"Analytics Tells Us What To Do In the 2nd Half"
- Thread starter Farsight Clone
- Start date
Huge defender of year 1.
Not good enough defensively to be a title contender.
Surpassed expectations for year one given the context, but I don’t see how they were going to be top 25 in defense even healthy.
No idea where Pope’s coaching has went though. Year 1 Pope beasts Year 2 Pope
He's looking at that **** upside down or something.
I’m not sure they even know how to use the analytics they have properly to be honest. They certainly don’t coach the team to what analytics say from other people. We don’t have their data but we can infer some things just by the teams we play.
"Gooooo, Runnnnnnn" really wasn't much of a game plan, to be honest!
That team was never in the conversion for final four.
True. Analytics can be a good tool. It’s not coaching. But you have to use tools you have properly. You don’t drive a nail with a screwdriver.
And you certainly can’t use Ken Pom analytics. It’s a pace metric.
He doesn’t count offensive rebounds as a possession.
For example; a team comes down and misses and gets the rebound and shoots again and misses and gets the rebound and then makes the shot. To Ken Pom that’s 1 possession for 2 or 3 points.
That looks good on paper. That’s 2 or 3 points for one possession. But it was 3 possessions. Players were credited with 2 rebounds and 3 shots.
One could argue that since he’s doing it for both teams it balances out in the big picture. And it probably does?
Just depends on what you are using the numbers for.
The bottom line though is that the numbers are skewed.
maybe but we saw the healthy version of that team them beat Duke and Gonzaga
Analytics are just another tool in a coach's toolbox, but if the machine is broken beyond fixing then no tool will help. This team is very poorly constructed both position and chemistry wise. They can say they love each other till the moon falls, but body language says otherwise.
Going forward, Pope is either going to have to go with all defense or all offense. I think we saw the all offense approach Saturday, so I guess we try the rock fight approach now.
Will add, I don't like analytics. I know they can help but they are easily skewed. A good coach uses his "eye" to do his job. At this point I think Pope might be proverbially blind.
Going forward, Pope is either going to have to go with all defense or all offense. I think we saw the all offense approach Saturday, so I guess we try the rock fight approach now.
Will add, I don't like analytics. I know they can help but they are easily skewed. A good coach uses his "eye" to do his job. At this point I think Pope might be proverbially blind.
On the call in show tonight Pope said he won't start Lowe because he wants to nurse his injury. I am sure analytics told him that too =\
Someone needs to talk to Coach Pope about External Validity or Generalizability. That is when the analytics data may not necessarily apply to the team in front of you at the moment.
If most of your analytics are gathered against mediocre teams, you can't blindly apply the data to the better team you are playing...similar to in medicine. You don't treat the elite athlete the same way you treat the geriatric patient with the same diagnosis.
Analytics must be stratified by opponent tier, not generalized across the entire spectrum of opponent quality. In medicine, you inform your clinical decisions with data, and the art of medicine is applying the data correctly to the correct patient. I think coaching is similar, and at the moment, I believe Pope is misapplying the data.
Anyone else have thoughts about this?
I have thoughts about it — absolutely. Suppose Pope did find that trends in data indicated productivity decline occurs after a certain volume of minutes of consecutive play. It isn’t necessarily correct to conclude, “Oh, this means I can only play every player X amount of minutes.” This would be an over generalization and over simplification of the narrative that data offers. You’d have to consider the effect this would have on decreasing rhythm during potential “hot streaks.” You’d also have to factor in that production from an athlete here at a 75% efficiency may still be better than 100% productive efficiency of the replacement player. Those are just tiny factors to consider — I think you are correct for sure.
.
I am very skeptical of most coaches' ability to interpret the data they are given. Anyone can run some data through R or some other program and create some statistics, but do you actually know what those statistics are actually capable of telling you? Let's face it, most of the population will hear that there's a 60 percent chance of rain then complain about a bad forecast when it doesn't rain. Can't imagine that most coaches have much more experience with data interpretation than the average person.
Analytics are fine for finding long term trends that may not be obvious in real time or finding something that goes against long held beliefs (sacrifice bunting in baseball), but relying too heavily on analytics is not coaching. Have that data in your arsenal, but coach based on what you see in real-time.
Good post.
I am very skeptical of most coaches' ability to interpret the data they are given. Anyone can run some data through R or some other program and create some statistics, but do you actually know what those statistics are actually capable of telling you? Let's face it, most of the population will hear that there's a 60 percent chance of rain then complain about a bad forecast when it doesn't rain. Can't imagine that most coaches have much more experience with data interpretation than the average person.
Analytics are fine for finding long term trends that may not be obvious in real time or finding something that goes against long held beliefs (sacrifice bunting in baseball), but relying too heavily on analytics is not coaching. Have that data in your arsenal, but coach based on what you see in real-time.
I agree with you 100%. That’s the danger of over-trusting analytics. Most numbers are averages with a lot of uncertainty around them (confidence interval), and that uncertainty grows when opponent quality and game context change. When the model becomes a rule instead of a guide, I believe coaching judgment gets weaker, not stronger.
Cal and Pope are complete polar opposites.
Cal: I don't give a **** what they do, we are sticking to our plan come hell or high water and we aren't changing!
Pope: We have no plan, or any ideas of what to do. But once we are down 20 points we can figure it out!
That’s good for ya. Builds character