Crunching the numbers. Defense is still the key.

actually considering the percentages are nearly identical, it seems it's not a defense issure or offense one - but rather a hustle issue. (Rebounds and getting loose ball - thus affecting the number of shots we allow).

But I wouldn't look at the last 10 games, we're still a work in progress - the last few games have been amazing though.

 

Defense is always key. In every sport. Check the percentages. Defense wins championships. Go ask the Spurs and NY Football Giants

 

That is kind of interesting. And here's how they rank by offensive and defensive efficiency. I use .976*(fga + .44*fta - orb + to) to approximate possessions:

Code:

          [b] OFF (rnk)       DEF (rnk)[/b]
season     105.3 (18)      102.4 (2)
last 10    107.2 (16)      101.6 (2)

Even though we were 9-1 in the last ten (before the Atlanta game), we won a lot of very close games so statistically the numbers aren't dominating.

 

hmm... Hayes play less = better defense... coincidence?

 

Here's how our defensive efficiency and Chuck's minutes have gone throughout the season. Using 5-game moving average.

 

Nice Graph. It shows what kind of an impact Scola and Landry are having defensively. They are so much better on Offense and don't give up a lot on defense. They've really kept their fouls down as well lately which makes Hayes essentially expendable.

 

WOW great stuff man! Keep it up.

 

I'm not sure how you read that from the graph. Just to be clear, the lower the "defensive efficiency", the better the defense. There's no indication from these numbers that less minutes for Chuck leads to better defense, as lui1107 suggested.

We did do a pretty good job defending in the previous 3 games, without Chuck play a whole lot. That's encouraging, but then again we weren't facing particularly strong offensive teams so I don't how much we can make of it.

 

I'm reading it like less minutes from Chuck isn't hurting our defense. Thus, playing Scola and Landry more minutes isn't hurting our defense either. I think as Scola and Landry have played more (as opposed to earlier in the season when Scola's defense was atrocious) their defense and positioning has certainly improved. That's a good thing because they have an offensive game as well.

 

Actually, the stats further proves that this team needs to be better on OFFENSE.

JVG already proved that you can't win even while giving up 90 points... when you only score 88. The Rockets desperately needs to improve their offensive efficiency while still maintaining their defense.

I feel it's already getting better with our rookies all playing pretty efficient on that side of the ball. And if T-Mac can find his touch, Bonzi capable of finishing around the basket, and Battier can revert back to his usual shooting level, I'd really be optimistic come playoff time.

 

Did you graduate from MIT as Morey did?
Good job!

 

Agreed on whether the graph is telling us anything or not. The 5 game moving average doesn't seem to have smoothed out the oscillations and there does not seem to be any discernible trend in defensive efficiency against Chuck's minutes.(changes are within the same band). Perhaps the message is that defensive efficiency is not greatly affected by Chuck's minutes.

 

Good defense wins championships.

Signed,

The Spurs and The Pistons

 

Perhaps, though personally I think there is a relationship in the graph that improvements in defensive efficiency tends to coincide with increase in Chuck's minutes.

Here's another way of looking at the data:



I plot Chuck minutes in each game against what I call "Adjusted DRTG", which is just our DRTG for the game minus the opponent's efficiency on the season. Not sure how statistically significant it is, but you can see the general trend. And, if you're curious, the correlation coefficient remains negative if I remove data points from each end, though if I just remove one from each end it flattens some. However, if I remove more than one from each end the slope increases.

Here's the raw data, if people want to play with it:

Code:

[B]Date          DRTG     adj DRTG  Chuck's min[/B]
2007-10-30    98.4      -12.6     26.1
2007-11-01    98.6      -13.0     32.5
2007-11-03    89.6      -18.0     23.4
2007-11-05    117.5     6.0       25.2
2007-11-06    95.3      -12.7     20.4
2007-11-09    95.4      -8.3      22.5
2007-11-11    94.8      -8.4      22.4
2007-11-13    112.0     6.7       19.1
2007-11-14    102.1     -8.9      27.9
2007-11-16    109.7     1.7       13.3
2007-11-17    114.7     2.3       17.1
2007-11-21    107.2     -4.3      31.2
2007-11-23    113.2     12.8      11.4
2007-11-24    87.9      -18.7     27.0
2007-11-26    79.1      -21.8     24.2
2007-11-28    98.2      -14.2     22.7
2007-11-29    116.5     5.8       9.7
2007-12-01    115.8     9.5       18.2
2007-12-05    108.2     2.9       15.7
2007-12-07    98.9      -3.0      15.7
2007-12-09    108.1     -1.6      24.3
2007-12-10    111.8     7.5       10.5
2007-12-12    88.0      -23.3     41.5
2007-12-15    118.6     7.1       28.7
2007-12-19    101.0     -9.1      24.1
2007-12-20    100.7     -5.9      24.6
2007-12-22    104.9     2.9       23.7
2007-12-23    103.8     -7.5      26.5
2007-12-28    88.5      -16.8     27.1
2007-12-29    83.9      -25.8     34.0
2007-12-31    112.2     1.5       22.4
2008-01-02    110.0     0.4       14.4
2008-01-04    103.6     -6.5      17.8
2008-01-05    105.0     1.2       36.0
2008-01-08    95.5      -12.4     24.0
2008-01-09    102.7     -1.1      25.6
2008-01-11    91.6      -9.9      19.3
2008-01-13    99.2      -10.9     18.7
2008-01-15    118.0     13.7      17.7
2008-01-19    98.4      -9.6      18.1
2008-01-21    91.0      -9.1      14.7
2008-01-23    116.8     16.7      14.6
2008-01-25    89.9      -17.7     7.0
2008-01-27    111.1     -0.5      23.7
2008-01-29    110.5     -0.2      0.0
2008-02-01    106.7     2.2       12.6
2008-02-02    99.3      -4.4      7.4
2008-02-04    97.1      -4.4      11.6
2008-02-07    93.1      -12.0     3.2
2008-02-09    97.5      -6.2      20.0

 

it's a very weak correlations, meaning 13.5% of the variation can be explained by the the x variable - not statistically significant at all. I do see with the moving average what you mean by the relationship, but the correlation on the raw data isn't good.

And correlations have to be thought of carefully because there could be another factor that influences both (meaning they are not independent variables). For example, chuck minutes could go up in certain situations for example when we are playing teams that do not score a lot of points or have good offenses.

 

That's a good point. I tried to address that by using an "adjusted DRTG". So, it's not how much we hold them to, but rather what we hold them to compared to how they usually perform.

 

the .135 number may not show a great correlation, but the 5-game moving average shows a clear connection b/t chuck's minutes and defensive efficiency.

early in the season, he got more minutes, and efficiency was low. then he lost minutes, it went up. then that cycle happened again, and then during chuck's highest minutes, our efficiency reached it's peak (or valley if you will) and then slowly increase as chuck's minutes dropped. only the last few data points really have some of our better D w/o chuck playing. so maybe it's good that it's not slipping any more w/o him.


and overall, yes our D has still been very good this season. considering how little deke has played, it's fairly impressive our D is still up at the very top.

 

I get all the trend and numbers but...whats "DRTG" again?

 

Someone just said that since we made Scola a starter at the PF spot with Yao that our PPG average is 100.

DD