I don't want KemPom to explain how he calculates it. I want Twins to break it down for me then answer my followup questions and explain how his formula factors in those things to make it so accurate.
Tracking Iowa's RPI, KenPom, & Sagarin
- Thread starter NikeHawk21
- Start date
I don't want KemPom to explain how he calculates it. I want Twins to break it down for me then answer my followup questions and explain how his formula factors in those things to make it so accurate.
For instance.
How does his formula calculate the randomness of when opponents catch fire and start making everything they throw up?
How does his formula calculate when opponents go ice cold and miss open shot after open shot.
Does his formula give half credit for playing good defense but the player scores anyway?
Does the formula factor in a difference of a team giving up wide open shots early in the shot clock vs making the other team work for a good shot?
Is giving up a wide open dunk the same as forcing a fade away deep 2?
Those are off the top of my head. I'll wait for Twins genius response then think of more.
How does his formula calculate the randomness of when opponents catch fire and start making everything they throw up?
How does his formula calculate when opponents go ice cold and miss open shot after open shot.
Does his formula give half credit for playing good defense but the player scores anyway?
Does the formula factor in a difference of a team giving up wide open shots early in the shot clock vs making the other team work for a good shot?
Is giving up a wide open dunk the same as forcing a fade away deep 2?
Those are off the top of my head. I'll wait for Twins genius response then think of more.
Robowe
Well-Known Member
But for every lucky contested shot someone makes over the course of the season an unlucky open shot will miss which makes it not flawed. It's called the law of averages.
Robowe
Well-Known Member
That's the thing with advanced stats. You see it in baseball all of the time. Not all hits are the same (swinging bunt, double off the wall). So, there are stats that try to take that randomness (luck) out of the calculation. Stats like WAR (Wins Above Replacement), FIP (Fielding Independent Pitching), Weighted Runs Created, BABIP (Batting Average of Balls in Play) try to break down these things while taking the randomness out.
However, there is still interpretation needed to be done to calculate things like defensive zone rating in baseball. Or, more importantly, some things that are not factored in (for example, when a fielder things a curveball is being thrown, so they take a step one way, and the ball is hit one step the other way - who's fault is that?).
Here's a link to a long explanation on baseball advanced stats. It will make you dizzy.
http://www.cbssports.com/mlb/news/the-eye-on-baseball-fans-guide-to-advanced-baseball-statistics/
Here's my favorite part when talking defense (which I think can apply to basketball as well). Specifically, variation and shifts (which in basketball is the same as zone vs man to man) IMO.
To be sure, though, WAR does have weaknesses. Specifically, the defensive component of WAR -- Ultimate Zone Rating (UZR) for fWAR (more on UZR below) and Defensive Runs Saved (DRS) for bWAR -- aren't as definitive as we would like. Each is among the very best methods we have for assessing defensive value, but they're far from perfect. First, single-season defensive data is especially prone to random variation, and second, the publicly available defensive data may not fully correct for things like extreme defensive shifts (which are very much a part of the game these days).
Laws of averages is extremely flawed. Unless you think if you get 3 bad breaks you will always get exactly 3 good breaks. If he has a perfect formula (he might for all you or I or anyone else who isn't named Will Hunting knows) than its accurate if the law of averages is also accurate, which it's not. It still makes it as good of a formula as it could possibly be. But it also still makes it flawed.
Why don't you tell me where Iowa was after the Omaha game? If it's considerably better now (which I assume it is) then that will say good things about the stat. If it's the same, it won't bode well for it.
I have no idea where we were after the Omaha game I don't remember. Just remember not to bring up points per game because that's flawed
I like it. So if a single season is way less accurate than say 5 seasons, what do you think half a season is? Twins has been quoting these stats as indisputable since 5 games into the season.
Robowe
Well-Known Member
But, that's where it's not necessarily true. If Bohannon has his hand up closing out on a guy it's different than Pemsl closing out right? Isn't it easier to shoot over the smaller player (if both have their hands up?) So, does the 6 foot 1 guy get dinged on that? That's not the law of averages.
Remember in the Michigan game when JBo made a two pointer, but then Michigan hit a 3 (in the corner where JBo was trying to close out)? You even mentioned in the game thread that he scored 2, but then gave up 3. Was the corner guy his responsibility? Were they in zone or man on that possession? Didn't he try to close out (I believe he did, but he was late)? Can't you see where all of these factors come into play?
I think this is all PC is trying to point out. And, in most cases (as Dean has pointed out many times), defensive stats in basketball are flawed. You can get a steal or a block when you are out of position (or beaten badly) if the offensive player makes a bad play (bad pass/shot). Should the defense get credit, or the offense lose credit?
So that's not something you can look up? I've seen you respond to my posts seconds latter with stats you've looked up. You can't find KenPom rankings from previous weeks huh?
I don't pay for KenPom so no I can't look up specific dates that I know of. Just now and previous year totals
My theory was the offense can still score against good defense. Good defense isn't necessarily decided by whether a team scores or not. Just like bad defense isn't decided by whether a team scores or not. An example would be Michigan. They missed a lot of 3s even tho our perimeter defense was bad.
I don't think advanced team defensive statistics are any more flawed than other stats out there
So ppg is flawed then too?
Yep. Not all steals are created equall. Not all assists are created equal. You have to physically see the play play out to judge how much credit or blame a player or a team deserves. That's why people always say stats don't tell the whole story.
They certainly can be. Cook's points were off last game because of a bogus goaltend call. If a guy banked in 3 3s and made a half courter at the buzzer, do his points show how good of a shooting night he had.
Well better not reference ppg moving forward then
They are as good as they possibly can be (for all we know).
Definitely not in blanket statements.
