"My rule was I wouldn't recruit a kid if he had grass in front of his house.
That's not my world. My world was a cracked sidewalk." —Al McGuire

Monday, August 11, 2008

Is the Aggressive Defense Worth it?

During the two-part Big East Preview, there were a number of questions that popped up in the MUScoop.com thread. One poster, Pardner, who has helped contribute information regarding Official Bias, wondered if we were going to see any change in the fact that TC's teams always outfouled the competition. So I decided to look into this a little bit further.

If you look at the 2007 Pomeroy Marquette page and the 2008 Pomeroy Marquette page, one can see Marquette's Defensive Free Throw Rate (FTA / FGA) ranking. In the last two years, Marquette ranked #236 (in 2007) and #262 (in 2008). Despite being a Top 15 team according to Pomeroy, Marquette was one of the worst teams in the country at preventing opponents from getting to the line. Last year Marquette fit in comfortably between Albany and Sacred Heart. yeesh

While one could say that this is a product of our guard-centric lineup and big man deficiency, I don't believe that roster mix is the reason. After all, Marquette put our opponents on the line more regularly than South Carolina Upstate. (did you even know that was a school? yes, they exist)

Instead, I attribute this more to our defensive philosophy, which is commonly known as an aggressive style defense that favors deflections and steals. In fact, my key assumption is that the there is a link between the aggressive style of defense, increased opponent turnovers and more fouls committed. Marquette had a stingy and thieving defense last year, ending up #11 in the country at steal percentage. We forced turnovers (#49 in the country), and when the defense was working we were off and running for easy points.

But was the defensive strategy worth it?

If you look at the Pomeroy Game Plan for Marquette, you'll see that Marquette had a 99% significant correlation between our Opponent's Free Throw Rate and our Defensive Efficiency. In other words, when our opponent had a good Free Throw Rate, our defense suffered. (As I say all the time, it sounds intuitive, but it's not the same for every team. It doesn't apply to Wisconsin) We allowed our opponents to have a FTR of 41% last year. That means that an average opponent, shooting 54 FGA, would also take 22 free throws. To repeat, this was #262 in the country.

Against Top 100 opponents, the story is even worse. Our opponent's FTR goes up to 46% (~25 Free Throw Attempts per game).

Yeah, but so what? The average impact of this Free Throw Rate was a worsening of our Defensive Efficiency of 26.2, with a range from 16.9 (low) to 35.5 (high). Keep in mind our season Defensive Efficiency was 87.8, so this impacted our defense by 20-40%. I calculate that the way we put opponents on the line cost Marquette an average of 17.8 ppg with a standard range from 11.5 ppg to 24.2 ppg.

As previously stated above, Marquette's aggressive defense was one of the best in the nation. How much did we benefit from our aggressive defense? Going back to the Game Plan, we also see a 99% significant correlation between our Opponent's Turnover Rate and our Defensive Efficiency. More turnovers from our opponent means that our defense was better. (Still sounds intuitive, but it doesn't apply to Pittsburgh). We forced opponents into a turnover rate of 23.4%, meaning that our opponents ended up with about 16 turnovers per game, or almost one out of every four possessions.

Against Top 100 opponents, the turnover story is also not as good. Better teams protect the ball more. We only forced opponents into a turnover rate of 21.7%. Remember that 20% is considered average (a turnover one out of five possessions).

We come back to the so what? The average benefit on our defense was an improvement of our Defensive Efficiency of 13.2 with a range from 10.2 to 16.3. On a percentage basis, this is about 10-20% of our defensive results. I calculate that the turnovers for our opponents cost them (and benefited Marquette) an average of 9.0 ppg with a standard range from 6.9 ppg to 11.1 ppg.

If the assumption is right, then the style of defense that's taught is to focus on deflections and steals, which leads to more turnovers and fouls committed. The problem is that this benefit to our defense is not enough to overcome the negative implications from the extra fouls we commit. The extra trips to the free throw line for our opponent are worse for us than the times our opponent commits a turnover.

Furthermore, since conventional wisdom tells us teams protect the ball better in the post-season, does Marquette's defense get exposed even further? Our ability to force turnovers went way down last year in the post-season, but I'll have to do more analysis before saying the defensive strategy gets worse for MU.

If the assumption is wrong, then I can't blame the defensive strategy and emphasis on turnovers for the extra fouls committed. However, the fact remains that there is a significant impact on team performance from allowing our opponents to head to the free throw line so much. One thing is certain, no matter what the implications of the defensive strategy may be, the coaching staff needs to emphasize that we stop putting opponents on the free throw line so much.


Steve Susina said...
This comment has been removed by a blog administrator.
Gene Frenkle said...

Well, we lost Dan Fitzgerald so that should free up about six-to-ten fewer free throws by our opponents per game.

Steve Susina said...

Since Cracked Sidewalks welcomes alternative opinions (presumably even those of listed contributors) I'll try with this alternative opinion to my esteemed colleague's post once again--I have certainly tried to strip all emotion from this comment and will try to keep it all business.

I think its a legitimate question to ask what the impact of fouls that are intended to put the opponent at the line rather than give them an easy shot from the floor. This is in addition to generating turnovers and deflections.

I'd like to compare MU's #262 ranking to Notre Dame (#4). I chose Notre Dame because there is an underlying assumption that #262 is bad and lower would be better. Improvement to the level of Notre Dame would put MU as one of the best teams in the country with respect to defensive FTA/FGA ranking.

I don't agree that this would necessarily be a good thing.

Start by comparing FTAs:
Notre Dame opponents: 488 FTAs, or 14.8/game
Marquette opponents: 767 FTAs, or 21.9 per game

As one might expect from the #262 team, MU gave up on average 7 more Free Throw Attempts per game than #4 Notre Dame.

IMO the logic of this stat suggests that that by putting opponents on the line Marquette opponents would wind up with fewer shots from the floor.

After all, every possession has to end in either a shot attempt, a foul, a turnover or the end of play. MU is limiting shots through aggressive defense--trying for turnovers, but also causing more shooting fouls (no FGA when the opposing player is fouled in the act of shooting).

And as it turns out, MU does indeed show 10 fewer opponent attempts than Notre Dame:

Notre Dame foes; 2125 FGA or 64.4 per game
Marquette foes: 1891 FGA or 54.0 per game

The net is that MU's more agressive style resulted in 7 MORE opponent FGA than Notre Dame, but 10 FEWER shot attempts.

All other categories essentially cancel each other out. ND had more defensive rebounds, Marquette had more steals, blocked shots were a wash, Tempo was pretty close (ND averaged about one more possession per game).

Assuming 80% opponent shooting on FTs and 50% on FGs, MU's agressive style would result in an average 6 extra points at the line and prevented 10 points from the floor, for a net benefit of 4 points.

I'm open to a give-and-take discussion of any flaws in my logic (and in the spirit of open debate I hope my fellow blogger is as well!), and look forward to further spirited repartee.

theKAYman said...

Great article and comments. Very insightful.

Oliver said...

Maybe I am missing something, but don't the benefits of a turnover include the lost points by the opponent and the benefits of a fast break (increased shooting percentage, increased number of shots and free throws, increased chance of a 3-point play)?

Your analysis only balances opponent free throws against our defensive efficiency.

In a less direct connection, it also doesn't include the effect on the opponent's defensive efficiency, assuming they get worn down by our aggressive play.

Rob Lowe said...

For MU84, let's just agree to disagree. Thanks for the comment on the blog.

For Oliver, the 2008 Game Plan suggests no statistical correlation between our Opponents' TO% and our Offensive Efficiency (AdjO). (That's the same as our opponents' Defensive Efficiency too.) If anything like increased shooting percentage, increased number of shots and free throws, increased chance of a 3-point play happened, then it would be captured in the AdjO, as well as in the MU eFG% and MU FTR. FYI, both of those rankings were pedestrian (#141 and #126, respectively). It' probably shows up in the OR%, which was very good for MU last year (#23).

Also note that I made an assumption regarding the defensive strategy resulting in more fouls committed and more turnovers. Even if this isn't true, it doesn't change the fact that we foul too much.

Let me make one final point about our opponents' Free Throw Rate. Against Top 100 opponents, we averaged a FTR of 46%. We played 15 games when an opponent achieved a FTR higher than 46% against us last year. We lost TEN of those games. Of the five games we won, three were blowouts (UWM, SHU#2, Providence), and two were some of the ugliest wins of the season (USF, SJU).

In fact, when keeping an opponent under 46%, we NEVER LOST.

Chris said...

This was an excellent post. Keep up the good work.

greyCat said...

If steals and blocks (leaving aside rebounding for a moment) can be interpreted as markers for the presence of aggressive defense Henry, then comparing the propensity to foul with the tendency to steal or block supports you -- Marquette's defense is indeed aggressive. The team's three most prolific theives in 2007-08 (McNeal, James and Hayward) were also ranked #1, #2 and #4 as prolific foulers. Pomeroy ranked McNeal, James and Hayward in his Top 500 for possesion-based steals (#44, #149 and #487 respectively). Not so odd when you think about it, but Barro and Burke, Marquette's two most efficient blockers, were the team's #3 and #5 ranked foulers, respectively.

I am reluctant to buy Marquette 84's characterization that high FTRate is a deliberate strategy to exchange free throws for field goal attempts, but rephrase that to something like a "recognized possible outcome" of aggressive defense -- a calculated gamble, and I would agree. I would expect a high free throw rate given up by a young team, but given Marquette's experience, that might be a (yellow?) flag going into next season.

Very good analysis Henry.