538 Elo forecast gives Raptors highest chance of winning finals (31%)

[Wannabe MEP]

Elo gives extra weight to recent performance. With the strangeness of the bubble, it's hard to know how relevant recent performance is. On the one hand, these games didn't really matter for some teams because their position is locked, and it's more like worthless preseason games. On the other hand, things are just really different now than they were a few months ago during the regular regular season, and so how teams have fared in the bubble could be much predictive of how they are going to play in the playoffs than how they were playing months ago.

After a small sample size of playoff games, it is starting to seem like teams playing well in the bubble have continued to play well in the playoffs, and vice versa. We'll see how much that continues.

[True North]

Th Raptors didn’t play today. After today’s results:

Their odds of winning the CHIP according to RAPTOR increased from 5% to 6%

Their odds of winning the CHIP according to ELO decreased from 37% to 34%

[J-Roc]

They purposely hide that ELO thing because they don't believe in it. They changed up their site this year to use RAPTOR and up till now it's been a failure. Like even if they updated the RAPTOR by next round to move us up, it's still a failure as it didn't predict the Raps to finish with the 2nd best record in the league.

These statistical models are not meant to be crystal balls.

What are they meant to be? I thought they have those percentages literally trying to predict future outcomes like crystal balls actually do.

[everdiso]

Th Raptors didn’t play today. After today’s results:

Their odds of winning the CHIP according to RAPTOR increased from 5% to 6%

Their odds of winning the CHIP according to ELO decreased from 37% to 34%

1. I'm looking forward to the end of the first round update. Right now the raps are benefitting from the easy 1st round matchup increasing their odds substantially. Things will be clearer once that is factored out.

2. The only thing that really effects tho is ELO, not as much the RAPTOR forecast. I really like 538 but given that this player metric is brand new, its interesting to me that RAPTOR views our players worse than the other impact metrics do....(especially strange since CARMELO was higher on most of these players last year - and was successful in its predictions)....

Our 7 man rotation from best to worst:

RPTR: 4.4, 3.0, 1.8, 1.1, 1.0, 0.3, -1.0 (10.6)
PIPM: 3.2, 2.8, 2.2, 1.4, 1.1, 0.5, -0.3 (10.9)
RPM: 5.2, 3.2, 2.1, 2.0, 1.9, 0.6, -0.2 (14.8)
BPM: 3.2, 2.7, 2.1, 2.0, 1.9, 1.7, 1.2 (14.8)

[Wannabe MEP]

These statistical models are not meant to be crystal balls.

What are they meant to be? I thought they have those percentages literally trying to predict future outcomes like crystal balls actually do.

[Wannabe MEP]

I really like 538 but given that this player metric is brand new, its interesting to me that RAPTOR views our players worse than the other impact metrics do....(especially strange since CARMELO was higher on most of these players last year - and was successful in its predictions)....

Our 7 man rotation from best to worst:

RPTR: 4.4, 3.0, 1.8, 1.1, 1.0, 0.3, -1.0 (10.6)
PIPM: 3.2, 2.8, 2.2, 1.4, 1.1, 0.5, -0.3 (10.9)
RPM: 5.2, 3.2, 2.1, 2.0, 1.9, 0.6, -0.2 (14.8)
BPM: 3.2, 2.7, 2.1, 2.0, 1.9, 1.7, 1.2 (14.8)

• Loss of Kawhi, and also advanced stats liked Danny Green a lot.
• RAPTOR is a mix of a +/- and box score stats. The Raptors don't have any box score superstars: no Harden, LeBron, etc.

[mrdressup]

That's from the starting pool. Our odds get instantly worse when one of the LA teams loses, and one must lose.

[Wannabe MEP]

That's from the starting pool. Our odds get instantly worse when one of the LA teams loses, and one must lose.

I'm not sure what you're saying. Explain?

[mrdressup]

That's from the starting pool. Our odds get instantly worse when one of the LA teams loses, and one must lose.

I'm not sure what you're saying. Explain?

A lot of the odds of winning the final are split between the LA teams. As soon as one loses the other's chances skyrocket. It's like the odds of winning a hand in poker. It changes with every street. Our chances of winning the final are not 31% if we lose in round two. We may win in round one and they could theoretically get worse than that of one of the LA teams if either loses in the first round. No matter what i would expect our odd of winning the final would be the lowest of any of the top teams if we get to the final.

[Wannabe MEP]

A lot of the odds of winning the final are split between the LA teams. As soon as one loses the other's chances skyrocket. It's like the odds of winning a hand in poker. It changes with every street. Our chances of winning the final are not 31% if we lose in round two. We may win in round one and they could theoretically get worse than that of one of the LA teams if either loses in the first round. No matter what i would expect our odd of winning the final would be the lowest of any of the top teams if we get to the final.

I don't think you understand how the math works. First of all, the "ELO RATING" shows the Raptors as the best team right now. So if we simply erased the other teams and pitted the Raptors against any one other team, the Raptors would be the favorite, according to Elo. That doesn't mean Elo is correct, but that's what Elo says. Now, you're right that things will change over time, so there's no guarantee that the Raptors would be the favorite in the final, but that depends on how the teams look getting there. If the Raptors squeak by every round while the Clippers blow everybody out en route to the Finals, then yeah, the Raptors would be underdogs vs the Clippers. But Elo says that the best three teams remaining are in the East, so most likely, the East team will be favored in the Finals by Elo.

RAPTOR is different. RAPTOR likes the LA teams, because it puts a greater emphasis on "superstars".

[lolwut]

They purposely hide that ELO thing because they don't believe in it. They changed up their site this year to use RAPTOR and up till now it's been a failure. Like even if they updated the RAPTOR by next round to move us up, it's still a failure as it didn't predict the Raps to finish with the 2nd best record in the league.

These statistical models are not meant to be crystal balls.

What are they meant to be? I thought they have those percentages literally trying to predict future outcomes like crystal balls actually do.

Say Team A has a 60% chance to win and Team B has a 40% chance to win.

Does that mean Team A will win?

[Wannabe MEP]

Also, yes, when one of the LA teams loses, he odds of the other LA go up. But that's just as true in the East. If, say, the Clippers beat the Lakers, and the Raptors beat the Bucks, then the odds of the Clippers go up a lot...as do the the odds of the Raptors. If the series are similarly competitive, then the Raptors' odds would actually go up more than the Clippers, because Elo thinks the Bucks are better than the Lakers (at least right now).

[J-Roc]

These statistical models are not meant to be crystal balls.

What are they meant to be? I thought they have those percentages literally trying to predict future outcomes like crystal balls actually do.

Say Team A has a 60% chance to win and Team B has a 40% chance to win.

Does that mean Team A will win?

Okay, then what's the point of trying to predict the more likely outcome? Because you're trying to predict what will happen?

[lolwut]

What are they meant to be? I thought they have those percentages literally trying to predict future outcomes like crystal balls actually do.

Say Team A has a 60% chance to win and Team B has a 40% chance to win.

Does that mean Team A will win?

Okay, then what's the point of trying to predict the more likely outcome? Because you're trying to predict what will happen?

That's the part you're misunderstanding. Predicting the more likely outcome is NOT the same as predicting what will happen.

Let's go with an even simpler example. A coin flip is 50/50. Everyone knows that. So...if I tell you a coin flip is 50/50, I'm not trying to predict which side of the coin it's going to land on. I'm simply stating the probabilities of this scenario.

[radeonboy]

All very nice, but the Clippers are winning it all. Any algorithm that doesn't reflect that all-but-objective fact is flawed.

I just don't see the Clippers getting it done.

[True North]

After Fri Aug 21 results.





After the Raptors win today the CHIP probability using RAPTOR increased to 7% but back down to 6% after the Celtics took care of their business

[True North]

Overnight the Raptors CHIP probability rose from 6% to 10% using RAPTOR

[everdiso]

Overnight the Raptors CHIP probability rose from 6% to 10% using RAPTOR

Strange. What drove that?

Overnight change in RAPTOR 'ship odds:

LAC 28% ----> 27% (-1)
BOS 20% ----> 14% (-6)
LAL 19% -----> 19% (--)
HOU 13% ----> 15% (+2)
MIL 10% -----> 11% (+1)
TOR 6% -------> 9% (+3)
MIA 2% -------> 2% (---)
UTA 1% -------> 2% (+1)

Celtics with the big drop for some reason.

[TheBoi10]

Overnight the Raptors CHIP probability rose from 6% to 10% using RAPTOR

Strange. What drove that?

Overnight change in RAPTOR 'ship odds:

LAC 28% ----> 27% (-1)
BOS 20% ----> 14% (-6)
LAL 19% -----> 19% (--)
HOU 13% ----> 15% (+2)
MIL 10% -----> 11% (+1)
TOR 6% -------> 9% (+3)
MIA 2% -------> 2% (---)
UTA 1% -------> 2% (+1)

Celtics with the big drop for some reason.

Maybe they updated the individual RAPTOR ratings and/or corrected minute distributions

[everdiso]

Overnight the Raptors CHIP probability rose from 6% to 10% using RAPTOR

Strange. What drove that?

Overnight change in RAPTOR 'ship odds:

LAC 28% ----> 27% (-1)
BOS 20% ----> 14% (-6)
LAL 19% -----> 19% (--)
HOU 13% ----> 15% (+2)
MIL 10% -----> 11% (+1)
TOR 6% -------> 9% (+3)
MIA 2% -------> 2% (---)
UTA 1% -------> 2% (+1)

Celtics with the big drop for some reason.

Maybe they updated the individual RAPTOR ratings and/or corrected minute distributions

I guess it could have been a delayed update to Hayward's status.