[RAPM] How do the Rockets compare with other contenders?

Can't really use RAPM for players with low minutes played (lack of minutes played is the weakness, which is why it's quite useful for measuring most good players, as they don't usually lack minutes played). In effect, lack of minutes = lack of data sets = lack of "regularizing".

For players like Andersen, Jones, Dmo it may as well just be thrown out.

 

I have to agree with you on this one.

All that matters are the results which can factor in RAPM, coaching and even luck

 

Even if you indulge in stats, you will certainly always be one step behind Morey.

http://www.*****.net/huqs-pen-im-bored-omri-casspi-pushed-one-sophomores-rotation/13245/

 

Well, you are certainly entitled to your opinion that win-loss is the truth. But even if you disagree with the OP. There's good reason why he uses this particular stat. There was once someone in 2008 who was intently analyzing players and teams using a primitive form of RAPM with several other analytical people. At that time, simply adjusted +/-. You can find his blog here. http://www.countthebasket.com/blog/

He stopped his blog in 2008. Because at that time, Daryl Morey hired him to continue his work. Today, he is the VP of Basketball Operations for the Houston Rockets.

So if you believe the Rockets are a smart organization, wouldn't it be nice to at least get a glimpse of how they think?

 

Thanks for this thread CXbby, and for the links various people have posted.

 

I think there a bunch of NBA coaches, scouts, and front office folks, including Morey, that disagree.

 

Suppose Morey watches a Player A in a game. He puts up 20 points, 10 rebounds, and look like future Shaq.

Now suppose Morey gets his scouts to watch Player A in 20 other games, chart his possessions, study his films. And finally these scouts give Morey a scouting report that says the player is inconsistent, has trouble dealing with bigger players, slacks off on defense too much, and overall grade of 2nd round pick.

Now do you think Morey will trust his own eyes? Or the combined views of all his underlings who provide him with spreadsheets of data?

 

I think he combines the stats from when he was watching and those from when the scouts were watching to have a larger sample size.

 

Most of the data from your hypothetical is still based on eye ball tests. Casspi is an example of a guy that they used eye ball test to determine he could be a PF.

 

You do realize that stats are based on eyeball tests, right? Do you know how box scores come into existence? It's because there's a person at every game who watches and chart the games with his own eyes. Stats like points and rebounds are manifestations what what some old basketball fan determine is worth watching and what is not worth watching. Someone deemed the task of rebounding the ball is worth watching, so he keeps track of rebounds. The same person deems letting the other player blow by you for a layup is not worth mentioning, so there's no "# of times getting owned by opposing PG" stat.

In fact, the new SportsVU is basically an eye that sees EVERYTHING. And then translate them into stats. Stats come into existence because of what we see. Not some magical, random numbers that appear out of thin air on your computer independent of the game.

 

The advantage of SportsVU being that the data allows you to mine for stats you may not have originally looked for, as once played the game is there forever.

 

Sure if the numbers were representative of their current play/worth. You can't accurately predict 2nd year performance based on small samples in a rookie season.

 

Really, the OP and some others believe RAPM is more relevant than Wins vs Losses SMH. Nothing is more relevant than wins and losses, period.

Play-off spots are not awarded to teams based on RAPM, the NBA still uses wins and losses regardless how antiquated they may be.

Teams do not advance in the play-offs based on RAPM the NBA still uses Wins and Losses.

The NBA does not award the team with the best RAPM the NBA Championship trophy, they give that to the team that wins the most games in the Championship series.

Wins and Losses are the most important stats by far.

As I said earlier, stats are just a tool to be used in conjunction with actually watching the game. Like all tools if you use them correctly they increase productivity. If you use them wrong, they are counter productive.

 

I don't understand what exactly you're arguing here. "The only stat that really matters is wins and losses" is just confusing analysis/forecasting with the ultimate team goal of winning games.

If the Rockets lose their first game, the manner in which they lost is more relevant for both improving the team and forecasting how they may perform in subsequent games than simply the fact that they lost.

 

I guess my point is (and I'm doing a poor job of communicating it) that a lot of stats (WARP, RAPM) are misleading and some times give a conclusion that is just crazy. At the end of the day the only stat that really counts is wins and losses.

 

This is just a very narrow minded (and lazy) way of thinking that I often hear when someone faces a stat whose interpretation falls out of their line of thinking. Don't like what a particlular stat suggests about a player they like? Just dismiss it out of hand by saying that stats don't matter.

Wins and losses are just another statistic that helps us understand what is happening on the hard court below us. As in any other statistic, it is prone to other confounding factors and should always be explained in context. Because wins and losses have their confounding factors, we have to look at other statistics to help us understand certain things.

For example, I can't really look at a wins/loss stat to tell me if Dwight Howard is a more impactful player than James Harden. I have to look at different statistics. RAPM is one of the statistics that try to explain that.

 

repped.

 

So Robert Horry is better than Hakeem because he won more championships?

 

If that's what you think your a fool. Try watching the game before trying to make a claim like that.