It was your word, not mine. Skewed data in small samples showing no real driving logic is also called coincidence. It doesn't matter if Brewer only started 3 of those 5, you actually made it easier for me. This shows even better how Skewed data in small sample causes misinterpretation and confirmation bias like you display. Now all I have to do is show the Philly game as the significant lowest game outlier. I didn't think to look, because I thought the data already showed massive skewing in the middle games. Dude ... you take out the Philly game as the severe low, and the Utah game as the severe high (with an inexplicable low Pace), then you can't look at those numbers without reasonably questioning whether or not there is statistical driver here at all Taking out the lowest and highest is very common in science experiments
You shouldn't make responses like this if you're not following the discussion closely. heypartner was the one that decided to nitpick one of my evidences 'as if' somehow it was my central point (that Bev has the worst DRtg, when it was only one of many data points I provided), ignoring the larger, more crucial theme - that Bev's return has coincided with our tanking in the defensive standings. And if we are going to attribute individual ORtg or DRtg to certain players, it makes much more sense to do it in the right context, hence the Ben Wallace or Tony Allen example. If Ben Wallace plays with Kobe and has both great ORtg and DRtg, you don't then go on to claim that Ben Wallace posts high ORtg because of his offensive prowess when he can barely average 5 points per game. That is obviously due to playing with one of the best offensive players in Kobe. But you can draw much more straight-forward connection between Ben Wallace's defense and his DRtg numbers, because defense is what he is known for and is his area of specialty. You could make an argument as to why individual rating is not the most precise stat out there, and I agree, but to suggest that just because players like Beverley could have high ORtg, we also shouldn't trust his DRtg is just wrong and logically doesn't hold up.
Why are you going back and forth to sites. You already said they are much different. NBA.com has everything you need. Do I also need to teach you to use a data query? Well, since I am showing 100% numbers from NBA.com queries, the onus is on you show any errors. There could be minor differences, but I double-checked. Otherwise, your memory isn't as good as my recent queries at nba.com You just made my point about using individual DRtg, once again. But doesn't matter. Now that I found the Philly game, we can do away with that adjustment. We just simply toss out the significant low game and significant high game out of the 18 and run the numbers on 16 games. Spoiler: further explanation of Utah game adjustment I didn't apply DRtg. I normalized all lineups involving the starters and EGo to use the actual starting lineup's points per minute, which wasn't anything special, it just wasn't horribly bad like the other Harden lineup. It wasn't about Bev...it was the starting lineup vs the Harden, EGo, Ariza, Capela and Ryno/Dekker lineup. That lineup gave up a significant high points per minute. I just adjusted them down to what the starting lineup did, actually what the Rockets did for the entire rest of game pretty much. It was a 13 pt difference.
Ha ha, I see that you're getting desperate. Why don't you break it down quarter-by-quarter while you're at it? Was the lighting in the stadium perfect during which we played 2Q in the Philly's game? How about the humidity? Noise from the crowd? I can also play this petty, 'find-an-excuse' game, you know. Just stop trying to move the goal post and admit your defeat. FACT : We played better defense without Beverley in the starting line-up, regardless of whether it was Brew or Gordon (103.0 and 105.8 DRtg) FACT : Our defense got worse with Bev's return (107.7 DRtg in his first six games as a starter, 111.2 if we include the last Utah game)
yes, I already did recalculation based on nba.com data which thoroughly debunks your delusional argument. FACT : We played better defense without Beverley in the starting line-up, regardless of whether it was Brew or Gordon (103.0 and 105.8 DRtg) FACT : Our defense got worse with Bev's return (107.7 DRtg in his first six games as a starter, 111.2 if we include the last Utah game)
Why don't you also throw out some quarters where we gave away too much or too little points? isn't that also an 'outlier?' LMAO you gotta be kidding me.
Nice spin job there, but the simple fact boils down to this : Pre-Beverley return (11 games) : 105.5 DRtg Post-Beverley return (7 games) : 111.2 DRtg simple as that my friend
Sorry, if you don't understand how it is common practice to throw out two huge outliers (a huge high and a huge low), then you don't know how to do science and statistics. Rockets are a 107.4 team when you toss out the Philly and Utah game Rockets are a 107.5 team when you include those two games. Rockets are a 107.2 team when either EGo or Brewer start (when not counting the Philly game) Rockets are a 107.7 team in Bev's games (when not counting the Utah game) That is about as statistically aesthetic as it gets. Just because you went looking for confirmation bias of Bev screwing up and posted it first, doesn't mean you get to say this very prudent statistical method shouldn't be used. When a very simple statistical method is employed to eliminate the noise of hi/lo outliers is so effective, this becomes the model that must be disproved. Your idea that one person is the cause of this becomes just an interesting hypothesis but not what you must disprove first. That's just the way it works. Are we merely Regressing to the Mean Our 105.4 running total after the Philly game (an 86) was artificially low, and it regressed to the mean (107.5) when we ran into an abnormally high game (a 132). It's that simple. And of course, we can enhance that method with Strength of Schedule adjustments like Hollinger uses for rankings. We check the opponent's ORtg vs the League. Are we keeping the league at their average, for instance? GSW is #1 in ORtg, TOR is #2, POR is #7 and Utah is #9. That will be 6 out of the 8 games since Bev came back. So, don't be surprised if GSW doesn't knock us up a notch again, just by doing their average thing. You're still having trouble using nba.com. The 7 games with Bev is 110.9. Hit refresh or something. http://stats.nba.com/teams/advanced/#!?CF=TEAM_NAME*E*hou&Season=2016-17&SeasonType=Regular Season&DateFrom=11/17/2016&DateTo=11/29/2016
No, that's not how it works because you're drawing the line arbitarily. As my previous sarcastic comment suggests, if you're going to throw out all the 'outliers', then also throw out outlier quarters, or outlier five-minute stretches during a quarter. Why only throw out outlier games? That's just an arbitrary rule that you made up. Also, you don't know whether 86 or 132 is a huge outlier or not since the season is not over yet. What if we have 70 and 145 game in the future? Then it makes no sense to throw them out since they are not the extremes anymore. Your argument would have some value if the sample size is set, but since that's not the case you shouldn't throw them out, because we don't know what a true outlier is. For instance, If we didn't have the 86 and 132 game, then 95 and 123 game would have seemed like outliers, no? If you still don't understand this point, then you're just very obtuse or ignorant about how real statistical analysis works. Yeah, I did my own calculation, huge difference. The point still stands. Pre-Bev DRtg >>>> (meaning better) Post-Bev DRtg
Let's revisit in 10-20 games time. Bev is overrated, but he is not the reason for our defensive struggles.
Agreed. The Rockets have had quite a BUZZ SAW of a schedule playing against some quality teams on the road...more games on the road then at home..new coaches..injuries..new systems yada yada...Utah may never shoot like that again this year against the Rox, Lakers fluke loss, OKC definitely a fluke loss..lost by Cavs by 6 points...all these games played on the road.
@heypartner Why are you trying to argue with a troll?! He is not interested in discussion, he is not going to be convinced by reasonable discussion. He clearly has a conclusion, and has a confirmation bias. Just put him on ignore, posters like this are not worth talking too.
Bev is a startable pg next to James Harden. He would have been reduced to a backup by pretty much any other team by now. He seemingly has all the skills you'd want next to James: Defense, brings the ball up the court, 3-point shooting. But, he is just barely good enough at all of those things, and wildly inconsistent (he may be the streakiest shooter on the team next to Brewer) especially with his tendency to foul players in the backcourt for NO REASON. He's our Draymond Green,except not as good of a passer, and in a position that requires less lock-down defenders. He is what he is, and until we find something significantly better, I'm okay with him.
Wow...all the advanced stats makes an eye test and an "observers" opinion mute. I love analytics and getting things "right", but things sure get hostile around here anytime you question stats. It kinda makes me miss the days of pts/reb/a/stl/blk and fg % being the only things analyzed. On the Beverly front I will go on record saying that upon his return, I was a bit leery that he'd take possessions away from JH13. Other than that I'm still way more comfortable starting Bev and having Gordon come off the bench. The Rockets are doing better than most people thought. Our star player is having an MVP type season. We have a lot to be thankful for being Rockets fans after the mess we saw last season. It's easy to point out negatives (or perceived negatives in the case). Dekker is doing good, Corey Brewer has actually started contributing again, Capela is doing good, Ariza has been way more consistent in his production this year. This is a great team to watch. #GoRockets Pass on the positivity!
Small sample size. Let's wait to see how things end up for Bev. that being said, he probably is overrated defensively, yes. It's weird, most people I know and go to school with are big Harden and Rockets haters for the most part, but LOVE Beverley. They think he's some sort of All-NBA Defense, Pit-Bull, Dennis Rodman-esque badass.
It's really way too small a sample size to make an discernible opinions about Beverley, and way too small to be used in any predictive manner. Most advanced stats need a full-season's worth of sample to really be valid, and even then may still be murky. In NBA its just tough to isolate a single player's performance analytically, because their performance is very influence/dictated by their teammates
heypartner clearly has an agenda for Beverley, lol. Engaging in one of the ugliest mental gymnastics I've ever seen. Like it or not, our defense looks really bad in the last few games, and I hope it has nothing to do with Beverley coming back, but I can't be so sure about that. He is definitely overrated on D tho
Nope, that's what psychologists call "projection." I'm not the one with an agenda for Bev. Clearly the OP who starts a thread calling him a "Fraud" has the agenda. I have an interest in stats. OP asked for a statistical analysis of some reason for a 110.9 DRtg during Bev's first 7 games, and I gave him a very simple, statistical litmus test that scientists use to flush out data. And now he doesn't want it.
LOL @ calling me a "delusional Beverley homer"! I have always thought that Bev's defense is overrated after his first season. But "fraud" is the kind of rhetoric an agenda poster like you would use to tear down Harden's teammates in order to prop up your idol. You didn't "destroy" me about the eye test thing. You never had a good answer to my pointing out that you had to watch the action in order to know how a defensive play broke down. No publicly available stats could show you that. That's why coaches watch videos, not just spreadsheets. I also pointed out how DRPM was flawed in showing a player's defensive impact. You never had an answer to that either. It is you who is being destroyed by heypartner in this thread. I do not have the time nor the patience to do what he does. From what you have posted, it is clear that you don't know how to watch a team play defense. All you try to do is pore over the internet to find numbers that would support your agenda and ignore all other facts.
lol someone clearly didn't read or understand what the whole discussion was about and wanted to just comments just to comment without even providing anything of substance. If you're going to write your first post, try posting something that is actually credible
