ISOLATION SCORING, Statistical analysis of LIN vs HARDEN and the top 50 scorers ISO rankings

Why only focus on one stat when several were mentioned? Because you wanna be difficult.

Also, Ast/TO ratio was mentioned and you still disregarded that portion of the post. Again, no surprise.

 

My point is that it pollutes the data in terms of determining bench vs starter play.

Beverley with full roster vs Lin with full roster is the only way to determine with any statistical validity their relative performance unless you plan on doing some very sophisticated statistical techniques in an attempt to clean it up. And even then, it requires a ton of assumptions and is of debatable value.

And even then, it doesn't address the sample size.

 

Maybe somebody should tell Clutch to lock all threads were there's insufficient data in the OP. I didn't start this thread. The entire OP is based on a small sample size. Go and chastise him for conjuring an idea with limited info.

Or how about addressing my first response to you in this very thread. You jumped in like you were high and mighty, but neglected to answer a simple questions. You dont want a legit conversation, you just wanna be difficult.

LAME

 

The context is irrelevant. You posted data, and then said it was indicative that Lin is superior off the bench. Regardless of my personal opinions, there is insufficient data to argue for either Lin as a starter OR off the bench from the data YOU presented.

Beverley might very well be the superior choice as the starter. However, until there's more convincing data your post was opinion more than anything unless you're presenting a reasonable statistical argument.

As for the original sample size presented in the original post at the start of the thread, I didn't say anything as the Synergy data presented is actually in line with 2012-13 and 2011-12 data. If anything, it's actually less impressive than Lin's previous PPP and 3 point shooting after the 2012 ASB.

Not to mention, VOA is more than adequately versed in statistics to defend himself.

If you wish to use statistics for or against a position, then expect it to be open to public criticism in a DISCUSSION thread.

 

Again, go and chastise the OP for conjuring an idea with limited data. You responded multiple times in this thread before my initial post and made no mention of sample size. You just wanna be difficult. .. and that's cool.

Your boy VOA is too much on his high horse to admit he exaggerated when he said FG% wad useless in a different thread. He subsiquently bypassed simple facts in this thread to carry on his lame argument from days ago.

Lame

 

I had to put that out there because a lot of jokers on here so quick to call you that.

Man I didn't know you were on Texans4life. What a terrible season huh lol.

 

Except small sample size does NOT apply to VOA's original argument. In fact, VOA made no argument at all.

He presented factual data, and said that the most efficient scorers in isolation were Lin and Harden to date. There's nothing to argue as it was a statement of fact (that happened to be correct).

He presented his data, and presented the source of his data. That is a correct use of data.

If I said that Harden scored 45 points, there is no sample size to debate. It's either a correct fact or it's not, and not subject to debate unless you're arguing the fact itself is incorrect.

You on the other hand were positing a "fact" with data that did not support that fact due to lack of sample size and polluted data. That is NOT the same.

And FG% is not entirely useless, but it is a flawed statistic in many ways. I could go on to outline why FG% is flawed in many ways, but I've already posted on that topic in multiple threads.

So, in summary...

Factual statement presented with data that is properly sourced and factually correct does NOT require sample size.

Interpretation of data that is presented as factually correct and not sourced is subject to discussion in terms of sample size, context and validity.

And your particular disagreement with VoA's points is something that you two can work out between yourselves.

 

<iframe width="640" height="360" src="//www.youtube.com/embed/rR3NhE8fBs8" frameborder="0" allowfullscreen></iframe>


I'd rather have Lin workit.

 

That was end of quarter on a hot Lin quarter in a one-on-one situation where he was feeling it. But that not his norm. The big part of that was 12 of his 27 points were scored in the 4th. Lin's a strong 4th quarter player.

 

1) There's only been 25 games this season, 33% of which Lin has missed. James missed some and/or wasn't healthy for the majority of the games. Sample size is a factor and context are factors.

2) Obviously the OP was making some sort of argument regarding efficiency. The facts he used were to combat anyone claiming Lin or Harden was inefficient or one player was less efficient than the other (a VERY common discussion in this forum).

1) The OP didn't simply state one player's stats. He compared stats against other players. VERY different than making a simple statement of ____ scored ____.

2) If you said Harden scored however many points, you'd likely add the fact of the amount of games.. OR it would be understood how many games you meant due to context. Sample size is would still be an issue, especially if you were trying to make a bigger point (i.e. Harden is a great scorer). The OP was making a bigger point than a simple point total from one game.

My data indeed supported the point I was making... sample size was just an issue. It's EXACTLY the same as the OP.

Everyone knows FG% has flaws, but it's far from useless. Your boy VOA stated it was COMPLETELY USELESS and has yet to man up and say he was WRONG.

So someone can use Dwight's 1-1 3PM from the GS game and say he's a good 3pt shooter without sample size? Don't be silly

Go ahead and say Dwight's 1 3PM isn't properly sourced, then chastise the OP for comparing stats between players when one has been hobbled most of the season.

 

There's some cool guys at T4L.

The season has been disappointed, but the fact Kubes and Schaub will be gone makes everything worth it. We even have a #1 pick to boot.

 

Do you *REALLY* want to go down this road? I mean, seriously?

First, VoA made the following statements...

So, while a little sensationalistic, he asks the question as to what the data has shown so far, cites the source (Synergy), and says factually that had the scoring been less efficient the season could have been worse, which is factually true since if their efficiency is lower then the Rockets score less.

Note that he did not state any conclusions regarding who is the better player, or more talented player, or make projections of their performance into the future. He just asked what the data shows and provided it.

IE, statement of fact.

Now, while VOA *did* offer an opinion in terms of the validity of statistical analysis in the context of 25 games...

...that is an opinion that he stated, and is not directly relevant as his original post is NOT an analysis of the data versus a presentation of the existing data (IE, statement of fact).

Additionally, the use of a 25 game sample size depends on the data you are attempting to interpret, the context of that data, and the methods you use to analyze that data.

In other words, you can use a 25 games sample size for certain types of data more effectively than other types of data, assuming you have the expertise to handle that data appropriately.

 

I just wanted to mention that Lin is 7th in the league in drives per game and he has the best FG% in those drives out of the top 25 players in that category. (8.0 drives per game / 64.2 FG%)

Parsons - 6.4 drives per game / 54.8 FG%

Harden:

 

Another one bites the dust. The tuna turned into sashimi or should I say "canned tuna".

 

You're good. Just don't like your repeated dodging of the fact, "FG% isn't USELESS." Will dodge again, but no biggie.

 

Cheerlead much?

 

The data itself is not screwed. The data is the data. You can argue the system that Synergy uses if you wish, or how they model that data, but the data is the data.

Analysis or conclusions drawn from that data can be problematic. And misusing data is NOT a technicality, it's making a conclusion that is flawed when you specifically referenced a data set.

It is NOT a technicality when you say, "And based upon this data, it's clear that..." when the data doesn't support that argument in any significant way. The use of data as a strawman runs rampant through online boards, and in my view should be challenged more often versus less. It not only makes for better discussions, but forces people to THINK about how they are consuming that data.

VoA presented an analysis of data that I had no issue with, and as such I left the debate up to those who wished to participate. As long as VoA didn't make a specific claim related to the interpretation of the data, I sat back and watched the discussion.

While I may have my own views in terms of why a data set shapes itself in a particular way, I am not obligated to comment on every single theory presented by posters. Sometimes I do, sometimes I don't. On a lot of threads I prefer just to read.

YOU posted data, did not attribute the source, and then made an interpretation and conclusion based upon that data. I challenged your use of the data because I believe you misused it.

You're on a public forum. Expect your conclusions to be challenged.

 

So comparing stats between a healthy and unhealthy player is fair?


Alllllrighty then