agree. When people are looking for this sort of thing in the data, I think rest and fatigue will be considerations.

I've often wondered about the impact of back-to-back playoff games (rarer now) on the establishment of momentum.