NBA players vote Tyrese Haliburton as most overrated player in the league

[MrGoat]

That's a very healthy sample size, actually.

We get political polls based on a much smaller % of population taken to be a good sign of public opinion. Just for example.

As someone who makes a living off of statistics I'll tell you this. Political polls are based on usually around a thousand people or so which is actually a much better sample size than this even if the percentage of the population is way smaller. Starting at about a sample size of 30 you'll have roughly a 95% chance of the results roughly matching the population depending on the amount of parameters you have though still the bigger the sample size the better and the more confident you can be the sample is representative of the population, the problem with political polls is finding a truly representative sample thanks to archaic sampling methods. That could also be a problem here because we don't know how these players were sampled, the distribution of which teams were sampled could matter here

I mean, you said 30 and they used 90 which also happens to be one fifth of the league. I never heard of political polls that managed to get to 1/5 of the entire electorate. As a professional maybe you know better than I do though.

What I'm trying to say is the percentage matters less than the actual sample size, which is the main reason why with extremely large samples they don't try to match the percentage because that would be cost prohibitively expensive, but a corrupted sample can still hurt the numbers. You need a legitimate unbiased sample and that's not always an easy task

[Black Jack]

As someone who makes a living off of statistics I'll tell you this. Political polls are based on usually around a thousand people or so which is actually a much better sample size than this even if the percentage of the population is way smaller. Starting at about a sample size of 30 you'll have roughly a 95% chance of the results roughly matching the population depending on the amount of parameters you have though still the bigger the sample size the better and the more confident you can be the sample is representative of the population, the problem with political polls is finding a truly representative sample thanks to archaic sampling methods. That could also be a problem here because we don't know how these players were sampled, the distribution of which teams were sampled could matter here

I mean, you said 30 and they used 90 which also happens to be one fifth of the league. I never heard of political polls that managed to get to 1/5 of the entire electorate. As a professional maybe you know better than I do though.

What I'm trying to say is the percentage matters less than the actual sample size, which is the main reason why with extremely large samples they don't try to match the percentage because that would be cost prohibitively expensive, but a corrupted sample can still hurt the numbers. You need a legitimate unbiased sample and that's not always an easy task

Okay well - this is sort of a silly poll topic anyway. I do think it seems like 1/5 of the league, presumably from different teams, is a solid snapshot.

I'm not going to call for Tyrese to be banished to the G league here

[MrGoat]

I mean, you said 30 and they used 90 which also happens to be one fifth of the league. I never heard of political polls that managed to get to 1/5 of the entire electorate. As a professional maybe you know better than I do though.

What I'm trying to say is the percentage matters less than the actual sample size, which is the main reason why with extremely large samples they don't try to match the percentage because that would be cost prohibitively expensive, but a corrupted sample can still hurt the numbers. You need a legitimate unbiased sample and that's not always an easy task

Okay well - this is sort of a silly poll topic anyway. I do think it seems like 1/5 of the league, presumably from different teams, is a solid snapshot.

I'm not going to call for Tyrese to be banished to the G league here

I'm not saying it isn't, quite the opposite actually because that still is a huge percentage of the population when it comes to sampling. I'd even say unless it wasn't an extremely biased sample it's probably quite relatively accurate as to the whole NBA player's feelings on the matter. 13/90 players picked Haliburton, I'd be very shocked if that total number didn't rise by a decent bit if they didn't poll everyone. But it's still an observation based on the stats and I wonder who the players were with a sample like this that is much easier to track

[BruttoNostra]

I mean, you said 30 and they used 90 which also happens to be one fifth of the league. I never heard of political polls that managed to get to 1/5 of the entire electorate. As a professional maybe you know better than I do though.

What I'm trying to say is the percentage matters less than the actual sample size, which is the main reason why with extremely large samples they don't try to match the percentage because that would be cost prohibitively expensive, but a corrupted sample can still hurt the numbers. You need a legitimate unbiased sample and that's not always an easy task

Okay well - this is sort of a silly poll topic anyway. I do think it seems like 1/5 of the league, presumably from different teams, is a solid snapshot.

I'm not going to call for Tyrese to be banished to the G league here

I just participated in a similar discussion with a bunch of other data guys in some social network.
One of them works for a company that sent out a survey email and they got hundreds of responses (out of 50K emails sent), and the question is whether you could treat the results of the survey as statistically significant - and the answer is no, because it's highly possible that you will answer the survey only if you got a really bad experience and want to "punish" the company/its representative for giving you a bad service (or vice versa).
Say, out of 50K:
- 2% (1K) got a great experience
- 96% (48K) are OK, they actually forgot about the experience a second after the deal was closed
- 2% (1K) got really frustrated, told their friends and seek a revenge
It wouldn't surprise anyone if in the first group the response rate would be 10%, in the second one - 0.1%, and in the third one - 50% (I think, it's most of the time how frequently I respond for similar emails/messages)
So looking at the survey, you get 648 responses, 500 of them are negative.

Not saying it's what happened here, but the list of possible biases are:
- the poll was sent to all players, but since only 90 answered, some teams might have 6-7 voters, while others 0, and certain teams have certain agendas/biases
- the poll was sent to exactly 90 players (3 per team) who indeed answered, but the question is how they were chosen - vets vs rookies, starters vs bench guys, including what was already mentioned - what about american vs euro players?
- recency bias #1 - the top 4 players all played in the postseason/play-in - if the poll was conducted recently, some of the voters might have forgotten Beal or PG or Zion even exist - maybe if it was conducted a couple of months ago, those three would have led the poll?
- recency bias #2 - a single bad game/quote by a player might "leave" his name in the voters conscience until the next bad game by another player. Think about how a single game here in RealGM moves a player from a fat euro / "Nico was right" to a GOAT level and vice-versa.
- how the questions were sent. Say, it's a web link with a dropdown list of players to choose from (per team or overall) - the order they appear there matters. Or if it says who currently leads the poll, it would be easier for the next voter to be lazy and choose one of the "leaders" as well

[Homer38]

Another reason you can't take players vote seriously....

[MrGoat]

That's a very healthy sample size, actually.

We get political polls based on a much smaller % of population taken to be a good sign of public opinion. Just for example.

As someone who makes a living off of statistics I'll tell you this. Political polls are based on usually around a thousand people or so which is actually a much better sample size than this even if the percentage of the population is way smaller. Starting at about a sample size of 30 you'll have roughly a 95% chance of the results roughly matching the population depending on the amount of parameters you have though still the bigger the sample size the better and the more confident you can be the sample is representative of the population, the problem with political polls is finding a truly representative sample thanks to archaic sampling methods. That could also be a problem here because we don't know how these players were sampled, the distribution of which teams were sampled could matter here

As an example, did they balance foreign players with American born? We don't know and I think that matters,

Sure, there could even be other factors not mentioned yet or accounted for that very well mean the sampling isn't truly random. That's actually even more likely in such a small population as the NBA although also easier to audit if allowed. But I'm too lazy to try to think of what other potential arguments/methods they could use skew it but there are plenty

[Black Jack]

What I'm trying to say is the percentage matters less than the actual sample size, which is the main reason why with extremely large samples they don't try to match the percentage because that would be cost prohibitively expensive, but a corrupted sample can still hurt the numbers. You need a legitimate unbiased sample and that's not always an easy task

Okay well - this is sort of a silly poll topic anyway. I do think it seems like 1/5 of the league, presumably from different teams, is a solid snapshot.

I'm not going to call for Tyrese to be banished to the G league here

I just participated in a similar discussion with a bunch of other data guys in some social network.
One of them works for a company that sent out a survey email and they got hundreds of responses (out of 50K emails sent), and the question is whether you could treat the results of the survey as statistically significant - and the answer is no, because it's highly possible that you will answer the survey only if you got a really bad experience and want to "punish" the company/its representative for giving you a bad service (or vice versa).
Say, out of 50K:
- 2% (1K) got a great experience
- 96% (48K) are OK, they actually forgot about the experience a second after the deal was closed
- 2% (1K) got really frustrated, told their friends and seek a revenge
It wouldn't surprise anyone if in the first group the response rate would be 10%, in the second one - 0.1%, and in the third one - 50% (I think, it's most of the time how frequently I respond for similar emails/messages)
So looking at the survey, you get 648 responses, 500 of them are negative.

Not saying it's what happened here, but the list of possible biases are:
- the poll was sent to all players, but since only 90 answered, some teams might have 6-7 voters, while others 0, and certain teams have certain agendas/biases
- the poll was sent to exactly 90 players (3 per team) who indeed answered, but the question is how they were chosen - vets vs rookies, starters vs bench guys, including what was already mentioned - what about american vs euro players?
- recency bias #1 - the top 4 players all played in the postseason/play-in - if the poll was conducted recently, some of the voters might have forgotten Beal or PG or Zion even exist - maybe if it was conducted a couple of months ago, those three would have led the poll?
- recency bias #2 - a single bad game/quote by a player might "leave" his name in the voters conscience until the next bad game by another player. Think about how a single game here in RealGM moves a player from a fat euro / "Nico was right" to a GOAT level and vice-versa.
- how the questions were sent. Say, it's a web link with a dropdown list of players to choose from (per team or overall) - the order they appear there matters. Or if it says who currently leads the poll, it would be easier for the next voter to be lazy and choose one of the "leaders" as well

It's a good point but then, how do you get an unbiased sample? Even election phone polling is based on who will actually pick up a phone and answer political questions - I for one sure never do and hang up on pollsters because I can never be sure if it's some sort of data farming operation. If there was a pollster that's verified and I know they are legit I probably would be more likely to respond, but even then I would wonder sort of like is this data going to be stored on me later?

Email polls are total junk.

Basically there's no way to know about this specific poll unless they state their methods. I doubt they're super strict about correcting for bias, it's just a silly NBA players poll not something that people will take action on.

[Beethoven]

In my personal opinion (from what I gather from work and life in general)

I dont believe any of colleagues who are actually successful, positive, accomplished, etc take part in any of this type of vote-casting or anything playground that tries to belittle someone or throw shade on someone who is trying their best and got some level of accomplishment or something.
It is always those in the rear who are not really noticed, cant get anywhere, cynical people who conjures up stuff like this about others getting some acknowledgement or perhaps even surpassing them at work.
In general, of course.

[BruttoNostra]

Okay well - this is sort of a silly poll topic anyway. I do think it seems like 1/5 of the league, presumably from different teams, is a solid snapshot.

I'm not going to call for Tyrese to be banished to the G league here

I just participated in a similar discussion with a bunch of other data guys in some social network.
One of them works for a company that sent out a survey email and they got hundreds of responses (out of 50K emails sent), and the question is whether you could treat the results of the survey as statistically significant - and the answer is no, because it's highly possible that you will answer the survey only if you got a really bad experience and want to "punish" the company/its representative for giving you a bad service (or vice versa).
Say, out of 50K:
- 2% (1K) got a great experience
- 96% (48K) are OK, they actually forgot about the experience a second after the deal was closed
- 2% (1K) got really frustrated, told their friends and seek a revenge
It wouldn't surprise anyone if in the first group the response rate would be 10%, in the second one - 0.1%, and in the third one - 50% (I think, it's most of the time how frequently I respond for similar emails/messages)
So looking at the survey, you get 648 responses, 500 of them are negative.

Not saying it's what happened here, but the list of possible biases are:
- the poll was sent to all players, but since only 90 answered, some teams might have 6-7 voters, while others 0, and certain teams have certain agendas/biases
- the poll was sent to exactly 90 players (3 per team) who indeed answered, but the question is how they were chosen - vets vs rookies, starters vs bench guys, including what was already mentioned - what about american vs euro players?
- recency bias #1 - the top 4 players all played in the postseason/play-in - if the poll was conducted recently, some of the voters might have forgotten Beal or PG or Zion even exist - maybe if it was conducted a couple of months ago, those three would have led the poll?
- recency bias #2 - a single bad game/quote by a player might "leave" his name in the voters conscience until the next bad game by another player. Think about how a single game here in RealGM moves a player from a fat euro / "Nico was right" to a GOAT level and vice-versa.
- how the questions were sent. Say, it's a web link with a dropdown list of players to choose from (per team or overall) - the order they appear there matters. Or if it says who currently leads the poll, it would be easier for the next voter to be lazy and choose one of the "leaders" as well

It's a good point but then, how do you get an unbiased sample? Even election phone polling is based on who will actually pick up a phone and answer political questions - I for one sure never do and hang up on pollsters because I can never be sure if it's some sort of data farming operation. If there was a pollster that's verified and I know they are legit I probably would be more likely to respond, but even then I would wonder sort of like is this data going to be stored on me later?

Email polls are total junk.

Basically there's no way to know about this specific poll unless they state their methods. I doubt they're super strict about correcting for bias, it's just a silly NBA players poll not something that people will take action on.

Now, that's a good point! So many polls around us, in media etc are very biased, I realised it during my Master.
As for election polls in particular, I think (not sure for the US, though) in many cases they choose a representative sample* who are called and required to answer since they are being paid. Or ar least it's how some of the more established institutes are doing it. If some local tv station were randomly calling me - I would hang out, just like you did.

*there is some funny story from 2016 elections, IIRC, with an afro-american guy who also represented some other minority group (maybe LGBT?), and since he answered Trump and his vote is so pivotal for the poll results (so many groups he is representing), it skewed the results in an interesting way

Edit: found the link:
Original (paywall): https://www.nytimes.com/2016/10/13/upshot/how-one-19-year-old-illinois-man-is-distorting-national-polling-averages.html
A free one: https://theweek.com/speedreads/654767/how-single-19yearold-man-illinois-wildly-skewed-major-poll-towards-trump

In some polls, [the 19-year-old is] weighted as much as 30 times more than the average respondent, and as much as 300 times more than the least-weighted respondent.Alone, he has been enough to put Mr. Trump in double digits of support among black voters. He can improve Mr. Trump's margin by 1 point in the survey, even though he is one of around 3,000 panelists.He is also the reason Mrs. Clinton took the lead in the USC/LAT poll for the first time in a month on Wednesday. The poll includes only the last seven days of respondents, and he hasn't taken the poll since Oct. 4. Mrs. Clinton surged once he was out of the sample for the first time in several weeks.

[doogie_hauser]

Reckon this poll can only work to Hali and the Pacers favour tbh.

I think he enjoys being the pantomime villain of the NBA to be honest

[Optms]

Harden for sure voted Giannis lol

[Drygon]

Read on Twitter

Read on Twitter

[littlerock2277]

I will gladly take this overrated player on the raptors any day of the week. Ja morant or Jaren Jackson Jr should be leading this list.

[slickrickstyles]

Reckon this poll can only work to Hali and the Pacers favour tbh.

I think he enjoys being the pantomime villain of the NBA to be honest

Has a great legacy to follow...Reggie Miller once released a novel called "I Love Being The Enemy" while still in the league haha

[doogie_hauser]

Reckon this poll can only work to Hali and the Pacers favour tbh.

I think he enjoys being the pantomime villain of the NBA to be honest

Has a great legacy to follow...Reggie Miller once released a novel called "I Love Being The Enemy" while still in the league haha

Feels to me like the small market/country community of Indianapolis has always had a us vs The world type of chip on their shoulder, and it's only natural they tend to love players who love to get under opposition skins like Reggie and Hali etc

[Scottie Pimpin]

Tyrese seems pretty likeable so I found this surprising. After seeing his dad storming the court to trash talk Giannis, I wouldn't be surprised if he does this with other NBA players and this influenced their decision to pick Tyrese lol.

[xxSnEaKyPxx]

Read on Twitter

[Sgt Major]

Read on Twitter

[axeman23]

I wonder is 1 of those players was Giannis...

[axeman23]

Read on Twitter

The new "Take THAT for data!"?