SVG questions the analytic side of the game

Morey faints.

 

Duh, of course you can't rely on numbers and analytics alone. The idea is to supplement traditional scouting methods with advanced statistics.

That said, SVG raises a solid point about how Synergy and some of the other resources available calculate their numbers. There are some holes.

 

That's a bit misleading. He's not questioning the usefulness of analytics, he's questioning the method to how a particular data is obtained, which is a valid and different discussion.

 

Stats are only as good as the people recording them.

 

Exactly. Garbage in, garbage out.

 

Stan would work well with Houstons stats teams, dozen+ guys that could sync everything to his thinking in order to make sense, Jeff has repeatedly said Daryl is good at providing stuff that makes sense to him.

His final complaint seems more about things like the memphis disaster.

 

It's a legitimate complaint, and its the same issue faced in the corporate world and everywhere else they use data analysis. If the data isn't good, the analysis will be misleading and you can make bad decisions. You can't just hire some kids at $9/hour to watch games and categorize plays; you need people who really know basketball. But, that problem is not insurmountable. Hire good people and establish a good process and you can get better data.

 

SVG is not averse to analytics. He likes to say provoking things, so he's putting the "anti-stats" hat on at this conference. The fact is he used advanced stats quite a lot as coach of the Magic, when many teams weren't using it.

He's absolutely correct that there needs to be agreement on basic concepts between the numbers analysts and the coaches.

 

I am neither agreeing or disagreeing with SVG but weren't those the same statements made for the original moneyball in baseball?

 

Basketball is a sport that is probably to complex for complete analysis. Even if you had a group of experts go through the tape of NBA all of the NBA games and completely categorize the plays, another equally expert group's analysis of the same data set would probably profoundly differ from the conclusions of the first group.

I guess what I am trying to say is that you go to war with the military you have. We will never have perfect metrics to use when analyzing basketball. We will never know when the metrics are even "good enough".

 

The example play he gave would absolutely be considered an isolation by Synergy's classification. Not to say that Synergy's methodology doesn't have its limitations, but the writer who used the stats in their article that Van Gundy is pulling his example from also may not have the basketball knowledge to properly analyze the numbers they pull from Synergy. So when Van Gundy thinks "isolation" he may think perimeter iso + post-up iso, whereas a writer just writing down the 17% number from Synergy is just getting perimter isolations.

The teams who use Synergy are aware of these distinctions, and of course, there's video associated with every play directly through Synergy.

 

So does this mean SVG now has 0% chance of becoming our next head coach?

 

On the contrary, I doubt Morey would disagree with SVG on this at all -- data integrity is vital to good analysis. Of course Morey would want the data to be speaking the same language as the coaches.

 

SVG's comments remind me of when I criticize Lin, however lightly, on the BBS.

It goes like this:

"Jeremy's defense was kinda weak tonight."

Lin's fans' responses: "No it wasn't! He was the only bright spot out there!"

Me: "Look, I was at the game, I saw with my own eyes, Isiah Thomas' footprints are all up and down Lin's back."

Lin's fans: "No way! Where's the stats?" (they send Excel spreadsheet) "Look at these stats! Lin played better than the rest!"

Me: (getting mad) "Look, stats are great, but I saw with my own eyes, he looked like a fool out there tonight."

Lin's fans: "You're the fool. We give you stats and you just say 'oh, I actually watched the game without rose-colored glasses'."


Yeah, I'm not surprised about the one comment: "Oh, I'm just getting statistical info, I'm not really watching..."

I believe stats can help you get the most out of your players / the draft, but I think overreliance on them can leave giant gray areas, too: the mid-range game, defense, etc.

I do think any coach for this team would have to buy into the system (watch the scene from "Moneyball" when Beane argues with Howe in the dugout about Hattie: "His fielding does not matter"; we wouldn't want that different-philosophy situation here) OR be a future Hall of Fame-type coach. Anything in between? I have my doubts.

 

LOL. Most replies I have seen have not been quite that civil.