lessthanjake wrote:AEnigma wrote:lessthanjake wrote:Why in the world would we use a number that we know is wrong when we actually have the right one? That makes zero sense. You’re just wanting us to incorporate a known error into the analysis because it helps LeBron.
No, what I want is for you to stop doing an extremely error-prone approach for Jordan and then advertising it as some proof of his superiority. The more you refuse to do the same for Lebron, the more obvious it becomes you recognise that these margins of error are a lot larger than you care to admit.
I’ll repeat what I’ve already said in this thread: To the extent you keep asking me to make these calculations for LeBron, I don’t think you quite realize how long these calculations took (or maybe you do, and that’s why you’re asking that I do it). I spent a lot of time doing this for Jordan. It was a huge undertaking. I’m not obligated to spend a substantial portion of my day doing something just because you would like me to, and being unwilling to do so is not being evasive. You’re free to take the time to make a point you want to make.
I am aware, I have done it myself. That is why I am confident here lol. But you are also partly missing my point.
(And I also think you are overstating the extent of the time involved, but I acknowledge the time I take to calculate something might be faster than the time it takes others.)
I for one don't really care about it much because we both agree LeBron's teams played a higher pace with him off the court and therefore that an equal-pace assumption would spit out an incorrect value for him (which would overestimate his on-off). You’re just asking me to spend a ton of my time running numbers to confirm something I’m not contesting. The fact that the equal-pace assumption would overestimate LeBron’s on-off doesn’t in any way mean it is doing the same for Jordan and I think you really just need to step back and acknowledge that.
A.) I did acknowledge that there are unknowns which could change the implication of Jordan’s numbers.
B.) Again, I do not think that the pace variations themselves are as important as you understanding that there is an innate aspect of these calculations which does not match up with what you see listed on these different websites. But the only way you can understand that is to do a comparison and contrast.
I am sympathetic to not caring to do the calculations yourself, but I am less sympathetic when you continue to use it as an excuse to trumpet Jordan’s averages as a tier above.
You say the method I’m using is “extremely error-prone” but your only basis for that is that it would overestimate LeBron…but you also admit that it would overestimate LeBron precisely because you know that there *was* a material difference in pace for LeBron’s teams when he was on and off the court. We have no basis to believe that the same is true for Jordan, so we have no basis to believe that it would be as inaccurate for him (or even that, if it were inaccurate, it'd be inaccurate in the same direction).
No, I am saying you are making an assumption that on/off pace is the sole differentiating fact in “overestimation”.
And, of course, I identified this potential source of error upfront in my post presenting my data, so the implication that I’m not addressing the issue is obviously unwarranted. The correct way to deal with this issue isn’t to bizarrely use known incorrect information for LeBron—as if we somehow have any basis whatsoever to believe that the direction and magnitude of the error in assuming equal "on" and "off" pace is the same for both of them and therefore would cancel out if we use LeBron information that contains the error—but rather just to understand and acknowledge that the equal-pace assumption in the Jordan data is a potential error source which could go either way. Of course, again, that is something I clearly identified and pointed out upfront, so I don’t really know what you’re arguing about, except to try to lobby to use known-incorrect information for LeBron because it’d be more favorable to him. It’s not more accurate to introduce more error.
Well “accuracy” went out the window as soon as you started making broad claims about what hand calculations for Jordan signify, and it sailed on past the freeway when you glossed over Basketball-Reference being meaningfully lower than other sources
even with that supposedly “overestimating” pace distribution.
And your logic is that it is “like for like,” but we have no reason to believe that Jordan’s teams played at a substantially different pace with him on and off the floor, and certainly no reason to believe that they did so in the same direction and magnitude that LeBron’s teams did. Incorporating a known error in favor of LeBron when there’s no reason to think the same error exists for Jordan is just silly.
Now you are back to throwing things at the wall again. You did not know the pace disparity, you are assuming basketball-reference does, and you are assuming that the sole reason these numbers do not match up is because that on/off pace difference is significant enough to drag all of Lebron’s on/off numbers down.
I eagerly await the moment when someone less invested in maintaining Jordan’s pedestal decides to take the time to look at what I am talking about.
Basketball-Reference’s on-off data uses play by play data (which is why it only exists from 1996-1997 onwards), so…yeah, it does know about a pace disparity when someone is on and off the court.
Yet it mysteriously does not match other sources, its specific pace listings do not match other sources, and its outputs do not even pass an immediate face-value smell test (again, +10 over 6 games -> +1.0 per 100 possessions
).
Which brings us back to critical thinking and the apparent indifference to questioning any assumptions so long as they continue to let you publicly inflate Jordan.
Usually there at least can be common ground in the data used: someone pulling from BRef will do so 
