Leafs chances for making the playoffs this year...

[bustaheims]

Actually, from what I've read about sports, momentum is often nothing.  The best known study, I believe, is from basketball.  They showed that in basketball, shooters don't actually get "hot" --- having made your last shot does not make it more likely you will make your next one.

Momentum is definitely a hard thing to pin down. I think there's some legitimate basis to it when comes to hockey in terms of puck control and driving the play, but, it's something that's really contained to individual games. People like to try to attach it to longer stretches, but, at that point, it really becomes way too nebulous a concept to hold value.

 

[bustaheims]

There's no way that site would have seen the Leafs going 7-2-1 in their last 10.

Well, for one, they're not predicting individual segments of the season. They predict the remainder of the season as a whole, and calculating the odds of different scenarios playing out. Secondly, I'm sure, in their 15+ billion permutations, there were a few million that saw the Leafs go 7-2-1 in their last 10.

 

[Corn Flake]

Actually, from what I've read about sports, momentum is often nothing.  The best known study, I believe, is from basketball.  They showed that in basketball, shooters don't actually get "hot" --- having made your last shot does not make it more likely you will make your next one.

Momentum is definitely a hard thing to pin down. I think there's some legitimate basis to it when comes to hockey in terms of puck control and driving the play, but, it's something that's really contained to individual games. People like to try to attach it to longer stretches, but, at that point, it really becomes way too nebulous a concept to hold value.

I look at it more as confidence being high/low rather than momentum.  It can happen in game and it can stretch over multiple games.  Teams that are winning are usually confident. And when they are they do their job with more ease than they would if they weren't.  When teams are down, players are hesitant, "squeeze their sticks too tight", have to think rather than just react, etc.  Teams that get over confident start to get lazy and then start losing, as we've seen this Leafs team do many many times over the last... forever.

 

[princedpw]

The algorithm cannot determine momentum.  It doesnt know if you've won or lost the last five games.  Momentum is everything

Actually, from what I've read about sports, momentum is often nothing.  The best known study, I believe, is from basketball.  They showed that in basketball, shooters don't actually get "hot" --- having made your last shot does not make it more likely you will make your next one.

While that's likely true in hockey too (getting a goal doesn't mean your next shot has a higher shot%) it has almost nothing to do with what I would call momentum in hockey. When a team keeps possession in the zone and starts to get more shots on net and the opposition starts to collapse in, it seems to carry to the next shift and the next after that for a while. If no goal comes after a while then it seems to settle a bit. But I would still call that momentum.

I see.  I suppose I wouldn't have called that "momentum" -- I would have called that having a consistent territorial advantage and/or consistent puck possession.  Certainly, having a consistent territorial advantage and puck possession does usually lead to scoring more goals.  Keeping a team boxed in so all they can really do is dump the puck out without possession and then change is a great way to win a game.

 

[Zee]

There's no way that site would have seen the Leafs going 7-2-1 in their last 10.

Well, for one, they're not predicting individual segments of the season. They predict the remainder of the season as a whole, and calculating the odds of different scenarios playing out. Secondly, I'm sure, in their 15+ billion permutations, there were a few million that saw the Leafs go 7-2-1 in their last 10.

Sure but their percentage of Leafs making the playoffs means nothing.  It's 70% chance based on all their simulations that the Leafs will have a good enough record in their final 22 games to make the playoffs.  Their method of determining wins/losses is largely random chance.  It would be interesting to take a screenshot of all their predictions for each team in the league right now and see which ones actually make the playoffs and which do not.

Ok here we go:

 

[Heroic Shrimp]

There's no way that site would have seen the Leafs going 7-2-1 in their last 10.

From that statement alone it's clear that you don't really seem to understand what probabilities truly are.  I'm sorry if it sounds condescending, but it's true.  Given that, I think it goes without saying that considerations of the number crunching of raw objective data that leads to the calculation of such probabilities will be even further beyond your understanding.

I have a decent background in probabilities and the calculation thereof, and I have no problems with the means they use to calculate their probabilities.  You must understand that although it is it not a perfect model, there is no perfect model.  But it's a decent model and worth using as a reference for what it may be worth.

 

[Zee]

There's no way that site would have seen the Leafs going 7-2-1 in their last 10.

From that statement alone it's clear that you don't really seem to understand what probabilities truly are.  I'm sorry if it sounds condescending, but it's true.  Given that, I think it goes without saying that considerations of the number crunching of raw objective data that leads to the calculation of such probabilities will be even further beyond your understanding.

I have a decent background in probabilities and the calculation thereof, and I have no problems with the means they use to calculate their probabilities.  You must understand that although it is it not a perfect model, there is no perfect model.  But it's a decent model and worth using as a reference for what it may be worth.

I think their system is severely flawed.  Look at that list above, they have Washington, Ottawa and Columbus at 18%, 26.7% and 57.4% chances of making the playoffs respectively.  They all have the same amount of points currently.    So they're saying there's a remote possibility of Washington making the playoffs despite behind 1 out of the wild card spot, but Columbus has better than 50% chance of making it even with the same number of points.  Something doesn't add up.

 

[CarltonTheBear]

So they're saying there's a remote possibility of Washington making the playoffs despite behind 1 out of the wild card spot, but Columbus has better than 50% chance of making it even with the same number of points.  Something doesn't add up.

Like I said with the TOR-MTL example, games played and ROW play a big factor here. Columbus has a game in hand on Washington so if they win that they are 2 points ahead of the Capitals. And the Caps would actually need to to be a point ahead of Columbus to pass them because of the tie-breaker.

 

[GhostofPotvin29]

There's no way that site would have seen the Leafs going 7-2-1 in their last 10.

From that statement alone it's clear that you don't really seem to understand what probabilities truly are.  I'm sorry if it sounds condescending, but it's true.  Given that, I think it goes without saying that considerations of the number crunching of raw objective data that leads to the calculation of such probabilities will be even further beyond your understanding.

I have a decent background in probabilities and the calculation thereof, and I have no problems with the means they use to calculate their probabilities.  You must understand that although it is it not a perfect model, there is no perfect model.  But it's a decent model and worth using as a reference for what it may be worth.

I think their system is severely flawed.  Look at that list above, they have Washington, Ottawa and Columbus at 18%, 26.7% and 57.4% chances of making the playoffs respectively.  They all have the same amount of points currently.    So they're saying there's a remote possibility of Washington making the playoffs despite behind 1 out of the wild card spot, but Columbus has better than 50% chance of making it even with the same number of points.  Something doesn't add up.

So if 3 teams have the same number of points right now, they should have the exact same odds of making the playoffs?

 

[bustaheims]

I think their system is severely flawed.  Look at that list above, they have Washington, Ottawa and Columbus at 18%, 26.7% and 57.4% chances of making the playoffs respectively.  They all have the same amount of points currently.    So they're saying there's a remote possibility of Washington making the playoffs despite behind 1 out of the wild card spot, but Columbus has better than 50% chance of making it even with the same number of points.  Something doesn't add up.

Washington is well behind in the tie-breaker (only 19 ROW) and have played more games than Columbus. As things stand right now, Columbus only has to gain a point on Detroit while staying at least even with Ottawa and Washington the rest of the way while not having New Jersey or Carolina earn 3 more points than they do (as long as they maintain the head to head tie-breaker with the Canes), and they're in thanks to the fact that they have more ROW. Washington has to earn at least 2 more points than Detroit and at least one more than Ottawa and Columbus, while not having New Jersey or Carolina earn more than 1 point more than they do. Factor in the potential difference in strength of schedule, and, yeah, things add up pretty nicely.

 

[Zee]

There's no way that site would have seen the Leafs going 7-2-1 in their last 10.

From that statement alone it's clear that you don't really seem to understand what probabilities truly are.  I'm sorry if it sounds condescending, but it's true.  Given that, I think it goes without saying that considerations of the number crunching of raw objective data that leads to the calculation of such probabilities will be even further beyond your understanding.

I have a decent background in probabilities and the calculation thereof, and I have no problems with the means they use to calculate their probabilities.  You must understand that although it is it not a perfect model, there is no perfect model.  But it's a decent model and worth using as a reference for what it may be worth.

I think their system is severely flawed.  Look at that list above, they have Washington, Ottawa and Columbus at 18%, 26.7% and 57.4% chances of making the playoffs respectively.  They all have the same amount of points currently.    So they're saying there's a remote possibility of Washington making the playoffs despite behind 1 out of the wild card spot, but Columbus has better than 50% chance of making it even with the same number of points.  Something doesn't add up.

So if 3 teams have the same number of points right now, they should have the exact same odds of making the playoffs?

No but I think the odds should be much closer than they are.  If Columbus loses the next two games do they suddenly drop from 57% to 18% because their record would be worse than the Caps?  I understand that they look at points and tiebreaker rules, but their random selection of wins/losses is where I take exception.  There's more to hockey then randomly selecting a winner based on record or home ice advantage.  Goalies get hot, players get hot, teams go into streaks. If you're only concerned with pure math and chance then sure it's a valid system of predicting who makes it in and who doesn't.  Other than that it has very little to do with anything.

 

[CarltonTheBear]

If you're only concerned with pure math and chance then sure it's a valid system of predicting who makes it in and who doesn't.  Other than that it has very little to do with anything.

What else do you think people are using this for?

 

[ZBBM]

Statistically, there is no such thing as momentum.  What happens is that when a team scores a goal and then comes out and has a couple of more good shifts, people attribute it to "momentum."  The vast majority of times, when there is no such pattern of good or bad play, people don't attribute it to anything (which is correct).  People are hard-wired to see patterns that are "caused" by human agency.

 

[Kin]

Statistically, there is no such thing as momentum.  What happens is that when a team scores a goal and then comes out and has a couple of more good shifts, people attribute it to "momentum."  The vast majority of times, when there is no such pattern of good or bad play, people don't attribute it to anything (which is correct).  People are hard-wired to see patterns that are "caused" by human agency.

Or there's this exchange from the terrific comedy The Trip:

"I feel like we just need to get you into the right movie. I think your career has got a lot of momentum."

"Yeah, but you gain momentum when you're going downhill."

 

[Optimus Reimer]

If we break down the remainder of the season into 7 game chunks, the Leafs need to win 4  games in each chunk.  That would be 12 more wins giving them 94 points which is very attainable, even moreso considering Nonis will be a buyer at the trade deadline trying to improve the team for the playoffs.  What I would like to see is the Leafs move ahead of Tampa or Montreal to start the playoffs at home.

 

[Zee]

If we break down the remainder of the season into 7 game chunks, the Leafs need to win 4  games in each chunk.  That would be 12 more wins giving them 94 points which is very attainable, even moreso considering Nonis will be a buyer at the trade deadline trying to improve the team for the playoffs.  What I would like to see is the Leafs move ahead of Tampa or Montreal to start the playoffs at home.

I like how you think.  4 wins out of every 7 games is very playoff-esque.

 

[CarltonTheBear]

I remember reading that's the mentality the Flames used for their 2nd half the year they went to the Finals.

 

[GhostofPotvin29]

You heard it team, emulate the Calgary Flames!  ;D

 

[moon111]

How many four-point games do these teams have?  What if division rivals go into overtime?  Teams can hope for a hot streak, but if their competition has the SAME hot streak, they've accomplished nothing.

 

[Highlander]

I think that Kessel and JVR will be so reved by the Olympics that they will be on afterburners when they put on the Blue and White again. I think they will be on fire, if Bernier can stand tall I think we will make it..Knock on woody.