DFDepressed FanDepressed Fan



, all the time

The Disruptors

O.T.- what was you quick take from training camp yesterday?

A couple things stuck in my mind.

1. They're really, really young.
2. The receiving corp are the stars of the team, which is pretty funny considering how they were the weakness basically throughout the McNabb era.
3. It looks to me like Celek is running a ton of the old "Westbrook" plays, especially the shovel pass.
4. Vick really couldn't hit the broad side of a barn with a pass over 10 yards. It's pure luck whenever he completes anything other than a dump off. He was picked multiple times, and airmailed just about every receiver he tried to hit down the field.
5. Kolb didn't look bad to me, very accurate, strong arm.

The disturbing thing was watching Rocca, Dorenbos and some kicker (not Akers) playing soccer on field number three for a couple hours in the morning.

Honestly, I didn't take the time to read everything. I did skim most of it, though.

Blocks and steals should never be equal in my book, though. It is like using FG% instead of TS% basically. Should be .5 blk + stl + charges taken.

I like it, though, and it shows where many of the more disruptive(can be good or bad) players are.

I can't help but say rebounding should be considered as well in any defensive metric, even if to a lesser extent than being counted as a gained possession. I realize you aren't going for an all encompassing metric, just had to say it.

Yeah, I'm just looking at this as a piece of the defensive puzzle. As for weighting blocks, I think using positional averages helps in that regard, if we were looking for a way to normalize so, say 2.00 is the average, regardless of position, then I'd get into a modifier for blocks.

How'd you come up with .5? Just off the top of your head, or are only 50% of blocks recovered by the defending team? That seems low to me, and I'd argue that even blocked shots that go back to the offensive team have some value, how much is anyone's guess (deeper in the shot clock, intimidation forcing teams to take jumpers rather than be embarrassed, etc).

Joe reply to Brian on Aug 11 at 8:24


I know just bringing up Berri makes some people angry, but whatever. He has the value of a steal as 0.033, a block at 0.019, and 0.033 for DREB.

So he has the value of a block at about 57.5% of that of a steal. That is what his regressions yielded(so that should be pretty much the best value to assign it relative to steals if one considers just box score data) Berri then has team defensive adjustments as well, but they spread the worth of a charge to all players on the team and stuff.(he only uses box score data)

You could use .575 or just .5, which is pretty much the same

A defensive rebound is worth the same as a steal. Not saying I agree with this.

Personal fouls also should not be ignored in an all catching defensive statistic.(which this isn't supposed to be) Berri takes the approach of evenly distributing free throws made across the team. So if Sammy had 10% of the PFs on the Sixers, he would be charges with 10% of the FTM. The weight is -.017.(marginally more than than half a possession) Not saying you should do this. Just an example.

Do you have any ideas on the best way to include rebounds in this metric? Pure rebounding numbers don't do it for me, it would have to be a form of defensive rebounding rate, pace adjusted to fit into the per-36 metric. I could definitely see finding a way to work it in, though, I think with stats that are available from these sites.

Brian -- great work (and I can't imagine how long it took you to put all that together). I'm always interested in new ways to evaluate defense in the NBA, and "defensive plays" is one of them.

A few comments:

* Back in the late 70's, there was a book called "Basketball Jones" (not the same as several other present-day books) that rated the NBA players of the day in several categories, from 1 (poorest) to 5 (best). For defense, the categories were defensive anticipation and defensive position. I remember that my favorite player at the time, Julius Erving, was a 5 for anticipation and a 1 for position. Dr. J was made for the "defensive plays" stat; there were some years when he was top 10 in both steals and blocks.

* The distinction between defensive anticipation (or defensive plays) and defensive position would also distinguish teams like the Warriors from teams like the Spurs. Position is probably more important, as the majority of defensive possessions don't wind up in a defensive play but rather a shot and possible rebound attempt. In those cases, straight-up defense, help defense, and ability to challenge the shot are important.
All of those are areas where teams like the Warriors are porous and teams like the Spurs excel.

* In some sense, for fastbreaking teams, steals and blocks are more valuable (as individual plays) than other defensive stops because they are more likely to lead to easy baskets on the other end. I don't know if anyone's ever done an analysis of "offensive efficiency after a defensive play," but I'd guess it would be higher than normal. The logic doesn't apply to charges drawn (since there is a stoppage in play), whose main value (other than the stop itself) is increasing the foul total of the opposing player.

* Obviously, the best defenders are those who make defensive plays and play sound positional defense. Historically, there was no one better (in my basketball-watching lifetime) than Bobby Jones, who got tons of blocks/steals/charges drawn but still played excellent positional defense. Still, there are some players who make few defensive plays but still play good positional defense. Historically, Joe Dumars was someone like that. In recent years, Battier was thought to be that type of player (few steals, good positional defense), so it's interesting that he performed well in this metric. I haven't seen Joe Johnson enough to know whether he fits this mold as well; he has a good defensive reputation, but it may not be deserved.

* Finally, when it comes to the current Sixers, I would say Iguodala and Jrue are both of the type that make defensive plays without excessive gambling. It's impressive that Iguodala is usually among the top ten in the league in steals without being a gambler on defense (he gets a lot of steals by stripping people in the post or deflecting passes). In his prime, Iverson was really a gambler. I'm surprised that Carney didn't show better, because I've always thought of him as a blocks/steals guy. And of course, the fact that Green and Kapono performed horribly matches what others have found, including the analysis at basketballgeek.com, where they ranked 426th and 430th (last) out of 430, respectively, in adjusted individual defensive efficiency. Some guys are bad defenders regardless of the metric (the anti-Bobby Jones).

Thanks, it took too much time :)

I love that ranking system, position and anticipation. That's really what I was looking for here, hidden value in the form of guys who made these plays playing within a defensive system that was both successful and fundamental. Thomas really stood out to me, as did Josh Smith, obviously. Houston has a bunch of guys who ranked well, including Jeffries (who was also good on the Knicks), Lowry, Battier and Ariza. In fact, so many guys on Houston rated well here, I have to think their advanced statistical scouting must use some derivative of this type of stat to target players.

Blocks, charges and steals are like dunks. They are highlight reel positive plays- but only a small part of your measure. You can be a good defender and not rack up huge numbers in these categories- although there is a correlation.

In a vacuum, I agree. The line of logic goes like this, for me.

1. I want to build a team that creates opportunities in transition through defense.
2. When I'm evaluating a player, I want to see the player's adjusted DEF/36 and add some weight to it because I value those plays.
3. How does his adjusted DEF/36 rank at his position.
4. Finally, I want to see how he made those plays. If he did it by gambling, or playing in a system that didn't care about fundamental defense, the numbers don't have as much meaning, but if he did it while still playing solid fundamental defense for a solid defensive team, then he's more valuable.

A good defensive player who doesn't make defensive plays is still a good defensive player, but a good defensive player who makes a ton of defensive plays is more valuable IMO. I don't think dunks are a good comparison.

If a player is scoring a ton on dunks it makes for very efficient offense. That is not a bad thing. In some cases it corresponds with good offense (like MJ) and sometimes it does not tell you as much (Bynum and Howard), but IMO I'd rather have my team score by a dunk than any other way.

Just like I'd rather have a steal or a block- but its only a small part of defense.

Eh, I'm still not buying the comparison. We aren't lacking for metrics to judge efficiency on offense, especially for scoring. Dunks falls into that. I think DEF is measuring something that isn't readily apparent through other stats.

I don't think its possible to ever come up with a defensive metric that comes close to what is available on offense. It would have to include things like deflections and "holds"- which I would define as percentage of your opponent's touches that do not result in a score, foul or assist. I'd also make a stat for "super-holds" where your opponent attempts a move either in the post or off the dribble and fails to get a score/foul/assist.

Yo Brian,
Tickets for the game on Sunday v. France at MSG start at $15. Game starts at 1pm. You still thinking about going? How about any other NYC depressed fans?

Yeah, I'm definitely interested. Haven't gotten tickets yet, $15 sounds good to me.

ESPN is projecting 31 wins for the Sixers. Here's the link.

I'd say 34, but 31 is not that far off. Anywhere from 28-42 wins would not surprise me. The team has huge question marks but a lot of players with specific talents.

Also, the preseason schedule has been released. Two games in Philly, one in Newark. The rest are all over the place.

tom moore on Aug 10 at 13:44

Sitting in classroom waiting to see USA workout.

Are they working out in NYC now?

You should try and get yourself some press credentials like the guys at Libertyballers have... maybe try through Mike Preston.

Are there defensive stats that use the persons a player's guardings shooting percentage or something into account? so that things that dont show in the box score can be counted.

I've never seen stats that count opponent's production on a granular level. 82games.com has a cumulative PER against stat, but I don't know their methodology for obtaining it. I've been told the number on this page (OPP production) goes deeper than just, "well, Iguodala was playing SF, so what did the opposing SF do," and takes actual matchups into account, but I've never been able to verify that.

There are so many variables that go into "what did his man do," though, that I don't know how you could possibly get a reliable number. It's subjective, when you get right down to it. Who's fault was the blown pick-and-roll? Was it part of the defensive scheme for him to leave that guy in the corner? Was it his fault that he left to cover for someone else? Who do you credit with good defense if a double results in a steal, the doubler, or the man?

What I'm trying to do here, and what I've always struggled with, is a way to get more value out of the defensive stats that are available, and then use that s a springboard to dig deeper into defensive performance on an individual level, and what it means to team defense. This stuff drives me crazy sometimes.

Great job; if the sixers hired a young assistant and used this data to build their roster than our g.m.!s vision may of been realized.

One name not listed who I thought would be perfect in our defend and run style is Taj Gibson, who along with T. Thomas would be perfect here. Maybe that is what they hope Thad can become.

Expand/Contract all comments

Leave a comment