RealGM

See previous topic: viewtopic.php?f=64&t=2260717!

Looking to collect more rankings and will update during the all-star game and also maybe if there is another pre-list topic for the RealGM top 100 thread this year?

Please provide suggestions on the Google Spreadsheet I'm looking to use to collect your rankings:
https://docs.google.com/spreadsheets/d/1sWU6bn6R9fd6tFLI_ghiGgOinthUYIJMLXGjdVDff1k/edit?usp=sharing

Also any comments on the visual / discuss the rankings! It seems that Mikan felt like a obliged inclusion for many
Surprised to see Curry so high!



Edit, updated by replacing the missing values with a hypothetical ranking of 32 or lower (since Wade had a vote at #31):

I am wondering how is n/a treated here. Kawhi has 5 votes to be top 30 but Harden got 11 votes. Wouldn't that mean Baylor, Kawhi and Jokic should rank behind as those people did not vote for him can have him 35, 40, or even more behind?

Giannis being so high seems strange to me in terms of how the rankings were calculated. I get the sense that you may have averaged the rankings players received without averaging in all the lists people did where he got left off. I mean maybe I'm wrong but generally speaking I think lists like this only work somewhat accurately for a much lower number than you ask to be ranked and then you have to use a points system rather than averaging the rank they are given. So if you want to do a top 20 for instance you still ask for people to rank 25 or 30 players then do a 30-29-28 points system. I commend you for the effort put into making this graph though.

I have to say I’m surprised at how many of these I’m the most extreme high or low on. Giannis, Kobe, sure, but no one else has David Robinson in the top 13, let alone the top 10? CP3 vs. Magic used to be a pretty trendy forum argument, now no one else has Paul in the top 15? No one else has Duncan at 3 or KG as high as 7? Literally 13 of my picks I’m either highest or lowest on, LOL.

Cavsfansince84 wrote:Giannis being so high seems strange to me in terms of how the rankings were calculated. I get the sense that you may have averaged the rankings players received without averaging in all the lists people did where he got left off. I mean maybe I'm wrong but generally speaking I think lists like this only work somewhat accurately for a much lower number than you ask to be ranked and then you have to use a points system rather than averaging the rank they are given. So if you want to do a top 20 for instance you still ask for people to rank 25 or 30 players then do a 30-29-28 points system. I commend you for the effort put into making this graph though.

Yeah, if you’re averaging the votes, you have to give people 31st or something when they’re not listed. The one that’s really egregious is Baylor. He’s only listed on 3 ballots and yet he’s ranked ahead of people with several high votes due to the calculation method.

iggymcfrack wrote:I have to say I’m surprised at how many of these I’m the most extreme high or low on. Giannis, Kobe, sure, but no one else has David Robinson in the top 13, let alone the top 10? CP3 vs. Magic used to be a pretty trendy forum argument, now no one else has Paul in the top 15? No one else has Duncan at 3 or KG as high as 7? Literally 13 of my picks I’m either highest or lowest on, LOL.

Worry not, you'd look more normal if I'd voted

Thanks for putting this together WestGOAT.

However, as per comments seen in post #5 itt, I would agree: there should some way of accounting when players are NOT on a poster's list at all.
13 out of 16 posters who replied don't have Elgin Baylor anywhere in their top 30.......yet he shows up at #23. Similar, but lesser issues arise with players like Mikan, Kawhi, Jokic, or Giannis.

If you want a simple solution, assume #31 for all posters who don't list them. Or better yet [if you're willing to take the time], ask each of us WHERE we rank each of the players listed whom we did NOT have on our ballots.

At any rate, SOMETHING to make it such that each of the players listed have equal number of votes.


Another potential solution might be to ask posters for their top 50 players.......but then only list the aggregate results for the top 30 that result (likely everyone---or nearly so---with have each of these lower consensus players SOMEWHERE in their top 50).

WestGOAT wrote:See previous topic: viewtopic.php?f=64&t=2260717!

Hi! You did not add The People's List, aka, the ChatGPT list. I am adopting it as mine.

Me: Who are the top 10 NBA players of all time?
ChatGPT: The top 10 NBA players of all time are:

1) Michael Jordan
2) LeBron James
3) Kareem Abdul-Jabbar
4) Kobe Bryant
5) Magic Johnson
6) Tim Duncan
7) Shaquille O'Neal
8) Larry Bird
9) Wilt Chamberlain
10) Hakeem Olajuwon.

viewtopic.php?p=103726630#p103726630

trex_8063 wrote:Thanks for putting this together WestGOAT.

However, as per comments seen in post #5 itt, I would agree: there should some way of accounting when players are NOT on a poster's list at all.
13 out of 16 posters who replied don't have Elgin Baylor anywhere in their top 30.......yet he shows up at #23. Similar, but lesser issues arise with players like Mikan, Kawhi, Jokic, or Giannis.

If you want a simple solution, assume #31 for all posters who don't list them. Or better yet [if you're willing to take the time], ask each of us WHERE we rank each of the players listed whom we did NOT have on our ballots.

At any rate, SOMETHING to make it such that each of the players listed have equal number of votes.


Another potential solution might be to ask posters for their top 50 players.......but then only list the aggregate results for the top 30 that result (likely everyone---or nearly so---with have each of these lower consensus players SOMEWHERE in their top 50).

yeah, getting a ranking (at least ballpark) for the left off seems like a good way to go. but maybe there are too many to do that? if we want to do that for the top 30 guys we didn't vote for, i'll say:

31. Mikan (i'll go in with the mikan at #30 crowd)
34. Stockton
35. Nash
36. Baylor

WestGOAT wrote:See previous topic: viewtopic.php?f=64&t=2260717!

Looking to collect more rankings and will update during the all-star game and also maybe if there is another pre-list topic for the RealGM top 100 thread this year?

Please provide suggestions on the Google Spreadsheet I'm looking to use to collect your rankings:
https://docs.google.com/spreadsheets/d/1sWU6bn6R9fd6tFLI_ghiGgOinthUYIJMLXGjdVDff1k/edit?usp=sharing

Also any comments on the visual / discuss the rankings! It seems that Mikan felt like a obliged inclusion for many
Surprised to see Curry so high!

Pl cahnge counting system b/c yours not ideal. This could be better, but has his flaws too.

1st 30pts
2nd 29pts
3rd 28pts
4th 27pts
.....
28th 3pts
29th 2pts
30th 1pts

Baylor cant be 23th with 3 mentions in TOP30

So with my counting system

1. Lebron 16
2. Jordan 16
3. KAJ 16
4. Russell 16
5. Duncan 16
6. Wilt 16
7. Hakeem 16
8. Magic 16
9. O'neal 15
10. Bird 15
11. Garnett 16
12. Kobe 15
13. Curry 15
14. Oscar 15
15. West 15
16. K.Malone 15
17. Dirk 14
18. D-Rob 15
19. Durant 14
20. Erving 14
21. Paul 13
22. Moses 14
23. Wade 14
24. Barkley 15
25. Mikan 9
26. Giannis 8
27. Stockton 9
28. Petitt 8
29. Nash 8
30. Harden 11

I actually like top 20, some small disagreements here and there but it's very good list.

After 20th spot, it gets fuzzier because some players shouldn't have been there with limited number of votes.

I'm somewhat surprised by Duncan being so uncontroversial. 3-9 is a very impressive range considering most media and casual fans have him like 9th at the highest.

Then his rival KG comes in with the complete opposite story. Votes inside the top 10 accompanied by a vote outside the top 20 and then a bunch of votes all over the place. It's interesting no more than 2 people voted him at any given spot. The next highest ranked played who has no more than 2 votes at his most popular spot is Moses Malone, over 10 places back. It shouldn't comes as too much of a shock since his impact is arguably GOAT tier but his accomplishments and play-off resume are barely top 20, if that.

dygaction wrote:I am wondering how is n/a treated here. Kawhi has 5 votes to be top 30 but Harden got 11 votes. Wouldn't that mean Baylor, Kawhi and Jokic should rank behind as those people did not vote for him can have him 35, 40, or even more behind?

Very valid question, and I haven't decided yet how to approach that issue, do you have any suggestions?

For now I did a simple arithmetic mean of the rankings, but I'm open to try something else as this approach has considerable drawbacks as pointed out in this thread.

Edit, replacing the missing rankings with a hyopthetical ranking of <32 also seems to work okay:

Cavsfansince84 wrote:Giannis being so high seems strange to me in terms of how the rankings were calculated. I get the sense that you may have averaged the rankings players received without averaging in all the lists people did where he got left off. I mean maybe I'm wrong but generally speaking I think lists like this only work somewhat accurately for a much lower number than you ask to be ranked and then you have to use a points system rather than averaging the rank they are given. So if you want to do a top 20 for instance you still ask for people to rank 25 or 30 players then do a 30-29-28 points system. I commend you for the effort put into making this graph though.

For example this points system, but then how would you take into account the hypothetical difference between a player that would have been ranked 31 vs one that would not be ranked at all for example?

Do I PM you my top 35 rankings?

Also you could don't count the best and the worst assessments, b/c they are least based. For example in gymnastic sport:

Gymnasts are awarded a score based on execution, artistry, technique and composition. The score starts at 10 and deductions are made for faults in execution, composition and artistry of presentation. The highest and lowest judges' scores are dropped.

WestGOAT wrote:

dygaction wrote:I am wondering how is n/a treated here. Kawhi has 5 votes to be top 30 but Harden got 11 votes. Wouldn't that mean Baylor, Kawhi and Jokic should rank behind as those people did not vote for him can have him 35, 40, or even more behind?

Very valid question, and I haven't decided yet how to approach that issue, do you have any suggestions?

For now I did a simple arithmetic mean of the rankings, but I'm open to try something else as this approach has considerable drawbacks as pointed out in this thread.

Edit, replacing the missing rankings with a hyopthetical ranking of <32 also seems to work okay:

Cavsfansince84 wrote:Giannis being so high seems strange to me in terms of how the rankings were calculated. I get the sense that you may have averaged the rankings players received without averaging in all the lists people did where he got left off. I mean maybe I'm wrong but generally speaking I think lists like this only work somewhat accurately for a much lower number than you ask to be ranked and then you have to use a points system rather than averaging the rank they are given. So if you want to do a top 20 for instance you still ask for people to rank 25 or 30 players then do a 30-29-28 points system. I commend you for the effort put into making this graph though.

For example this points system, but then how would you take into account the hypothetical difference between a player that would have been ranked 31 vs one that would not be ranked at all for example?

This is fantastic. 1000x better. I like the look of it being able to see the voters who left off each player too.

SpreeS wrote:Also you could don't count the best and the worst assessments, b/c they are least based. For example in gymnastic sport:

Gymnasts are awarded a score based on execution, artistry, technique and composition. The score starts at 10 and deductions are made for faults in execution, composition and artistry of presentation. The highest and lowest judges' scores are dropped.

I like letting every vote count. We already have done a lot of projects where we rank guys through the median opinion. This is just an average of every poster’s opinion and we can see them all represented visually.

WestGOAT wrote:

dygaction wrote:I am wondering how is n/a treated here. Kawhi has 5 votes to be top 30 but Harden got 11 votes. Wouldn't that mean Baylor, Kawhi and Jokic should rank behind as those people did not vote for him can have him 35, 40, or even more behind?

Very valid question, and I haven't decided yet how to approach that issue, do you have any suggestions?

For now I did a simple arithmetic mean of the rankings, but I'm open to try something else as this approach has considerable drawbacks as pointed out in this thread.

Edit, replacing the missing rankings with a hyopthetical ranking of <32 also seems to work okay:

Cavsfansince84 wrote:Giannis being so high seems strange to me in terms of how the rankings were calculated. I get the sense that you may have averaged the rankings players received without averaging in all the lists people did where he got left off. I mean maybe I'm wrong but generally speaking I think lists like this only work somewhat accurately for a much lower number than you ask to be ranked and then you have to use a points system rather than averaging the rank they are given. So if you want to do a top 20 for instance you still ask for people to rank 25 or 30 players then do a 30-29-28 points system. I commend you for the effort put into making this graph though.

For example this points system, but then how would you take into account the hypothetical difference between a player that would have been ranked 31 vs one that would not be ranked at all for example?

This is inevitable, but the main thing would be ask people to rank deeper than you want to actually display (e.g. if looking for a consensus top 30, ask for rank to 50 from each voter, if wanting 50 ask 100). The alternate (and inevitable unless you want to chuck out data at the end for, say the 31st and 32nd candidates or whatever is just beyond what you're after) is fringe/missing the cut candidates are ranked less accurately, off less input. And just acknowledge this.

But for those inevitably off the board, a rank just off the board is a solid, "satisficing" solution.

I already ranked 50 on the sheet

Wouldn't mind going around 70 deep for reference purposes.

WestGOAT wrote:

dygaction wrote:I am wondering how is n/a treated here. Kawhi has 5 votes to be top 30 but Harden got 11 votes. Wouldn't that mean Baylor, Kawhi and Jokic should rank behind as those people did not vote for him can have him 35, 40, or even more behind?

Very valid question, and I haven't decided yet how to approach that issue, do you have any suggestions?

For now I did a simple arithmetic mean of the rankings, but I'm open to try something else as this approach has considerable drawbacks as pointed out in this thread.

Edit, replacing the missing rankings with a hyopthetical ranking of <32 also seems to work okay:

Cavsfansince84 wrote:Giannis being so high seems strange to me in terms of how the rankings were calculated. I get the sense that you may have averaged the rankings players received without averaging in all the lists people did where he got left off. I mean maybe I'm wrong but generally speaking I think lists like this only work somewhat accurately for a much lower number than you ask to be ranked and then you have to use a points system rather than averaging the rank they are given. So if you want to do a top 20 for instance you still ask for people to rank 25 or 30 players then do a 30-29-28 points system. I commend you for the effort put into making this graph though.

For example this points system, but then how would you take into account the hypothetical difference between a player that would have been ranked 31 vs one that would not be ranked at all for example?

That's why I said you go like 10 past whatever sample you are trying to rank. So if you want to do a top 30 then you ask for lists of 40 or 45.