How do you all feel about analytics?

[Wurider05]

Apparently we are using these now to rate players etc. How do you all feel about the process? Does it make a difference or is analytics just a fad? Could this be why a lot of B/C level players got paid major cash this summer?

[JTB]

Bunch of B.S.!!!

Like Chauncey Billups said on nba tv "if Reggie miller misses 3s all game or even if his shot has been off the last 5 games, he's not going to leave him open for a wide open game winning shot just because his shot has been off"

Teams aren't going to leave Korver open for a game winning shot just because he missed 10 3s in a row prior to that situation.

[Guest]

Bunch of B.S.!!!

Like Chauncey Billups said on nba tv "if Reggie miller misses 3s all game or even if his shot has been off the last 5 games, he's not going to leave him open for a wide open game winning shot just because his shot has been off"

Teams aren't going to leave Korver open for a game winning shot just because he missed 10 3s in a row prior to that situation.

Yep and like Charles said 'Per 36' - saying this guy would score this many point if given so many minutes - "He ain't playing that many minutes for a reason" lol.

I guess it has it uses to a point but making it a focus of setting lineups and who gets last shot is crazy.

[kerrypack]

i am glad that you asked.  let me send this post in before i pass out by being too drunk.

[kerrypack]

so anyways, I don't even remember the point of this topic, so let me get started talking.  Everyone should be quantified based on expected value.  Our metrics are likely the border of garbage.  the metrics would need to be adjusted based on team strategy.

[DJlaysitup]

"Advanced Analytics"....reminds me of an old Boss I had.  I would tell him what was going on and then he would "ponder" a little bit...and basically repeat it back to me with a couple word changes and he would flower it up a bit.   At that point I'm sure he felt he thought it up.

[Admin]

Yep and like Charles said 'Per 36' - saying this guy would score this many point if given so many minutes - "He ain't playing that many minutes for a reason" lol.

I guess it has it uses to a point but making it a focus of setting lineups and who gets last shot is crazy.

It's actually been proven that a players numbers tend to hold true when his minutes increase, far more than they fail to meet that level. Of course it gets less accurate the further from 36 you go, but say 12-14 minutes going to 30ish it's very accurate.

[turnermx]

Good analogy!

[Guest]

Lots of good insight here packed into jumbled thoughts. And probably right about the Hawks metrics.

I think I actually have a pretty good analogy for analytics and it has to do with BBQ (no coincidence that I'm smoking a butt overnight at the moment). If you want to do it right, you've got to have a good rub and enough experience to know what kind of fuel to use (charcoal, lump charcoal, wood - hickory, cherry, apple, etc.) plus how long to smoke it for. There are lots of good BBQers who have that down almost to a science and don't need any newfangled devices like digital thermometers that can measure the temperature of the grill and meat. Those devices, when correctly calibrated, are your data quantified more precisely. It's good information, although admittedly old school BBQers don't necessarily need to know the exact temperature because they have experience or some innate ability to get the understanding of where the grill and meat are. Old school BBQers don't need to say "oh the meat is about 150 degrees right now! since it has been smoking for 8 hours at roughly 225 degrees" as they likely have some other phrase they'd use to mean the same thing like "oh that meat has a nice give to it, it's firming up right now". Really, those two quotes could mean the same thing but are told in different language. At the end of the day, you want the best tasting meat no matter what the process is. If you're old school BBQ, then the digital thermometers (data, not analytics) are extraneous to you and is a waste. So why use them?

The way analytics should be thought of is by taking data and applying it making better BBQ for you. For a simple example, your data are your temperatures (of meat, grill, and outdoors), type and characteristics of meat, your grill, and type of fuel. One poor application of analytics in this example would be to go to a website and see that you should "smoke at 225 for 12 hours" and then do that for your BBQ. You are taking the too literally and not considering all factors. Lots of times this is what one calls a rule of thumb, it works *on average* but we aren't trying to say it will always work out. Sometimes you need to smoke for more or less time. Maybe you need to run the smoker a bit higher in temperature? Who knows, but if you want to correctly apply analytics then you go through and put your hard hat on. Through enough practice and critical thinking for analytics, you can eventually find out that for your particular grill you may need to run about 230 for 10 hours. There's still a fudge factor there because you cannot quantify everything, but you'll start to figure out that "hey, not all analytics applications will apply to *my* situation" and as soon as that light bulb clicks you eventually realize that you are applying a crude form a analytics once you make adjustments. Based on experience, you reform what you have done in the past and eventually you get better at this.

The whole preceding paragraph has, to some degree, been what old school BBQers have gone through. That's great! There's not just one way to learn how to BBQ. And if your only goal is to make the best damn BBQ ever, then pick whichever method is easiest. Do it and be done with your life. But you ever realize that lots of "old school" BBQers on BBQ Pitmasters (think of competition level BBQers) typically do not have well run BBQ businesses? Yeah, that is where the analytics REALLY starts to show. A reason, among many others, why the old school guys don't do well in business is because their method of cooking is too costly for the amount that they can sell their BBQ for. And so they do not appropriately value their product. This can be a problem if you want to be a successful business, but it isn't a problem if you only care about making *the best* regardless of price you wish to sell for or how much it costs you.

Now there are two types of old school BBQers. Those that HATE digital thermometers or other tools used to collect data and those that use them. Those that use digital thermometers may not get a huge benefit from them, maybe they never changed their methods? But at least they have more information.

old school BBQers = old school NBA players/executies (that typically hate stats)

digital thermometers = statistics (box score, +/-, sportVU tracking)

"smoke at 225 for 12 hours" = analytics from a website (Wins Produced, Win Shares, new ESPN metrics)

your particular grill you may need to run about 230 for 10 hours = in-house analytics for NBA teams that do it well

business = CBA which makes it so that teams cannot spend willy-nilly

These things parallel. And I did not touch on a big problem that exists which are f***ing morons running around and telling you "HEY IF YOU WANT TO SMOKE THE BEST BUTT THEN MAKE SURE YOU RUB IT WITH TALCOLM POWDER AND THEN WHEN YOU'RE SMOKING IT TAKE IT OFF THE SMOKER AFTER 6 HOURS STICK IT IN A COOLER FOR AN HOUR THEN PUT IT BACK ON". That right there is technically analytics in my example. And it is realllly bad for anyone who knows their ass from a whole in the ground. You'll encounter those types of people all over the place if you get into BBQ. You also encounter those people all the f***ing time when discussing basketball analytics. Don't let those bad examples misconstrue what analytics are. You have good people and bad people with analytics, but analytics itself is not good or bad. It's all about your application.

If you want to be a basketball fan or purist and say "aw to hell with analytics, I know how to do this shit!" then that is fine. Please don't run my business though because you are likely a poor evaluator of values and costs.

I prefer charcoal ole skoo BBQ ..$hit takes better seriously. They can find a billion new ways to cool meat but I like ole skoo stuff..just me

[DJlaysitup]

Concerning Wilt Chamberlain in the day:

 

Scout:  "He is a big F'er and he dominates."

 

Head Scout:  "Chamberlain is a real force, if we can get him he will improve our team immensely in multiple ways."

 

General Manager:  "Mr. Chamberlain demonstrates talents that are not found in even the top few NBA Athletes.  His physical presence on the court is impressive and his overall game changing abilities are second to none."

 

Owner:  "What do you think coach?"

 

Coach:  "He is a big F'er and he dominates."

[ATLSmith]

It's a good tool.  The Hawks should use it.

Edited October 18, 2014 by ATLSmith

[Wretch]

They have their merits.  The fallacy lies in leaning on advanced metrics to project rather than reflect...and I stress LEANING on.  As a supplement to "normal" player scouting and preparation for games, it's a good idea.  As a primary evaluation tool, no.  For odds makers and sports betters...sure.  For GM's and coaches trying to build contenders...?  Leaning on this sort of thing to predict what WILL happen or decide who's the "better" player is a mistake.

 

One of the things I really dislike about some of these stats is how there's no weight given to the quality of play.  You have to LOOK at the way a player is performing and the circumstances surrounding his stats....not just look at the stats.  I mean...you take per and you say Role Player X could theoretically score 17ppg if you'd give him 36 minutes.  Problem is...

 

Who is he playing against during his minutes?  Is he getting the bulk of his PT in garbage time?  And how is he getting these stats...?  A halfcourt Hail Mary, a garbage time wide open 3, a circus shot 3 off the glass, and a game winner over the outstretched arms of Scottie Pippen all have the same weight.

 

In short...good supplementary, bad primary evaluation tool.

[MaceCase]

It's actually been proven that a players numbers tend to hold true when his minutes increase, far more than they fail to meet that level. Of course it gets less accurate the further from 36 you go, but say 12-14 minutes going to 30ish it's very accurate.

That's because it's a good comparative tool better used to establish "rate" rather than predict production.  The Chuck way of thinking is just a lazy way of saying he doesn't believe that common sense can apply to math but as Fanatic has pointed out though, you'd never want Chuck as your GM. 

 

James Harden is my only argument on the topic.  A LOT of "old-school" heads saw nothing special in the guy, and especially on this board, ironically saw him as not being close to holding a candle to Joe.  Joe of course was the perfect example of a guy putting up stats in 40+ minutes whereas Harden was getting just over 30 minutes a night.  Now the per36 brought down Joe's production and brought up Harden's and I foolishly thought that maybe people would go "hmm perhaps this argument has merit as their rate on regular minutes are rather similar"

 

Nope, wrong.  The "old-school" heads circled their wagons and defended that Alamo all the way through an All Star selection and playoff appearance before they gave up the ghost. 

[BrazilianHawk]

I have to think the Millsap and DMC signings were all about the metrics. Therefore I consider analytics a great tool that should be used in all circumstances.

[benhillboy]

I like em, they don't bite. Being halfway educated, Math and Science holds a little weight to me. But a fool would rely heavily on them (Morey) as well as totally disregard them (Chuckster).

Coming from a franchise who has gotten more production out of marginal players than anyone by far, I'm sure Bud has that perfect mixture of knowing the what's, when's, why's, how's, and where's of the tools he uses when evaluating players or coming up with gameplans. They defended and utilized the corner three to perfection with f&@kin Bruce Bowen.

[benhillboy]

That's because it's a good comparative tool better used to establish "rate" rather than predict production. The Chuck way of thinking is just a lazy way of saying he doesn't believe that common sense can apply to math but as Fanatic has pointed out though, you'd never want Chuck as your GM.

James Harden is my only argument on the topic. A LOT of "old-school" heads saw nothing special in the guy, and especially on this board, ironically saw him as not being close to holding a candle to Joe. Joe of course was the perfect example of a guy putting up stats in 40+ minutes whereas Harden was getting just over 30 minutes a night. Now the per36 brought down Joe's production and brought up Harden's and I foolishly thought that maybe people would go "hmm perhaps this argument has merit as their rate on regular minutes are rather similar"

Nope, wrong. The "old-school" heads circled their wagons and defended that Alamo all the way through an All Star selection and playoff appearance before they gave up the ghost.

You know you finna get North cranked up.

The comparison was always funny to me, though. James' points per shot at any given time can be almost twice Joe's.

[Peoriabird]

I have to think the Millsap and DMC signings were all about the metrics. Therefore I consider analytics a great tool that should be used in all circumstances.

Thabo and Korver signing also

[Guest]

I have to think the Millsap and DMC signings were all about the metrics. Therefore I consider analytics a great tool that should be used in all circumstances.

Really? To hear some tell it, they were signed because we couldn't get any stars and no other team wanted them, they had no other offers and the Hawks overpaid for them, same with Thabo and Korver [green font]

[Admin]

Really? To hear some tell it, they were signed because we couldn't get any stars and no other team wanted them, they had no other offers and the Hawks overpaid for them, same with Thabo and Korver [green font]

 

You're so pretty! I love that you used [green font] instead of using the [ sarcasm ] tags [ /sarcasm ] around your text to make it sarcastic. 

 

And yes these signings are terrific. Jury is out on Baze but Thabo, DMC and Sap were all terrific signings and the trade for Korver has been enormous as well. 

[Atlantaholic]

There's a saying were I come from that states: Stats are like thongs, they show everything except what's essential.