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

[SnowblindNYR]

Why not? I submit that the psychological side/approach/impact is a HUGE determining factor in why games aren't pre-decided on paper. The problem you're going to encounter is that momentum measured at an individual level is almost unmeasurable without real-time CT scans on the players monitoring brain activity, and on a team level things like chemistry and synergy (also non-quantifiable to statistics, but very real factors) are going to play a huge part in whether or not momentum manifests itself in anything quantifiable to stats. Coaches, nevertheless, preach about things like "building off a good shift/PK/whatever", "getting mojo back", "breaking out of a funk", hot/cold streaks, etc. And that's because, if "harnessed", momentum is something that you can build on - even if "only" psychologically.

Because it's possible that any small difference in mental state in some players in the aggregate is not significant. It's possible that the effects are not large enough or they're large but it doesn't matter since they're not good enough to take advantage of momentum or the other team's negative momentum. Social psychology is a very qualitative field that has a lot of subjectivity to it.

 

[Ohashi_Jouzu*]

Because it's possible that any small difference in mental state in some players in the aggregate is not significant. It's possible that the effects are not large enough or they're large but it doesn't matter since they're not good enough to take advantage of momentum or the other team's negative momentum. Social psychology is a very qualitative field that has a lot of subjectivity to it.

But again, that's not what you're looking for. Presupposing that momentum exists, you shouldn't be interested in the times when the mental state of some players is NOT significant. You should be concerned about the times when it DOES manifest, and subsequently how/why. There are so many different instances to which players/coaches/colour commentators attribute the term that it's hard to know where to begin making a list to search for. But when you have a team that doesn't even average 4 goals/game trailing by 3 goals in the 3rd period, and that team somehow finds a way to string together 4 goals to win the game in the last period, for example, it's kind of hard to categorically deny the potential manifestation/impact of what most people in the game would call "momentum".

Unfortunately for someone who wants to do this over a "meaningfully large" sample, this would involve watching entire games and/or relying on participant "testimony" to narrow down a time frame (or time frames) during the games to study which might yield something noticeable/detectable in terms of before-during-after. There aren't macros or spreadsheets that can help you watch games with a pulse on the emotion or flow of the game, so most statisticians won't touch such a project with a 10 foot pole. They'll simply say "don't even bother, because I can't quantify it for you using the easiest to gather league data and a rudimentary algorithm from a Stats 101 textbook, so it doesn't exist. Next."

 

[Doctor No]

But again, that's not what you're looking for. Presupposing that momentum exists, you shouldn't be interested in the times when the mental state of some players is NOT significant. You should be concerned about the times when it DOES manifest, and subsequently how/why.

The problem with this approach is narrative bias. In a world where momentum doesn't exist - in fact, supposing that all of the games were perfectly random - then there will still be patterns exhibited in the random results. The ones that "looked like momentum" would then be justified as such by players, media, and the like.

XKCD did a strip on this awhile back that I like.

It's a bit more cynical than I am (they're momentum atheists, whereas I'm momentum agnostic).

Put differently, if you flip twenty unweighted coins 1,000,000 times apiece, you'd expect one time where all 20 coins simultaneously turn up heads. Fans of those coins might attribute unnatural abilities to the coins.

So where does that leave us? If momentum exists (and is meaningful), then it has to manifest itself in results at a level beyond what would be seen from random data.

 

[Ohashi_Jouzu*]

The problem with this approach is narrative bias...

I get that. I do. But we're talking about something that's observable. We're not just relying on the participants' "interpretation" of what they feel. Passes start hitting the tape more. Players start getting to pucks earlier. A whole slew of execution-related aspects that result in the ice appearing "tilted" that commonly available stats fail to encapsulate.

For example can you find me the passing success % for a team's first 10 minutes of a period vs. the last 10? Can you tell me how many more pucks players got to first in those initial 10 minutes vs. the last 20? What if it was actually just a 5 minute window of ice tilting that "got the team back into the game", and what if it didn't produce any goals but simply contributed to wearing the other team down by the end of the game? etc. No stats for this available anywhere. And even if you came up with numbers, what would these numbers mean/produce vs. another team, on another night, at a different point in the schedule, with different members on the roster?

Obviously I haven't provided an exhaustive list of potential underlying currents that could lead the boat towards an understanding of momentum. My point is more that there are plenty of excuses for dismissing the notion of momentum, but most of them stem from inability to completely describe a hockey game's flow through numbers. Still an observable phenomenon, though, whatever the "momentum atheists" (cool term) feel more comfortable calling it.

 

[Doctor No]

I agree with your points. To the extent that momentum exists and plays a role in games, the abundance of data that we're seeing now (and will continue to see more of, with NBA-style tracking), it will appear in the data. God knows that plenty of people are looking for it.

 

[SnowblindNYR]

The problem with this approach is narrative bias. In a world where momentum doesn't exist - in fact, supposing that all of the games were perfectly random - then there will still be patterns exhibited in the random results. The ones that "looked like momentum" would then be justified as such by players, media, and the like.

XKCD did a strip on this awhile back that I like.

It's a bit more cynical than I am (they're momentum atheists, whereas I'm momentum agnostic).

Put differently, if you flip twenty unweighted coins 1,000,000 times apiece, you'd expect one time where all 20 coins simultaneously turn up heads. Fans of those coins might attribute unnatural abilities to the coins.

So where does that leave us? If momentum exists (and is meaningful), then it has to manifest itself in results at a level beyond what would be seen from random data.

Just did the math there's a 61% that it happens at least once.

 

[Ohashi_Jouzu*]

I agree with your points. To the extent that momentum exists and plays a role in games, the abundance of data that we're seeing now (and will continue to see more of, with NBA-style tracking), it will appear in the data. God knows that plenty of people are looking for it.

Well, until someone can quantify confidence (which I think is a huge component of performance, "contagious" to different extents within different groups, and a contributing factor in maintaining momentum as a group long enough to turn it into something "statistically relevant"), I don't think anyone is going to "find" it.

 

[Random Forest]

Well, until someone can quantify confidence (which I think is a huge component of performance, "contagious" to different extents within different groups, and a contributing factor in maintaining momentum as a group long enough to turn it into something "statistically relevant"), I don't think anyone is going to "find" it.

Why not? If confidence is a factor, and I would agree that it is, the data will bear it out presuming it captures the types of things you mentioned above.

 

[Dr Quincy]

If I said you're amazing, sexy and smart vs you're a piece of trash, how can a loser like you exist... Which situation would you play better in?

The answer is: If you called me a piece of trash and I played well you'd say it was because I wanted to prove you wrong. If you called me amazing and I played well you'd say because you gave me confidence. If I played poorly when you called me a piece of trash you'd say your comments shattered my confidence. If I played poorly when you said I was amazing you'd say that you didn't make me want to prove you wrong and I got complacent.

That's the problem with this kind of thing.

 

[SnowblindNYR]

A regression I ran that I'm sure has been done a million times before but is interesting nonetheless.

Is the old adage that defense wins championships true?

The answer is both are important but it does seem that a low GA again correlates better with playoff wins.

Variable Coefficient Std. Error t-Statistic Prob.​

GF_AVERAGE 4.732958 1.447029 3.270811 0.0013
GA_AVERAGE -6.520304 1.730167 -3.768599 0.0002
C 8.471477 5.092776 1.66343 0.0982

 

[Random Forest]

The answer is: If you called me a piece of trash and I played well you'd say it was because I wanted to prove you wrong. If you called me amazing and I played well you'd say because you gave me confidence. If I played poorly when you called me a piece of trash you'd say your comments shattered my confidence. If I played poorly when you said I was amazing you'd say that you didn't make me want to prove you wrong and I got complacent.

That's the problem with this kind of thing.

Good post. Narrative bias is everywhere in sports.

For example, after the Christmas break, I heard multiple announcers suggest something to the effect of, "They came out flat after having almost a week off". Problem is, I also heard other announcers suggest, "They came out with energy; they were flying after being well rested from the break!"

There's always going to be a narrative to make sense of an observation.

 

[Hardyvan123]

I agree with your points. To the extent that momentum exists and plays a role in games, the abundance of data that we're seeing now (and will continue to see more of, with NBA-style tracking), it will appear in the data. God knows that plenty of people are looking for it.

I still think that there will be missing data and information that is unknown that can and will have an affect on teams and heck even random luck is more likely to have an affect in the NHL these days since skill isn't as necessary to win games as is systems, goal tending and blacked shots ect...

team flights and when and how long they fly may or may not have an impact as well, even flu bugs running through a team, something we can't track, can have a very profound impact.

 

[hatterson]

skill isn't as necessary to win games as is systems, goal tending and blacked shots ect...

I'm curious as to why you don't count those things as skill.

Blocking shots is more than just flopping on the ice and hoping a puck hits you. A system is more than just a simple "go here" and involves both skill from the players to understand it as well as well as coaching skill to devise and communicate the system. And I'm pretty sure Doc No will come in here and punch me if I try and say goaltending isn't a skill.

No, it may not be dekes and flashy stick work, but it's certainly still skill.

 

[Doctor No]

And I'm pretty sure Doc No will come in here and punch me if I try and say goaltending isn't a skill.

I've been off the ice since December 9th (and have a game Friday) so if it's a skill, then I'm hoping it has a long memory.

 

[1990*]

I am way, way out of my depth on this sort of thing, but would not a reasonable avenue of investigation be whether positive plays by a player engender more of them in any given game, stretch of games, etc.? For example, does a player who scores a goal in Period 1 have a better chance of bagging one in successive periods that night versus any other period on any other night? Does a player who completes a successful pass make more of them over the next ten such attempts, versus one that is picked off/missed?

It's not quite "momentum", per say, but we start to approach a holistic picture once the behavioral component is introduced; if it really is as contagious as we're led to believe by the psych research on the matter, we should be able to find that positive patterns of teammates scoring/making passes/making hits/(insert action here) spread to the entire line and perhaps team.

The beauty of this is that the opposite should also hold true; that is, if momentum "exists", so should contagious cold streaks. Does Shane Doan having multiple shot blocks against/posts/etc. spread to his linemates, or is he on an island? Does a quick goal against to open a game send a team back on their heels by all metrics we have available to us?

I don't know, am I off-base here?

 

[Plante]

pretty much this and this forum probably isn't the place for the question because hockey, like all sports, has an emotional base that isn't going to show up in large grouping of statistics that looks for trends or predictive behavior .

It's an eye test thing really.

Also momentum is more of an in game thing than a carry over one.

Each game does stand on it's own more or less but in game emotions and momentum can sweep teams (and players) up at times.

I second this sentiment.

 

[Doctor No]

I second this sentiment.

Did you factor narrative bias into your sentiment?

 

[SnowblindNYR]

I am way, way out of my depth on this sort of thing, but would not a reasonable avenue of investigation be whether positive plays by a player engender more of them in any given game, stretch of games, etc.? For example, does a player who scores a goal in Period 1 have a better chance of bagging one in successive periods that night versus any other period on any other night? Does a player who completes a successful pass make more of them over the next ten such attempts, versus one that is picked off/missed?

It's not quite "momentum", per say, but we start to approach a holistic picture once the behavioral component is introduced; if it really is as contagious as we're led to believe by the psych research on the matter, we should be able to find that positive patterns of teammates scoring/making passes/making hits/(insert action here) spread to the entire line and perhaps team.

The beauty of this is that the opposite should also hold true; that is, if momentum "exists", so should contagious cold streaks. Does Shane Doan having multiple shot blocks against/posts/etc. spread to his linemates, or is he on an island? Does a quick goal against to open a game send a team back on their heels by all metrics we have available to us?

I don't know, am I off-base here?

You're not off base and if I was a professional NHL stats guy I'd probably look at that but some of that would be a lot of work and some of that I don't even know if it's possible to do (the passes one). As is, I compiled game results and even that took a while.

 

[Plante]

Did you factor narrative bias into your sentiment?

No


Looking back though I still agree with it. Momentum isn't game to game or period to period. It's an emotional aspect that can last for shifts at a time. Humans are emotional beings and certain events can rev us up or bring us down. Hard to discuss in words, but on the ice you can feel it.

 

[SnowblindNYR]

This seems to be a relatively controversial topic that seems to elicit strong emotional responses. I didn't mean it to be. Is the reason it's controversial because it's somewhat related to another controversial topic that a elicits strong emotional response, fighting? There is so much heated debate whenever fighting is questioned and one supposed benefit of fighting is momentum.

 

[Doctor No]

Looking back though I still agree with it. Momentum isn't game to game or period to period. It's an emotional aspect that can last for shifts at a time. Humans are emotional beings and certain events can rev us up or bring us down. Hard to discuss in words, but on the ice you can feel it.

I think that's part of the problem - as a human, and as a hockey player, I agree with you 100%. I believe that momentum is a real thing and effects games.

The problem is that I'm subject to narrative bias, too (whether I like it or not). When I believe a game was influenced by momentum, it's always after the fact, and I never sit there after games thinking about "wow, that one play really didn't change the momentum at all."

Similarly, there are a lot of times during games where I think that something's going to change the momentum ("wow - that was a big save! I bet we turn things around now."), but don't keep track of the percentage of times where I'm right.

 

[Stand Witness]

I didn't read through the whole thread but I saw a few mentioning that 'momentum is something that is purely eye test because it is emotional based and you can't quantify that'. I would have to disagree and I think there would be a way to quantify momentum but it might not be as traditional as some would want it. In fact I believe the most basic form of momentum is already being tracked.

Simply looking at the possession and increased #'s a team receives after a goal is scored should quantify momentum. As momentum increases your possession and the number of shots taken should as well.

Same as a players ability to be 'clutch'. Some people who don't like numbers might not like that you could actually determine it but if you look at the performance of players while they are behind in the game you should figure out which players are by definition 'clutch' and it will ultimately tell you the 'most clutch' players too.

 

[SnowblindNYR]

BTW, I should have mentioned this a while back. I meant momentum between games and not between shifts, maybe I should change the name of the thread.

 

[Hawksfan2828]

Oh absolutely there is momentum between games - moreso in the playoffs...

Momentum is merely a change in confidence or as far as a team - morale.

I mean if momentum didn't exist then teams wouldn't be able to come back when they were down in a series 3-1 or even 3-0...

 

[Doctor No]

Oh absolutely there is momentum between games - moreso in the playoffs...

Momentum is merely a change in confidence or as far as a team - morale.

I mean if momentum didn't exist then teams wouldn't be able to come back when they were down in a series 3-1 or even 3-0...

Short answer: you're hand-waving. Saying that something exists because it "has to exist" isn't much of a proof.

Longer answer: wouldn't teams with a 3-0 have a lot of momentum? So how does the fact that (some) teams have come back from a 3-0 deficit prove the existence of momentum?

And what does the existence of teams that get swept (after being down 3-0) prove? Momentum, or no momentum? Couldn't you say that "Teams that are up 3-0 sometimes lose a series, and sometimes sweep. This proves that momentum either exists or doesn't exist."

Longer answer: if you're flipping fair coins, and have flipped heads three times in a row, there's 1-in-16 chance that you'll then flip four tails in a row. Is that proof of momentum's existence?