One of the things that gets mentioned a lot is that post-entry passes into Yao and Yao's work in the low post result in a lot of turnovers. This has been cited as a reason as to why Rockets may improve offensively without Yao. A review of the facts suggest otherwise: With Yao on the floor, the Rockets average 14 TOs per 100 possessions. With him off the floor, the Rockets average 13 TOs per 100 possession. This is true this year as well as last year. It appears any effect Yao has on TOs is mild at most, under either of the last two coaches. It's not that Yao doesn't turn the ball over, or the entry passes don't result in turnovers sometimes, it's just that it seems the Rockets find other ways to turn the ball over when they are not throwing entry passes to Yao. Any kind of aggressive offensive act-- driving into traffic, posting in the paint, passing the ball around, basically anything other than chucking jumpers-- result in an increased number of TOs. Yao took a lot of TOs because he handled a lot of the aggresive, high TO risk work when he's in there. When he's off, the TO risk didn't disappear, it just got distributed to other players who had to take on the aggressive high TO Risk work.
If those numbers are from 82games, that's actually turnovers per 48 minutes. It's also rounded to the nearest whole number. In terms of TO% (turnovers per possession), with Yao on the court it would be between 14.9% and 16.0%. With Yao off the court, it would be between 13.6% and 14.7%. So, it potentially a minor difference, but it could also be very significant. Worst case, the difference between 13.6% and 16.0% is ranking 3rd and 16th respectively in team TO%.
My numbers were from 82games.com, so yes, they were per 48 minutes and rounded. I am curious as to why your numbers contain a range? Is it some sort of error/confidence interval? If we take the midpoint of each range, would be 15.45 vs. 14.15, about 1.3 TOs per 100 possession.
If it is a CI, the midpoint of each range would represent the best estimate i.e. 15.45 +/- a margin of error of .55. Unfortunately, as Durvasa mentioned, the separation of each range can be minor or significant and can only be determined when analyzed. All of our statements would be useless without the actual data from each game and the turnovers of each to run a t-test comparing both means of turnovers.
I assume 13 turnovers per 48 minute is actually between 12.5 and 13.5, and 14 turnovers per 48 minutes is actually between 13.5 and 14.5. Then I used the possessions per minute data for Yao on and off the court from basketballvalue.com, updated through the Wizards game.
Hmm durvasa. Not sure why you are using a range. If so you have to range everyone else as well. Carl has a very good point. I've always been annoyed with folks focusing in on Yao's turnovers and leaving everything else out. It's very misleading. For instance (based on my calcs): Rockets TO's per 100 with Yao in: 15.3 Rockets TO's per 100 w/out Yao: 14.6 Wow, without Yao's turnovers we move up from roughly ranked 16th to 8th in the NBA. Let's stop Yao playing because we are so much better without his turnovers!! Fire Adelman! Fire Les! Fire Morey! Trade Yao! Not so fast. Let's take a look at offensive efficiency which includes the impact of turnovers into the calculation (by my calcs): Rockets Pts per 100 with Yao in: 106.3 Rockets Pts per 100 w/out Yao: 103.2 So the team's offensive efficiency even removing Yao's turnovers actually performs worse than with Yao's turnovers. Hmmm. So wow we improve on our turnovers without Yao (up 8 spots) but offensively we actually fall 9 spots from 16th to 25th. Just for reference sake, (by my calcs) this season the league average offensive efficiency (which again includes the turnovers) is 107. If you look at Yao's individual offensive efficiency including turnovers, it is 115.6.
Good works, Guys! Very interesting stats. Tango, would you share how many pts Rockets give up per 100 possession with Yao and without? Thanks in advance.
That's harder to do exiaol! I'll see what I can do. I'll have to basically get on/off court figures to try and figure out points allowed with and without Yao. This means synthesizing data from 82games or other sources that have on/off court stats to do so. Maybe durvasa has some ideas.
I may not be understanding what exiaol is asking, but isn't that just Off Court Defense, shown http://www.82games.com/0708/07HOU16D.HTM ? Actually, I'm not sure how you calculated your offensive efficiency numbers with and without Yao. This is given in the above link, and it doesn't match. And I used that range because I'm pretty sure the "13" and "14" figures cited by Carl Herrera were rounded to the nearest whole number by 82games.com. The 13, for instance, could actually be any where between 12.5 and 13.5 with equal probability.
No, they don't match at all. It all depends on how possessions are factored. (1) 82games and basketballvalue are based on on/off stats. I have no idea how they are coming up with possession numbers. On basketballvalue they show you the minutes and possessions totals. There's a huge difference between the basketballvalue possessions per game vs. typical regression based possession equations. Rockets basketballvalue pace per game: 108 Rockets typical eqn pace per game (what I used): 90 Rockets knickerblogger pace per game: 89 That's a whopping 18 possessions difference per game which I have no idea how they came up with. Comparing various teams from basketballvalue, their game pace numbers for the teams I looked at are all inflated compared to what you find on knickerblogger.net. I have no idea why. Personally I'm finding it really hard to believe the Rockets are averaging 108 possessions per game this season. (2) My calcs are based on invidual (Yao) and team totals to calculate possessions. For the team without Yao I just pretended like Yao wasn't ever with the Rockets and removed his stats from the calculation of team offensive efficiency. Theirs is based on on/off court analysis which in theory should be a better reflection with and without Yao. But their pace numbers with Yao on/off are pretty screwy IMO. On/Off +/- stats have their usefulness but to be honest I've never liked where Rosenbaum and Pelton have tried to take these stats because there are just to many unanswerables with +/- stats because of issues like not really knowing the cause and effect since +/- stats you're really comparing the 5 guys on the court vs. and their subs vs. the opponents 5 other guys on the court and their subs as well.
So, the Rockets can put to bed that concern when pondering on whether to trade Yao... ...next case, er..thread...
I'm confused. That's not the number I get when I use basketballvalue.com's data. http://basketballvalue.com/teamunits.php?year=2007-2008&team=HOU According to this, Rockets have played a total of 2746 team minutes, and they've had 5191 offensive possessions and 5211 defensive possessions. That's comes out to a 90.9 pace factor. This more or less matches the pace I'd get using the stats at 82games as well (not including Wizards game): On: 1.065 pts/poss, 4061 total points, 2033 minutes => 3813.1 possessions Off: 1.081 pts/poss, 1332 total points, 654 minutes => 1232.2 possessions So: 48* (3814.1 + 1232.2 poss) / (2033 + 654 min) = 90.1 pace 82games and basketballvalue.com are based on play by plays. I believe they are more accurate then the methodology (I think) you are using.
IMO, we are a better TEAM with Yao Ming out of the lineup. Our defense gets stronger, we run faster, our endurance increases and our turnovers decrease. We lose a bit of offensive presence in the paint, but i'll give that up to take advantage of all the above mentions positives. The rockets will continue to roll without Yao. I'm excited.
We're a better team without Yao against crappy teams - but against really good ones we really need him. Take a look at our last loss against Utah. Yao is an offensive powerhouse particularly in close games because you can't foul him since he shots so well. You need Yao to beat a Phoenix, LA, or San Antonio. And we've won against GS with Yao as well. We can go far and knock of a team like Dallas without Yao - maybe even Utah. But we won't get past a SA or Lakers without him.
For the Rockets: 2746 minutes 5191 possessions. 2746min / 57 games = 48.17 min/game 5191 possessions / 48.17 min/game = 107.7 pace In your calcs, your total possessions = 5045.3. They list total possessions at 5191. You're missing 146 possessions which is almost 2 games worth.
Just noticed that your Yao on/off court minutes don't match what they have listed as well. They list 2044 on, 702 off = 2746 minutes total. You show 2033 + 654 = 2687 total minutes.
I know you guys are into math ... but is that hard to tell that Rox is a better team with yao? even with his TOs? I recall some1 posted a TO leader list, it consists mainly of great players : Nash, TD, H12, Yao, ....players like Battier and Hayes won't have a lot TOs, because they are seldom involve in offense.
