GDT: Anaheim at Columbus -placeholder edition

[Viqsi]

Pressure is off and once again we are playing ourselves out of the number one pick.

What is with this everpresent persistent delusion?

We are not getting the #1 pick. We never were getting the #1 pick. Or #2, or #3. We're fighting for #4 and below, because to do otherwise we'd have to win a lottery for the first time EVER in franchise history after FIFTEEN TRIES and outright LOSING AND MOVING DOWN rather than merely failing to win.

Seriously. Stop this. You might as well be bemoaning how we don't have unicorns magically making injuries go away and the puck go in the net because it's about as plausible.

* * *​

I disagree.

If we finish in worst case we'd pick 4th. But an 80+% probability of being in top 3.

Um. No. It is nowhere near that good. Calculating it exactly is extraordinarily difficult because subsequent results are based strongly on who won prior, but ultimately it ends up around 50-50 - y'know, close to what the odds were of retaining 1st overall in the prior setup. Try experimenting with the NHL Lottery Simulator if you don't believe me - run it many many many times and see how often Buffalo ends up in the top 3.

And how many times have we defied those kind of odds, for the worse? More than anyone else in the NHL's history.

 

[Doggy]

Question about losing the lottery...if we had not lost it the year we had the worst record in the league do you think we would have taken Yakupov?

If that's the case then the argument could certainly be made we won the lottery that year. Yakupov seems to be regressing while Murray is looks no worse than a #3 D-Man and probably more of a #2+.

 

[Viqsi]

Question about losing the lottery...if we had not lost it the year we had the worst record in the league do you think we would have taken Yakupov?

If that's the case then the argument could certainly be made we won the lottery that year. Yakupov seems to be regressing while Murray is looks no worse than a #3 D-Man and probably more of a #2+.

And the argument could also then be made that positioning yourself for the draft lottery is irrelevant. Especially since right now the best blueliners in that draft so far have been #5OA Morgan Rielly and #6OA Hampus Lindholm. Murray right now is in the Honorable Mentions alongside #9OA Jacob Trouba, #15OA Cody Ceci, and #22OA Olli Maatta.

 

[JacketsDavid]

Um. No. It is nowhere near that good. Calculating it exactly is extraordinarily difficult because subsequent results are based strongly on who won prior, but ultimately it ends up around 50-50 - y'know, close to what the odds were of retaining 1st overall in the prior setup. Try experimenting with the NHL Lottery Simulator if you don't believe me - run it many many many times and see how often Buffalo ends up in the top 3.

And how many times have we defied those kind of odds, for the worse? More than anyone else in the NHL's history.

It's 70%+ if worst record.
I ran the simulator just now 10 times and Buffalo was in top 3 - 8 times.
Maybe you can find a mathematical system - but if you give yourself 20% of first pick, then depending on who wins you'll have 20 opportunities (out of somewhere between 86.5-99 balls). If you lose again you'll have 20 opportunities (out of somewhere between 75-97 balls).
Obviously odds will be drastically worse if the 14th seed wins, but adds get much better if 2nd seed wins.
But it's 70% or so to be in top 3 with worst record. So I will stand corrected on 80% but it's still likely (70%) that you will be top 3 if you're the worst.

 

[The Four Seven]

Not surprised, given how the fight ended.

 

[BluejacketNut]

We can talk about poor attendance all we want, but I'm still amazed we draw a 2/3 to 3/4 full arena every game with what we've had to endure since our inception. There are only a hand full of cities in the NHL that would outdraw us if they had the same history we've had.

What happens is all the bad boys and girls get Bluejackets tickets for Xmas

 

[EspenK]

Question about losing the lottery...if we had not lost it the year we had the worst record in the league do you think we would have taken Yakupov?

If that's the case then the argument could certainly be made we won the lottery that year. Yakupov seems to be regressing while Murray is looks no worse than a #3 D-Man and probably more of a #2+.

I believe it has been reported that Murray was our guy regardless if we picked 1 or 2.

 

[EDM]

Dalton Prout is not a very good hockey player. But you have to be out of your mind to pick a fight with him the way Stewart did last nite. Seems that word should be pretty well spread around the NLH by now.

 

[Viqsi]

It's 70%+ if worst record.
I ran the simulator just now 10 times and Buffalo was in top 3 - 8 times.
Maybe you can find a mathematical system - but if you give yourself 20% of first pick, then depending on who wins you'll have 20 opportunities (out of somewhere between 86.5-99 balls). If you lose again you'll have 20 opportunities (out of somewhere between 75-97 balls).
Obviously odds will be drastically worse if the 14th seed wins, but adds get much better if 2nd seed wins.
But it's 70% or so to be in top 3 with worst record. So I will stand corrected on 80% but it's still likely (70%) that you will be top 3 if you're the worst.

When I say "many many many" times I mean 100+. 10 is nowhere near good enough for a reasonable conclusion.

EDIT: Hell, 100 probably isn't good enough by itself. I don't know Monte Carlo methods well enough to know when it converges, but I do know that when we're doing sampling runs at work for genetic analysis using MCMC sampling, we go for 3000 such samples. And SportsClubStats, which uses similar techniques to predict playoff standings, uses 10000. But I didn't have the patience to click the buttons that many times.

EDIT 2: My SportsClubStats info is out of date. They run just over two billion samples nowadays, apparently. Yowza.

 

[Johansen2Foligno]

Dalton Prout is not a very good hockey player. But you have to be out of your mind to pick a fight with him the way Stewart did last nite. Seems that word should be pretty well spread around the NLH by now.

You know what they say, if you can't beat 'em, beat 'em up.

 

[MAHJ71]

Not surprised, given how the fight ended.

 

[Hipster Letestu]

That was one of the cleanest punches I've ever seen landed in a hockey fight. Most are glancing shots off the shoulder/chest or a lucky jab that connects. That was a solid right cross that connected directly to his jaw. That was a knockout punch in a boxing match. Not surprised by this at all.

 

[major major]

It's 70%+ if worst record.
I ran the simulator just now 10 times and Buffalo was in top 3 - 8 times.
Maybe you can find a mathematical system - but if you give yourself 20% of first pick, then depending on who wins you'll have 20 opportunities (out of somewhere between 86.5-99 balls). If you lose again you'll have 20 opportunities (out of somewhere between 75-97 balls).
Obviously odds will be drastically worse if the 14th seed wins, but adds get much better if 2nd seed wins.
But it's 70% or so to be in top 3 with worst record. So I will stand corrected on 80% but it's still likely (70%) that you will be top 3 if you're the worst.

I did the math on some other thread, I can't remember where. But I came up with something like a 52% chance of top 3.

 

[JacketsDavid]

I did the math on some other thread, I can't remember where. But I came up with something like a 52% chance of top 3.

Has to be higher than that - you have 20% chance of first overall. In worst case scenario (14th place team getting #1 overall) you would have 20/99 chances for 2nd or 20.2%. Again in worst case if 13th overall won 2nd overall) you would have 20/97 chances for third or 20.6%.

So worst case it would be 60.8% and that assumes that the 1/100 and 2/99 chances win the first 2 spots.

There has to be some statistical nerd that figured it out but google couldn't help me
I know my original 80% is too high, and I know the floor would be 60.8%. So it's likely 64-68% of getting in top 3 if you have worst overall record.

 

[EspenK]

Has to be higher than that - you have 20% chance of first overall. In worst case scenario (14th place team getting #1 overall) you would have 20/99 chances for 2nd or 20.2%. Again in worst case if 13th overall won 2nd overall) you would have 20/97 chances for third or 20.6%.

So worst case it would be 60.8% and that assumes that the 1/100 and 2/99 chances win the first 2 spots.

There has to be some statistical nerd that figured it out but google couldn't help me
I know my original 80% is too high, and I know the floor would be 60.8%. So it's likely 64-68% of getting in top 3 if you have worst overall record.

Flipping your logic around it would be 20% for #1, 23.1 for 2nd assuming 2nd worse won and then 26.7 assuming 3rd worse got 2nd pick. That would make your best odds of being top 3 69.8.

This approach seems too simple but it probably frames the range somewhat realistically of between 60.8 and 69.8.

 

[major major]

Has to be higher than that - you have 20% chance of first overall. In worst case scenario (14th place team getting #1 overall) you would have 20/99 chances for 2nd or 20.2%. Again in worst case if 13th overall won 2nd overall) you would have 20/97 chances for third or 20.6%.

So worst case it would be 60.8% and that assumes that the 1/100 and 2/99 chances win the first 2 spots.

There has to be some statistical nerd that figured it out but google couldn't help me
I know my original 80% is too high, and I know the floor would be 60.8%. So it's likely 64-68% of getting in top 3 if you have worst overall record.

The mistake you are making is that you are calculating the odds at three different points in time and then adding them together. That will give you a different number from the odds of drafting top 3 going into the draft (one point in time). You have to multiply the odds of getting second (about 20 balls out of 90) by .8 because you have a .2 probability you won't even make it that far. And then multiply the third drawing by .625 I think.

I just did all the math here a few weeks ago, but can't remember which thread.

 

[EspenK]

The mistake you are making is that you are calculating the odds at three different points in time and then adding them together. That will give you a different number from the odds of drafting top 3 going into the draft (one point in time). You have to multiply the odds of getting second (about 20 balls out of 90) by .8 because you have a .2 probability you won't even make it that far. And then multiply the third drawing by .625 I think.

I just did all the math here a few weeks ago, but can't remember which thread.

I am not a stats whiz and I think from what you do you must be pretty good at it but I don't understand your comment about multiplying the second chance by .8 because if we don't make it that far it means we won 1st.

The three independent looks seem to make sense to me because that assumes we didn't win the previous round.

Apologies to blah but there is no Advanced Stats thread.

 

[major major]

I am not a stats whiz and I think from what you do you must be pretty good at it but I don't understand your comment about multiplying the second chance by .8 because if we don't make it that far it means we won 1st.

The three independent looks seem to make sense to me because that assumes we didn't win the previous round.

Apologies to blah but there is no Advanced Stats thread.

Believe it or not but the 20% chance of getting the first pick actually reduces your chances of getting the second pick.

 

[EspenK]

Believe it or not but the 20% chance of getting the first pick actually reduces your chances of getting the second pick.

Before you start the process I get that - I think. But if you lose the 1st draw you now have a better chance of winning 2nd although I suppose it now reduces your chance of winning third.

So before you start I guess I can buy your calculations. But for each draw individually I link JacketsDave's approach is correct.

And I'm now going to take an Advil

 

[JacketsDavid]

This approach seems too simple but it probably frames the range somewhat realistically of between 60.8 and 69.8.

So pretty close to the 64-68% range I said quoted above that

 

[major major]

Before you start the process I get that - I think. But if you lose the 1st draw you now have a better chance of winning 2nd although I suppose it now reduces your chance of winning third.

So before you start I guess I can buy your calculations. But for each draw individually I link JacketsDave's approach is correct.

And I'm now going to take an Advil

"Before you start" is what we are interested in here. What are the chances the last place team picks top 3, going into the draft. There's no difference between the way David and I are approaching the individual draws, he's just treating them as independent events, and I'm treating them as dependent, because they are dependent.

 

[JacketsDavid]

The mistake you are making is that you are calculating the odds at three different points in time and then adding them together. That will give you a different number from the odds of drafting top 3 going into the draft (one point in time). You have to multiply the odds of getting second (about 20 balls out of 90) by .8 because you have a .2 probability you won't even make it that far. And then multiply the third drawing by .625 I think.

I just did all the math here a few weeks ago, but can't remember which thread.

I think that's the way it works.

You have 100 (or 1000 balls actually because some are 1/2 percentages). So you have 20% at 1st pick.

After that winner then you draw again - and let's say a team with 10 balls (10%) won the lottery - so Columbus would not have 20/90 chances.

Then for 3rd place you would again do the process over.

So your numerator is always 20, and denominator has to get smaller each drawing (unless you win).

 

[major major]

As usual, someone on reddit had too much coffee and worked the entire thing out, and made a spreadsheet for everyone: https://m.reddit.com/r/theydidthema...st_what_are_the_odds_that_an_nhl_team_in_the/

And it's a 52.5% chance of picking top 3.

 

[EspenK]

I was thinking about this as I was driving to and from my errands this morning and I cam eup with this as a theory.

The odds on each individual draw as we have discussed range from 20% to 26.7 depending on who wins the first 2 draws. So to oversimplify this collectively we have 69.1 chances to win out of 261.5 balls. Wouldn't our best chances of being top 3 then be 69.1/261 =26.5. Dunno just thinking.

 

[JacketsDavid]

As usual, someone on reddit had too much coffee and worked the entire thing out, and made a spreadsheet for everyone: https://m.reddit.com/r/theydidthema...st_what_are_the_odds_that_an_nhl_team_in_the/

And it's a 52.5% chance of picking top 3.

I read it and can't follow.

They are saying there is a 17% chance we draft 2nd how is that possible to be lower than 20%?

I mean if we win the 20/100 balls it's 100% sure we'll be drafting in top 3 (e'd be first).

But if we lost first one w still get 20 balls on second attempt - and that would have to be greater than 20%.

Maybe this is why I'm an accountant instead of a statistician. I must be missing something.