Analytics

Lets talk some Hoops!
User avatar
Larryl9797
Site Admin
Site Admin
Posts: 1031
Joined: Tue Sep 12, 2006 3:24 pm
Location: Central Connecticut
Contact:

Re: Analytics

Post by Larryl9797 » Mon Oct 29, 2018 11:17 am

Capture.JPG
No back to our regular scheduled basketball talk. DM is correct he has so much hoops knowledge that he really does not need highly processed analytical mining. My guess any analytical flags don't really get looked at cuz of the feel and knowledge DM has for the game. ( Which is TONS and TONS)

User avatar
BigBlue92
CCSU Fans Regular
CCSU Fans Regular
Posts: 286
Joined: Thu Jan 21, 2010 8:32 pm

Re: Analytics

Post by BigBlue92 » Mon Oct 29, 2018 12:02 pm

Larryl9797 wrote:
Mon Oct 29, 2018 11:17 am
DM is correct he has so much hoops knowledge that he really does not need highly processed analytical mining. My guess any analytical flags don't really get looked at cuz of the feel and knowledge DM has for the game. ( Which is TONS and TONS)
Larry, I'm not trying to be a wise-ass but was your comment meant to be a joke or taken seriously? I honestly wasn't sure if it was intended as sarcasm.
At the NCAA level I would certainly hope that ALL coaches have a tremendous knowledge of the game which is precisely why virtually all programs are trying to get every little advantage possible through the use of various metrics that have been proven highly successful. I don't think anybody in the country would suggest that analytics should drive every decision that's made. That would be ludicrous. But equally ludicrous would be to NOT take full advantage of this tool. I have friends who coach at the D-3 level and I can tell you that analytics have become a huge part of their game-planning, assessment, scouting, and recruiting. It gives coaches a whole new lens through which to view a player's and a team's performance (including strengths and weaknesses). It also serves as a motivational tool that coaches are sharing with their players.

And perhaps the greatest proof that metrics are an indispensable tool is that Vegas has bought in 100%. As with anything, follow the money. If analytics were not an accurate predictor of performance then bettors wouldn't be salivating to get their hands on them.

User avatar
Larryl9797
Site Admin
Site Admin
Posts: 1031
Joined: Tue Sep 12, 2006 3:24 pm
Location: Central Connecticut
Contact:

Re: Analytics

Post by Larryl9797 » Mon Oct 29, 2018 1:25 pm

I was being serious.. I am sure that he uses Analytics to some extent but probably uses his player/coach senses a whole bunch more. Personally. Someone who has a high value of Analytics probably wont say this "Personally, I hate analytics. Maybe because I played, it drives me crazy."

I am not saying that stats and analysis is not a baseless science. I'm just saying that If DM had to choose between a Statistician or new Asst Coach, I know what DM would choose.

CCSU4-3
CCSU Fans Know it all
CCSU Fans Know it all
Posts: 792
Joined: Mon Mar 07, 2016 11:55 am

Re: Analytics

Post by CCSU4-3 » Mon Oct 29, 2018 4:46 pm

ccsuhoops wrote:
Sun Oct 28, 2018 10:52 am
Yeah, they are not gonna lose, but Boone is another over analytical manager and look where it got them. Bottom line is you still have to play the games and score more than the other team to win. I agree with above, analytics should be looked at but not used to make every single game time decision. Games still need to be played and won and numbers on a piece of paper won’t do that. It’s not PS4 it’s sports.
It got the Yankees over 100 wins. That's not too bad in my book.

Analytics aren't about results. Its about improving the chances of a positive outcome by percentage points.

If, based on evidence, I told you a play has a 54% chance of scoring 2 points, versus another play that has a 48% chance, which one would you prefer. No matter what, there is at least a 46% chance they do not score.

Wouldn't it be helpful to know that (hypothetical numbers) Tyler Kohl shoots 38% going to his left and 55% going to his right? That's analytics. Just like knowing that 58% of Ian Krishnan's three point attempts miss long instead of short.

These are things that analysis tells you. It informs your decisions. I do not understand why a coaching staff wouldn't want this information.

CCSU4-3
CCSU Fans Know it all
CCSU Fans Know it all
Posts: 792
Joined: Mon Mar 07, 2016 11:55 am

Re: Analytics

Post by CCSU4-3 » Mon Oct 29, 2018 4:54 pm

Larryl9797 wrote:
Mon Oct 29, 2018 1:25 pm
I was being serious.. I am sure that he uses Analytics to some extent but probably uses his player/coach senses a whole bunch more. Personally. Someone who has a high value of Analytics probably wont say this "Personally, I hate analytics. Maybe because I played, it drives me crazy."

I am not saying that stats and analysis is not a baseless science. I'm just saying that If DM had to choose between a Statistician or new Asst Coach, I know what DM would choose.
Steve Kerr had a long playing career and is one of the most progressive coaches in the NBA when it comes to analytics.

Interesting that analytics are good enough for Kerr and Greg Popovich and Brad Stevens, but not for Donyell Marshall.

Shane Battier was a player...now he's the Director of Analytics for the Miami Heat. All this "he played the game so he doesn't need analytics" is garbage. Just like in baseball..there are hundreds of current and former major leaguers who are very into this stuff. But somehow John Smoltz becomes the representation of "The Players".

User avatar
ccsuhoops
CCSU Fans Know it all
CCSU Fans Know it all
Posts: 3241
Joined: Wed Oct 17, 2007 12:10 pm

Re: Analytics

Post by ccsuhoops » Mon Oct 29, 2018 5:16 pm

CCSU4-3 wrote:
Mon Oct 29, 2018 4:46 pm
ccsuhoops wrote:
Sun Oct 28, 2018 10:52 am
Yeah, they are not gonna lose, but Boone is another over analytical manager and look where it got them. Bottom line is you still have to play the games and score more than the other team to win. I agree with above, analytics should be looked at but not used to make every single game time decision. Games still need to be played and won and numbers on a piece of paper won’t do that. It’s not PS4 it’s sports.
It got the Yankees over 100 wins. That's not too bad in my book.
True but how many times did he take out pitchers based on that info, only to get burned? Quite a few, they talk about it during the broadcasts and its all over social media, especially when he was wrong. :lol:

Bottom line, I get the need and I hope that CCSU does use them, but from that quote, I bet they don't. I've got my revisions about DM but I'm gonna hold them till after next season to see what he can do. It's taking all I got after watching them not cream a d3 team but I'll be good. Hey, at least they beat this d3 team by more points than the last d3 team they played last year. #-o

User avatar
BigBlue92
CCSU Fans Regular
CCSU Fans Regular
Posts: 286
Joined: Thu Jan 21, 2010 8:32 pm

Re: Analytics

Post by BigBlue92 » Mon Oct 29, 2018 5:41 pm

To put this in other terms, take a look at horseracing--a sport that many would consider a bit "backwards". There was time when all handicapping, breeding, and evaluation was based strictly on the clocker's times and whether or not a horse was visually impressive. Then a man came along who revolutionized the sport. His name was Andrew Beyer--a Harvard dropout who was a math whiz. He developed a metric for evaluating a horse's performance and placing a numerical value on it. It became known as the Beyer Figure. That was 40 years ago. The Racing Form began publishing those figures in their past performances. Then Brisnet came along and took horseracing analytics to new heights. Today if you go to Saratoga or any track you will see EVERY horse-player, trainer, owner, and fan walking around with the DRF and/or Brisnet PPs. EVERYTHING in horseracing is analytics. They're used to evaluate talent, determine trends, look for patterns, notice speed, determine pace, etc, etc. Will it allow you to pick a winner every time? Of course not. But understanding how to decipher the analytics of the sport will certainly give you a leg up over the guy who just uses the "eye test". (Attached is a sample Brisnet PP for ONE horse. ALL analytics.)
Attachments
708BCDD7-8ADE-42D2-B43A-F0A13DD1FB82.jpeg

RonnyRob15
CCSU Fans Know it all
CCSU Fans Know it all
Posts: 509
Joined: Sat Mar 10, 2007 9:35 am

Re: Analytics

Post by RonnyRob15 » Mon Oct 29, 2018 7:27 pm

You guys kill me.

The Dodgers have the biggest analytics team in the sport and the Red Sox were trend setters owned by a billionaire who made his money on stock market algorthims. The numbers are so dumb the Red Sox have won 4 world series in 15 years. They probably would have 6 if they stopped listening to the nerds.

If you don’t participate in analytics in professional sports you are the Baltimore Orioles. Running around with old school guys getting your arses handed to you every night.

This is pretty disappointing from Donyell but nobody stays here long enough to have a decent sample size and Central probably can’t afford the resource or even the software.

Let’s all just enjoy the parachute show at the half.

User avatar
ccsuhoops
CCSU Fans Know it all
CCSU Fans Know it all
Posts: 3241
Joined: Wed Oct 17, 2007 12:10 pm

Re: Analytics

Post by ccsuhoops » Tue Nov 13, 2018 10:22 am

Ok, more questions about the numbers, what the heck is t-rank, why does it differ so greatly from Ken Pom and why is it better than good ol rpi?

Thanks in advance :-)

CCSU4-3
CCSU Fans Know it all
CCSU Fans Know it all
Posts: 792
Joined: Mon Mar 07, 2016 11:55 am

Re: Analytics

Post by CCSU4-3 » Tue Nov 13, 2018 11:36 am

ccsuhoops wrote:
Tue Nov 13, 2018 10:22 am
Ok, more questions about the numbers, what the heck is t-rank, why does it differ so greatly from Ken Pom and why is it better than good ol rpi?

Thanks in advance :-)
T-Rankings.com is just another projection system. Bart Torvik is the creator and I use Barttorvik.com for a lot of my research.

RPI has a formula that is strictly records based, which was great in the 80s but we have far better data now. Both Kenpom and T-Rank (and BPI and others) used more advanced metrics which are much more predictive.

To use a baseball analogy: Player A goes 3 for 4 with 3 weak seeing eye singles. Player B goes 0 for 4 with four line drive rockets that outfielders made diving catches on. RPI would look at 3 for 4 and assume that player is more likely to get a hit tomorrow. The projection systems would look at exit velocity and the Statcast measurements of how good the catches were, to determine that Player B is more likely to get a hit.

No one uses RPI anymore, not even the NCAA selection committee. It's dead.

Post Reply

Who is online

Users browsing this forum: No registered users and 3 guests