Defense: Does it really even matter?

[OneNole]

Defense, a debate that has gone on since the beginning of the dds series. Does the defensive rating really even matter? Or is it a rating, which Dennis would have you believe plays little to no difference in the game? As Dennis would put it, "give me an 80 block 60 defender over an 80 defender 60 block guy any day." Well lets dive in and find out how important (if at all) the defense rating really is.

Similar to how Andrei ran his IQ test, I too created a fictional league and ran a control to determine the worst teams record wise in the league. I wanted to choose the two worst teams as well as a mid level defensive team to see how they would respond to the changes. To determine our control group, we would run 5 test sims (would have been 10 but after noticing little discrepancy between the sims we were confident 5 was plenty), strategies were set to 5 for each team, and default ratings were used. Once the sim was ran we determined our chosen teams would be the Charlotte Bobcats, Toronto Raptors, and Memphis Grizzlies.

The control group: (5 season average)
Team - Record - Points against - Opponent FG%
Charlotte 9-73 107.12 48.6%
Toronto 13-69 115.5 50.2%
Memphis 30-52 101.9 47.6%

As can be seen Memphis is a clear step ahead of Charlotte and Toronto who are just horrible defensively. Toronto allowing their opponents to shoot above 50% from the field.

Our next test had us edit the 3 teams defensive ratings of each player to 80 while keeping their steal and block intact.

80s Defense:
Team - Record - Points against - Opponent FG%
Charlotte 19-63 98.44 43.4%
Toronto 24-58 104.32 44.3%
Memphis 35-47 99.88 44.8%

As expected, a huge jump from Charlotte and Toronto which saw an 11% improvement in wins and 9% drop in points allowed as all 3 teams now belonged in the top 10 of total team defense. Memphis saw the smallest jump of the 3 which should come as no surprise as a couple of their players, Tony Allen especially, saw a decrease in their defensive rating as they were already well above 80.

Our next test saw us have each team continue with their 80 defensive rating but steals and blocks were dropped to 30 for all players.

80s Defense 30 Block 30 Steal:
Team - Record - Points against - Opponent FG%
Charlotte 10-72 104.44 45.9%
Toronto 12-70 111.28 46.9%
Memphis 10-72 110.22 48.6%

I am not going to lie, these results came as a slight surprise to me. While FG% defense (Outside of Memphis as should be expected for reasons listed above) team records plunged as opponents FG attempts.

Our last big test had us drop each teams defensive rating to 40 while making all PGs, SGs, and SFs have 70 steal and all PFs and Cs have 80 block.

40s Defense 80 Block 70 Steal:
Team - Record - Points against - Opponent FG%
Charlotte 11-71 110.66 51.1%
Toronto 7-75 118.08 51.8%
Memphis 12-70 115.14 52.6%

The 3 worst teams in the leage not only record wise but also defensively. All 3 teams allowed their opponents to shoot well above 50% and all 3 allowed their opponents to average over 110 ppg.

It does appear that the defensive rating does seem to impact a teams performance dramatically when it comes to opponents fg% but even with elite defense you still need to be able to force turnovers if you truly expect to impact the game. Who knows how much better these teams could have been if their strategies were set to take advantage of their defensive capabilities.

I will admit though, while I wasn't surprised of the impact that the defensive rating had, I was surprised that steals and blocks still remained some what important. So what say you? Did any of this come as a surprise to you?

(Note: Dennis, I also did a small test on 80 block/steal 60 defense vs 80 defense 60 block/steal and the 80 defense continued to have a better win % and saw Charlotte with their most games, same size was small however.)

[GreenBear]

Block ratings definitely have an impact on opponents' FG% because a block equates to a missed shot. So I'm not surprised by that at all. That's a big reason why I have Austin and Biyombo in my front court. They block a ton of shots. I lead the league in blocks, so it comes as no surprise that I'm top 5 in opponents' FG% as well. There's definitely a correlation there.

I would be curious if steal ratings (by themselves, so keep the block ratings at the control level) have an impact on opponents' FG% though. It should result in fewer shots by the opponent (because they turn the ball over more) and less points against, but does it also lower opponents' FG%? My hypothesis would say no, but this engine has surprised me before.

I will admit that I wasn't sure about the defensive rating, but I'm glad to see that it has an effect.

[KW]

I'm not all that surprised. There is some correlation between the DEF rating and the actual defensive impact (with plenty of exceptions but the general trend is there). I find that a combination of BLK, STL, IQ, DI, DEF, and "correct" weight will decide the overall defensive impact. I'm sure endurance and defensive rebounding also play a factor- and in a way, drawing fouls is its own kind of defense because it can take the best opposing players off the floor. There are a lot of players with DEF ratings in the 60s and even some 50s that are very good defenders, because of those other ratings. Obviously an exorbitantly low DEF rating (something in the 30s or 40s) might simply be impossible to overcome. But you're not going to see a defensive liability with a 75+ defensive rating unless the player is incredibly out of shape.

[Dennis]

I said that the defensive rating matters, but that I value block and steal higher. Also I was saying 60 def not 40 def. 40 def is like Harden-defense.

[OneNole]

Block ratings definitely have an impact on opponents' FG% because a block equates to a missed shot. So I'm not surprised by that at all. That's a big reason why I have Austin and Biyombo in my front court. They block a ton of shots. I lead the league in blocks, so it comes as no surprise that I'm top 5 in opponents' FG% as well. There's definitely a correlation there.

I would be curious if steal ratings (by themselves, so keep the block ratings at the control level) have an impact on opponents' FG% though. It should result in fewer shots by the opponent (because they turn the ball over more) and less points against, but does it also lower opponents' FG%? My hypothesis would say no, but this engine has surprised me before.

I will admit that I wasn't sure about the defensive rating, but I'm glad to see that it has an effect.

The 80 block 70 steal with 40 defense gave the worst defensive stats in the league. Obviously that may adjust some if DI/FCP was changed.

I saved each test so I may in the future run a test with adjustments to strategys to each team to see how much further of an impact it gives.

[andrei]

Love this study and results are very impressive. Huge discrepancy between defensive abilities of the teams. We can see that high defense complemented with Blocks in frontcourt and Steals in backcourt is amazingly better than either pure defenders or pure shot-blockers. Which is probably obvious but great to see it with numbers.

Thanks dude - great read

[Rizzo]

All this tells me is that Zhou Qi is going to be the best defender in the league!

[hardenwithnod]

Good stuff, these kind of articles definitely helps for DDS3 noobies such as myself, thanks!

[Marcos_Beck]

Thank God I've built a 77 / 82 / 88 / 74 / 84 defensive lineup lol.
Great one here Brett. Those articles running tests to figure out which rating impacts some part of the engine most are incredibly interesting and everyone should benefit from that.
Well done sir, keep the tests up with multiple strats and let us know the results!

[Chilikonkarne]

Great article and the outcome proves that dds3 is not as bad as everyone say it is