6-11, ESPN"s latest prediction of our season

[pmedic920]

I don’t see a problem here.

I predicted a 6 win season, many others did as well.

What’s the issue?

Sarc:{)

 

[knownone]

I know, but it's fresh download from ESPN just minutes ago. And ESPN claimed their prefictor's analytics based.

It is analytics-based. However, ESPN's model weights preseason predictions pretty heavily, so it will be off quite a bit as we advance.

 

[HawkRiderFan]

Have they updated their date since pre-season before running the simulation? It sure doesn't look like it

 

[Rainger]

Have they updated their date since pre-season before running the simulation? It sure doesn't look like it

ESPN is lazy shit, more concerned with woke social re-engineering than sports!

 

[Nunya]

GIGO = Garbage in...Garbage out.

 

[bmorepunk]

Have they updated their date since pre-season before running the simulation? It sure doesn't look like it

Each game may not be updated until the week before.

I know there's a bunch of people who want to act like the whole world is against them, but many of you seem to have poor understandings of what models and probabilities are.

The models used to generate these probabilities may or may not be useful. I have no idea what the equations, initial conditions, or weightings are.

The idea that you take all those probabilities and you get "6-11, ESPN"s latest prediction of our season" demonstrates a near zero concept of basic probabilty.

Those last two games are both close to 50% splits. If we just take that on face value and ignore dependencies, it's nearly equally probable that they 0-2, 1-1, or 2-0 on those games (with some small probabilities of one or more ties).

None of the numbers you are looking at are predictive of a single event. They are statistically modeled distributions of many events. Think of a very large number of parallel universes where the games are played. Your team is modeled to win in that percentage of that large number of parallel universes.

 

[Rainger]

Each game may not be updated until the week before.

I know there's a bunch of people who want to act like the whole world is against them, but many of you seem to have poor understandings of what models and probabilities are.

The models used to generate these probabilities may or may not be useful. I have no idea what the equations, initial conditions, or weightings are.

The idea that you take all those probabilities and you get "6-11, ESPN"s latest prediction of our season" demonstrates a near zero concept of basic probabilty.

Those last two games are both close to 50% splits. If we just take that on face value and ignore dependencies, it's nearly equally probable that they 0-2, 1-1, or 2-0 on those games (with some small probabilities of one or more ties).

None of the numbers you are looking at are predictive of a single event. They are statistically modeled distributions of many events. Think of a very large number of parallel universes where the games are played. Your team is modeled to win in that percentage of that large number of parallel universes.

I understand you are trying to do a some statistical justification, BUT see my previous post. ESPN writers are LAZY and as Nunya says garbage in garbage out.

WHY would ANY sports organization base their data on pre season projections???? We are 1/2 way through the season. Do the modeling based on the current situation.

It is just like the dip shit who has done 2 draft projections on CBS using the VEGAS pre season betting lines of draft position showing the Hawks pick at 5 and our pick from Denver at like 22.

That is just lazy shit, not taking any effort to say, well based on current records maybe just maybe the Hawks wont be drafting 5 (with their pick lol) and taking the player the lazy writer says.

 

[LeveeBreak]

I understand you are trying to do a some statistical justification, BUT see my previous post. ESPN writers are LAZY and as Nunya says garbage in garbage out.

WHY would ANY sports organization base their data on pre season projections???? We are 1/2 way through the season. Do the modeling based on the current situation.

It is just like the dip shit who has done 2 draft projections on CBS using the VEGAS pre season betting lines of draft position showing the Hawks pick at 5 and our pick from Denver at like 22.

That is just lazy shit, not taking any effort to say, well based on current records maybe just maybe the Hawks wont be drafting 5 (with their pick lol) and taking the player the lazy writer says.

yea...^^^^^ What is currently coveyed is not currently relevant.

 

[HawkRiderFan]

I think I understand probabilities enough to think that at 5-3, any model predicting the Hawks to go 1-8 the rest of the way based on the way they are playing, and their remaining opponents are playing is highly suspect. That's just numbers, not any "everyone's against us" feeling

 

[bmorepunk]

I think I understand probabilities enough

You don't.

 

[Jegpeg]

Those last two games are both close to 50% splits. If we just take that on face value and ignore dependencies, it's nearly equally probable that they 0-2, 1-1, or 2-0 on those games (with some small probabilities of one or more ties).

That is not actualy true. The following are nearly equally probable in a game that is a 50% split: Win both games, lose both games, Win the first game and lose the second, Lose the first game and win the second. Therefore the probability of winning 0 games is about 25%, winnging 1 game about 50% and winning both games about 25%.

 

[bmorepunk]

That is not actualy true. The following are nearly equally probable in a game that is a 50% split: Win both games, lose both games, Win the first game and lose the second, Lose the first game and win the second. Therefore the probability of winning 0 games is about 25%, winnging 1 game about 50% and winning both games about 25%.

You're right; I'm treating them as independent events so it should be P(A) * P(B) like you said.

 

[nwHawk]

Well, those paper tiger nothing without Hall Jets just beat the Buffalo Bills. Still think that will be an easy game?

 

[toffee]

Officially, we can tell ESPN to F their analytics, one game at a time:. This one just went down 38.2% my azz.

 

[SmokinHawk]

ESPN's analytics are clearly just aggregating stats and low precision metrics over the course of the season and averaging them out because this is not the same team we trotted out in week 2.

 

[HawkinNY]

Would love to see how the whole season was picked. Cause they prob had us at 2-3 wins with these picks.

 

[Ruminator]

For Seattle to win just 1 or 2 more games to fulfill ESPN's and others' predictions, half of the team would need to go blind, have lobotomies, or die in the next week or two.

 

[Rainger]

For Seattle to win just 1 or 2 more games to fulfill ESPN's and others' predictions, half of the team would need to go blind, have lobotomies, or die in the next week or two.

Well according to ESPN the Hawks won't win another game this year since they already have 6

 

[Sun Tzu]

You don't.

Yeah, this whole thread highlights that few of our regular posters have even a basic understanding of probabilities.