Kelly, I don't know why you keep bringing it up. I never claimed FG% alone is a good enough indicator for defensive strength. I only included it as an auxiliary to facilitate the discussion on team defense. Top 10 defensive team in which way? Last I checked, using 2005-6 stats, the Suns were ranked* - 28th in points allowed - 17th in defensive FG% - 20th in 3-point FG% defense - 6th in forcing turnovers (a top 10, woohoo!) - 6th in opponents' assists (another top 10, not bad) - 22nd in steals - 12th in blocks * All per game stats.
Because you refuse to look at per-possession for evaluation of defense . http://www.knickerblogger.net/stats/2007/d_de.htm They are #9 based on defensive points-allowed-per possession and just about as good as the Mavericks. EDIT: keep in mind that per-possession stats factors all the stuff you listed above into the calculation in translated into number of possessions except for assists and blocks.
Keep look at per-game stats from last season, and keep throwing in sarcastic "woohoo"s ... we'll just have a better understanding of why certain teams do better than others, while you stay a year behind and a quip ahead.
I have Always liked Jon Barry, but after I heard him defend the ROX, I know he is the kick as$ sum b!tch, I love the guy the only team bill like's is L.A. and only because he has to.
To all of the Suns fans out there, Phoenix defense sucks. If your gives up over 100 pts a game, your defense is not that good. I think 100 pts a game allowed is equalivlent to about 25 pts to 30 pts a game in football. You notice teams with great offenses and sub-par defense rarely win championships in basketball. If you look at all of the teams from last years champs going back to the 60s, you notice most of the teams that got to the NBA finals were in top 10 to 13 defensive teams which is mostly determined by points allowed per game. Could you imagine a team that's giving up over 25 pts. game winning the Super Bowl with run up and down the field style? It could happen, but I think teams will exploit Phoenix's nack for not playing defense. I think the Lakers from last year in playoffs gave an idea of how you need to beat Phoenix, slow the game down - control the possesions, beat them up on the inside, out man them down low and on the boards and don't give them easy shots. Even with Amare, the Lakers could use the same strategy again because Marion couldn't beat two or three big men under the basket at the same time. ---------------------------------------------------------------------- When Phoenix gets to the playoffs they might have to do against 4 of the leagues best defenses-----they will be games that don't make over 100, which gives them a greater chance losing. they will be playing against these defensive juggernauts in 7 game series. Hou. 89.7 ppa SAS. 90.6 ppa Dal. 92.5 ppa Det. 95.1 ppa To also look at these teams offenses have players/the squad to score over 100 a game against Phoenix great defense and still have chance to win, can Phoenix win a contest in low 90s to low 80s?
Throughout the Celtics championship years in the 60s when they were considered one of the great defensive teams and led by Defensive Player of the Millenium Bill Russell, they consistently gave up 110+ ppg. Does that make them an abysmal defensive team? No. You know why? Because the pace was ridiculously high back then. Points per possession. That's what really matters.
If you claim Suns' defensive achievements have always been looked down upon, is it a crime for me to cite some numbers registered with the most recent whole season? Further, if you knew I am going to use per-game stats as a standard procedure, why would it be a surprise now that I did it? As for sarcastic "woohoo"s, I am sure you know the tone in your previous replying posts to mine is anything but.
I don't see how that makes them more fundamental. Points scored/allowed per game is also a rate. It's possible for a team to actually allow more points per game on the season, and STILL have a winning record. They were preconceived (I consider them intuitive). Because per-game points scored and allowed are so heavily dependent on pace (even per-game stats + FG%, though to a lesser extent), they actually obscure the fundemental features of an offense and defense, as well as how offense/defense contributes to winning. I calculated how these stats correlated with one another to illustrate this point. win% = wins / ( wins + losses ) It's simply the percentage of games played that you won. Where did I show they were the same? Yes, there are slightly different number of possessions for each team. Generally, I take "Off/poss" as points scored divided by offensive possessions and "Def/poss" as points allowed divided by defensive possessions. Alternatively, you could take the average of offensive and defensive possessions, to use the same number in the denominator for both offense and defense. The first approach is more "real", but the second approach might be more desirable if you want to truly equalize possessions for both teams. Either way, the results are roughly the same. I don't know if I'd go so far as to say it's a sufficient condition for winning. But, yes, if you're ranked very high in offensive or defensive efficiency (per possession), then you're likely a pretty good team. I think that squares with our intuition -- if you're a very good offensive team or a very good defensive team, you're probably pretty good overall. There are exceptions, of course. Seattle, last year, was a very good offensive team (in terms of per possession stats), but their defense was absolutely terrible. By terrible, I mean bad on historical levels. Consequentially, overall they were a bad team. So, we have: Def/g = Def/poss * poss/g We know, intuitively, that poss/g should have little to no correlation to winning% (you have good slow teams, bad slow teams, good fast teams, and bad fast teams ... pace tells you nothing about how good a team is). And because Def/poss has a relatively high correlation to winning%, when you multiply that by pace you're essentially adding noise to the statistic. And you end up with Def/g. That's the best way I can explain it. What wins on the defensive end (or offensive end), fundementally, is determined by what you do on a per possession basis. When you try to modify that by multiplying by pace you're making it a less reliable indicator of winning%. This is what my intuition tells me, and this is what the numbers show.
durvasa...i haven't studied in depth all your posts on winning%'s, etc. But did you have any analysis that supports the above? And particularly not on a regular season basis. thanks
I posted the results for the last 5 regular seasons in a table earlier: Code: [SIZE=2] off/poss def/poss off/g def/g offmetric defmetric pace win% ----------------------------------------------------------------------- off/poss 1.000 def/poss -0.039 1.000 off/g 0.846 -0.158 1.000 def/g -0.173 0.831 -0.522 1.000 offmetric 0.863 -0.045 0.909 -0.334 1.000 defmetric -0.118 0.906 -0.372 0.934 -0.194 1.000 pace 0.224 -0.202 0.697 -0.701 0.507 -0.499 1.000 win% 0.684 0.641 0.490 0.441 0.593 0.537 [B]0.005[/B] 1.000[/SIZE] Correlation between pace and win% is essentially 0.
I was thinking post-season. And then not necessarily win%, but series win. I.e. - is there any correlation between PACE and ability to win series in the postseason, both overall, and more specifically to conference finals and championships?
Well, both team playing a given series necessarily play at the same pace, and one wins and the other loses. You could address that issue, maybe, by seeing how often the faster paced team (in the regular season) won a playoff series.
wnes, I missed this post before, so I'll respond now. The only extra information possession allowed per game would provide when coupled with points allowed per possession is pace. So, you must think that pace is somehow an indicator of defensive ability. I disagree. Purposely slowing the pace might be a defensive strategy for some teams, just as defending the pick and roll in a particular way might be a defensive strategy. However, wlowing the pace is only an effective defensive strategy insofar as it helps to limit the opposing team's points per possession (conversely, increasing the pace is only an effective offensive strategy insofar as it helps to increase your point scored per possession). All else being equal, if one team plays at a different pace than another team but still allows the same points per possession, it makes no sense to say they are better or worse defensively. I've given another example (hopefully more convincing) below to illustrate this. You said the following: "Why is it necessary to use per possession data in order to show which team is better defensively? Rockets ranks No.1 both in points allowed per game and opp FG%. Plain and simple." This implies that you think points allowed per game + opp FG% is as useful or more useful than simply looking at points allowed per possession when evaluating a defense. That's wrong. I didn't understand what this part means. If my initial example and subsequent posts hasn't convinced you yet, I'll give it one more shot. Let's say team A and team B allow the same TS% (points scored per scoring attempt), force the same number of turnovers per possession, and they have identical defensive rebounding%. If this is the case, necessarily both teams will be allowing the same points per possession. I'm claiming that all these factors (or, equivalently, points allowed per possession) cover everything that fundamentally matters in a defense. Now, it's quite possible that team A allowed fewer points per game than team B because they play at a slower pace. Does that make team A a better defensive team? No, it doesn't and here's why. The same basic principle holds for both teams -- they need to score more than they allow per possession to win the game. Yes, team A has less points to score, but it also has propotionately less possessions to score with. So, by slowing the pace, team A hasn't given itself any extra advantage relative to team B by what it's done on the defensive end. Pace, it turns out, really doesn't tell you a whole lot about what wins. That's why evaluating a team based on per game statistics is quite foolish. You take any basic stat -- points scored/allowed, turnovers forced, steals, rebounds, etc. -- and you'll find that they all correlate better with winning% when taken on a per-possession basis rather than a per-game basis.
Exactly...any data on that would be really interesting. I think instinctively, the answer would be that pace does matter (i.e. - the team that played the slower pace in the regular season more often wins the series), but that could be way wrong.
The team that wins the playoff series generally had the better winning% in the regular season (because that team would have home court advantage, and they are generally just a better team). And, as I've shown, pace has little to do with winning%. That's why I'd expect there wouldn't be much of a relationship here.
I can't believe this thread continues to go on and on. Here's last year's playoff results displaying regular season pace: Code: rnd1 win pace lose pace w-pace WCR1 SAN 88.5 SAC 91.9 L WCR1 PHO 95.8 LAL 90.9 H WCR1 LAC 91.7 DEN 93.9 L WCR1 DAL 87.8 MEM 86.3 H ECR1 NJN 89.8 IND 89.7 H ECR1 MIA 91.6 CHI 92.9 L ECR1 DET 86.8 MIL 91.2 L ECR1 CLE 89.8 WAS 92.3 L semi win pace lose pace WCS PHO 95.8 LAC 91.7 H WCS DAL 87.8 SAN 88.5 L ECS MIA 91.6 NJN 89.8 H ECS DET 86.8 CLE 89.8 L cf win pace lose pace WCF DAL 87.8 PHO 95.8 L ECF MIA 91.6 DET 86.8 H f win pace lose pace NBA MIA 91.6 DAL 87.8 H The team with the higher pace won the series equally as teams with the lower pace (actually higher pace teams have 1 more series win vs. lower paced teams). Higher pace teams and lower pace teams had about an equal % in winning a series as well as losing. In other words, there was little correlation to series wins related to pace. Of course durvasa's correlation data pointed that direction but who's counting right??? Let's just ignore that because it's "not intuitive". I had mentioned this earlier. Do you know who the lowest pace team in the NBA is this season? The Portland Trailblazers. Wow what a defensive powerhouse! They almost made it in the top 10 NBA teams that hold their opponents to the lowest points per game coming in at #11 at 96.6 PPG! Wow they are a good defensive team huh? NBA league median is 99 PPG! Oh wait they only happen to have a 16-24 record coming in at #23 in the W-L standings? How could that be??? Could it be that defensively they are NOT #11 but #27/30 allowing their opponents to score 111.1 PP100P???? Hmmmm.
Except it's not inuitive. Nor is it intuitive that the Heat actually had more possessions per game than Mavs last year, or that the Spurs did, too...heck, the Mavs almost had fewer possessions per game than the Pistons. The fact that the Mavs ranked 27th out of 30th in pace is shocking to me. Is possessions/game even the best measure? Is fast-break points per game better, or # of times a team shoots before 16 seconds in the shot clock. They're 24th in "offensive efficiency". They could also get better at that, stay the same defensively, and improve. Nobody is discounting the benefit of per possession stats. But...intuition sparks curiousity which fuels statistics. A run and gun team hasn't won the NBA championship in at least 2 decades. Why not?
You were the one that said "intuitively" pace should have made a difference in who the winner of a series was . I'm just poking fun at that at both our expense since the results of playoff series wins looking at pace basically reflects what durvasa's correlation data was already telling us - which is there was no corellation between pace and series wins. So let's just not measure anything then...EVER...because there's always something you else you can look at . Yes they are beneficial. Much more beneficial than PPG stats if you're comparing teams. You'll have to explain how this is relevant to the discussion. I can try to conjecture where you are going with this but I'll just make a fool out of myself guessing . So let's say I'm trying to buy a firewall. Product A and Product B are put to the tests around the country 82 times. Each test lasts 48 minutes. At the end of the tests Product A allowed only on avg 80 hacker intrusions at 1000 attempts in 48 min while Product B allowed on avg 100 hacker intrusions in 2000 attempts in 48 min. Which f/w is the better defender against hackers? A or B? It's B because it stops 95% of the attacks while Product A only stopped 92% of the attacks. If we used the per-possession concept the comparison here would be to NORMALIZE the hacker attempts to say 1500 attempts each: Product A: 120 intrusions per 1500 attempts Product B: 75 intrustions per 1500 attempts And don't go saying well there's other things to look at because the whole crux of this thread was regarding why PPG stats are misleading because they are inflated (or deflated) by pace. So there!
I think that wins in the playoffs are probably more determined by execution in half-court sets. Good playoff teams are better prepared, rotations tighten up, and teams play more consevatively on offense, leaving less opportunity for fast breaks. It would be interesting to see if good fast-paced teams tend to beat up on bad teams more in the regular season, while good slow-paced teams usually play steadier and better against good teams.
