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

[Ohashi_Jouzu*]

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.

Well that's a bit tougher. Once you've had a chance to sleep and get a practise in before the next game, it's kind of hard to draw on the same emotion that you may have channeled in the game previous. And pros are surprisingly good at leaving past games behind them and focusing on the next one. But I'm sure a close examination of enough playoff series' would reveal cases of a team settling into their roles and discovering match-ups/mismatches that work, and being able to translate that into strategy/execution that generates the ability to preserve some game to game momentum. Veterans and coaches play a big role in keeping a good thing rolling, too. Much harder to discern through observation than momentum within a single game, though, and probably even harder to reveal anything concrete statistically given what we have to work with.

 

[Canadiens1958]

Difference

Well that's a bit tougher. Once you've had a chance to sleep and get a practise in before the next game, it's kind of hard to draw on the same emotion that you may have channeled in the game previous. And pros are surprisingly good at leaving past games behind them and focusing on the next one. But I'm sure a close examination of enough playoff series' would reveal cases of a team settling into their roles and discovering match-ups/mismatches that work, and being able to translate that into strategy/execution that generates the ability to preserve some game to game momentum. Veterans and coaches play a big role in keeping a good thing rolling, too. Much harder to discern through observation than momentum within a single game, though, and probably even harder to reveal anything concrete statistically given what we have to work with.

But this is no different than leaving past periods, shifts, bad moves, etc behind within the game and focusing on the precise upcoming situations going forward.

As for coaching, observations or anything that falls under the umbrella of "Changes", to a degree this may be measured. Looking at optimization of variables such as shift length, match-ups, how a team defends, attacks, etc will impact performance in the next game.

 

[Ohashi_Jouzu*]

But this is no different than leaving past periods, shifts, bad moves, etc behind within the game and focusing on the precise upcoming situations going forward.

In a simulator with very simply programming maybe, but in reality/practise it's very different. When it comes right down to it, both the chemical and "human nature" sides of our emotions, confidence, and behaviour, and how we channel or react to them, is very different over a period of a couple of minutes/hours vs a period of time interrupted by sleep AND/OR other robust physical activity AND/OR preparations that may or may not involve a different approach than the previous game.

 

[BenchBrawl]

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?

I didn't read the thread, but this comment caught my eyes.It's not sure that a team leading a series 3-0 has momentum in-between the Game 3 and Game 4.Anxiety to choke could come into play and cut their legs.

That's not to mention there's in-game momentum and an underlying, more general momentum that can last days or weeks.

 

[Doctor No]

I didn't read the thread, but this comment caught my eyes.It's not sure that a team leading a series 3-0 has momentum in-between the Game 3 and Game 4.Anxiety to choke could come into play and cut their legs.

Well, sure, but are you confident enough to say that the team *down* 3-0 has momentum going into Game 4? That seems to be the claim being made.

 

[Poogs]

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...

what???

 

[Marotte Marauder]

There is definitely momentum (confidence or feeling good about your game). The issue is that the paper thin line defining which team has momentum is always being challenged.

Momentum may last for a shift before it's wrested away by the other team. Sometimes momentum will go from one shift to another and last a game or a series. Thus a team that's down 3-0 in a series can have momentum, how long it lasts will be seen as the game and series progresses.

 

[Poogs]

its sort of a dumb premise IMO. IF momentum exists (a big if), does a team really capitalize on it enough to have a tangible effect? Its so hard to quantify, it may not even exist, and if it does exist I cant imagine its effect being significant that I think you can spend your time a lot more efficiently

 

[SnowblindNYR]

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?

Even though I'm doubtful about momentum existing, I don't know if it's entirely fair to compare series comebacks to random flips of coins.

 

[Doctor No]

Even though I'm doubtful about momentum existing, I don't know if it's entirely fair to compare series comebacks to random flips of coins.

Allow me to explain.

(Some) people are claiming that series comebacks are evidence of momentum.

We *know* that fair coins do not contain momentum-based properties.

Therefore, in order to show that series comebacks are evidence of momentum, they must have some property that is distinguishable from fair coins. Game sequences have to exhibit patterns different from what we can expect randomly, *and* those patterns have to be attributable to momentum.

So far, this has not been demonstrated.

Clearer?

 

[Canadiens1958]

Fair Coins

Allow me to explain.

(Some) people are claiming that series comebacks are evidence of momentum.

We *know* that fair coins do not contain momentum-based properties.

Therefore, in order to show that series comebacks are evidence of momentum, they must have some property that is distinguishable from fair coins.

So far, this has not been demonstrated.

Clearer?

So momentum change could be as simple as replacing a poorly performing player with another? Replacing a fair coin on an undesirable streak with another fair coin that may generate a desirable streak? Doubtful.

Problem with the fair coins analogy is that there are no opposing coins that are just as fair, trying to dictate a different outcome at the same time.

Back to hockey. A seven game series is not about probabilities(fair coins) it is about statistics(mining data) and making appropriate analysis and changes. How are the analysis and changes reflected in the result? Is a balanced result W-L-W-L-W-L-W significantly different from a L-L-L-W-W-W-W string?

Second string may produce a more attractive narrative but it may also mask strategic or player selection weaknesses by both teams. Momentum(assuming the scenario reflects momentum) would cancel or balance in each instance.

So far no definition or example(s) of momentum have passed the litmus test.

 

[Doctor No]

Well - you said a bunch of things that are inclusive of my point, but then you agreed with me at the end, so I'll take it.

 

[SnowblindNYR]

Allow me to explain.

(Some) people are claiming that series comebacks are evidence of momentum.

We *know* that fair coins do not contain momentum-based properties.

Therefore, in order to show that series comebacks are evidence of momentum, they must have some property that is distinguishable from fair coins. Game sequences have to exhibit patterns different from what we can expect randomly, *and* those patterns have to be attributable to momentum.

So far, this has not been demonstrated.

Clearer?

I mean, it's only indistinguishable from fair coin flips because it's possible to flip H 4 times in a row. Not sure what that proves, it's technically possible to flip H a million times in a row.

 

[Doctor No]

I mean, it's only indistinguishable from fair coin flips because it's possible to flip H 4 times in a row. Not sure what that proves, it's technically possible to flip H a million times in a row.

Well, sure. How often do teams come back to win a series when down 3-0? Is it meaningfully different from 1 in 16 (the odds of flipping heads four times in a row)?

The fact that you used "technically possible" for one situation but not the other tells me that you see one as reasonably possible but not the other.

When you look at situations where a team is down 3-0, do you view coming back to win as reasonably possible or not? We have ways of quantifying these situations.

 

[hatterson]

I think it's worth noting in the 3-0 scenario that teams aren't always equal. Yes, sometimes better teams get down 3-0, but more often the team that is down 3-0 is not as good as their opponent. They ended up down 3-0 not just because they were unlucky, but also because they simply weren't as good.

Even if you reduce their odds of winning to 45% from 50% you are down to about 1/25 as opposed to 1/16.

Therefore, in order to show that series comebacks are evidence of momentum, they must have some property that is distinguishable from fair coins.

In this case I'd argue they don't need to show difference from fair coins, but rather unfair coins. Granted defining just how unfair those coins are is a non-trivial task.

 

[Doctor No]

Agreed - you'd want "coins" reflective of each team's baseline ability.

I'm having enough trouble explaining this with fair coins, however.

 

[Canadiens1958]

Coins

^^^ Issue is not fair or fair coins. Let me illustrate.

A football quarterback may start a game throwing three incompletions followed by four completions. Yet this may not impact the result of the game. Each pass may be viewed as non impacting sub events.

On the other hand in a playoff series, a team losing three games then winning four games and the series, is viewed as significant.

Point is that that each game was decided by consecutive or non consecutive sub events that happened to produce a statistically interesting result.

 

[Doctor No]

Okay. Now show that it *matters* (and isn't just stochastic).

 

[Canadiens1958]

Attributable

Okay. Now show that it *matters* (and isn't just stochastic).

Issue is what is attributable and how it is attributed.

Coin flip in an NFL playoff over the weekend was repeated because on the first flip the coin did not rotate. Objective and filmed standard that the flip was not fair with attributed cause of "unfairness".

Incompleted pass. Defensive and offensive sides attribute cause, importance and consequences differently. Likewise completed pass.

 

[Doctor No]

Okay. Now show that it *matters* (and isn't just stochastic).

(I guess I need to say it again).

 

[Ohashi_Jouzu*]

Okay. Now show that it *matters* (and isn't just stochastic).

(I guess I need to say it again).

Well, I suppose it matters because equating W/L to "fair" coin flips yada yadas over the process behind the W/L, and how that might impact teams moving forward. You can't flip a coin and get heads, then flip it again and get a BETTER heads. You can, on the other hand, win convincingly enough to make players on your team feel more confident, make the players on the other team feel less confident, create doubt in the other team's goalie resulting in a change in starter for the next game, etc, etc. In other words, the "quality" of the result can invoke (potentially momentum-affecting) changes in approach for either team involved - not just the result in black/white terms.

So I'm not sure if this is what Canadiens was talking about when it comes to "attributable events", but the events that contribute to the result (or "consecutive or non consecutive sub events that happened to produce a statistically interesting result") and their "quality" and/or "magnitude" matter, imo, and the coin flip analogy completely fails to account for that from what I can tell.

 

[hatterson]

Well, I suppose it matters because equating W/L to "fair" coin flips yada yadas over the process behind the W/L, and how that might impact teams moving forward. You can't flip a coin and get heads, then flip it again and get a BETTER heads. You can, on the other hand, win convincingly enough to make players on your team feel more confident, make the players on the other team feel less confident, create doubt in the other team's goalie resulting in a change in starter for the next game, etc, etc. In other words, the "quality" of the result can invoke (potentially momentum-affecting) changes in approach for either team involved - not just the result in black/white terms.

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.

 

[Doctor No]

I responded in the Football thread - sigh.

So I think that the question becomes "does momentum change the probability of the next flip?"

(Which is basically what hatterson has now said, better than I did ).

 

[Canadiens1958]

Coin Flip

Well, I suppose it matters because equating W/L to "fair" coin flips yada yadas over the process behind the W/L, and how that might impact teams moving forward. You can't flip a coin and get heads, then flip it again and get a BETTER heads. You can, on the other hand, win convincingly enough to make players on your team feel more confident, make the players on the other team feel less confident, create doubt in the other team's goalie resulting in a change in starter for the next game, etc, etc. In other words, the "quality" of the result can invoke (potentially momentum-affecting) changes in approach for either team involved - not just the result in black/white terms.

So I'm not sure if this is what Canadiens was talking about when it comes to "attributable events", but the events that contribute to the result (or "consecutive or non consecutive sub events that happened to produce a statistically interesting result") and their "quality" and/or "magnitude" matter, imo, and the coin flip analogy completely fails to account for that from what I can tell.

Fair coin flip stands on its own attributable standards of fairness. Coin did not rotate. Not disputable.Flip was repeated. Coin rotated and the result stood.

Incomplete pass. Fair football(as defined by NFL inflation standards). Sequence of three incomplete followed by four complete passes by the same quarterback, using a fair football.

Attribution.It is just as possible that the quarterback adjusted successfully to failure as the defensive team misadjusted to success.

Likewise in hockey even though the distinction between offence and defence is not as precise as in football. Did the successful team adjust and gain impetus towards victory or did the unsuccessful team "clutch defeat from the jaws of victory" and give away victory?

See the Cincinnati defeat vs Pittsburgh in the NFL playoffs. No momentum shift. Demonstratably Cinncinati "clutched defeat from the jaws of victory" three times before the winning FG was kicked. All Pittsburgh had to do was avoid losing by kicking a successful FG.

Hockey with a seven game series simply requires more events before a series result is determined.

 

[Doctor No]

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.